Techn. Hochsch., Diss., 2003--Aachen.

Zeolite catalysis plays an important role in many industrial applications due to their unique properties and has become widely used in the area of oil refining. Of particular interest is Zeolite Y, which can be hydrothermally treated into its ultrastable form, USY. USY offers a superior practicality, especially when dealuminated and metal-loaded. The importance of alkanes hydroisomerization arises from the continuingly stricter regulations imposed on the utilization of gasoline as an automotive fuel. The requirements to reduce the aromatics content in gasoline present a need to find an alternative way to maintain its research octane number (RON). An alternative to gasoline's high-octane aromatic content is to increase the RON for the paraffinic content of gasoline, which can be accomplished through hydroisomerization. Commercially, bifunctional metal-loaded zeolites are used to hydroisomerize the light naphtha stream produced at overheads of atmospheric distillation towers. However, no such process exists for the low-value heavy naphtha cut. This targeted process would, if successful, greatly improve refiner's profitability.In this work, bifunctional USY zeolite catalysts are studied in the hydroisomerization of a normal alkane (nC7, RON = 0). This nC7, found in heavy naphtha, has been used as the 'model' compound. The impact of different reaction conditions and catalyst properties on catalyst activity and stability, in addition to the catalyst selectivity to high octane isomers is one step towards determining optimum conditions and preferential catalyst formulations that favour octane maximization. Six platinum-loaded USY zeolite catalysts, four in-house and two commercial, were tested in an atmospheric glass fixed-bed reactor and a stainless steel reactor purpose-built during the course of this thesis. Reaction temperatures ranged from 170 to 250oC at pressures between 1 and 15 bar. The hydrogen to hydrocarbon molar ratio was fixed at 9, with feed space time ranging from 35.14 to 140.6 kg.s/mol. In-house catalysts were hydrothermally treated at different severities, while commercial ones were originally dealuminated through acid-leaching treatments.Results have shown commercial catalyst CBV-712 gave the best performance and highest octane values for product isomers (>30). In addition, there was no coke generation. The next best catalyst was the most severely steamed in-house catalyst (USY-D) that has shown a remarkable performance at high pressures, almost eclipsing the performance of CBV-712, yet produced higher levels of coke. Other USY catalysts tested were less robust during reactions, probably due to imbalance in their acidic to metallic functions, or diffusion limitations arising from their pore structures. The best catalysts were, nonetheless, highly sensitive to sulfur presence in the feed, which severely impacted their activity, especially their metallic functions, and thus require sulfur-free feeds in order to demonstrate their full capacities. Simple kinetic modelling of experimental data was performed using the initial rates method and estimation of kinetic parameters, whose values were in good agreement with previous literature.

京都大学 / 0048 / 新制・課程博士 / 博士(エネルギー科学) / 甲第19819号 / エネ博第325号 / 新制||エネ||65(附属図書館) / 32855 / 京都大学大学院エネルギー科学研究科エネルギー社会・環境科学専攻 / (主査)教授 東野 達, 教授 坂 志朗, 准教授 亀田 貴之 / 学位規則第4条第1項該当 / Doctor of Energy Science / Kyoto University / DFAM

Direct, efficient oxidation of alkanes is a long-standing goal of catalysis. Gas phase FeO+ can convert methane to methanol and benzene to phenol under thermal conditions. Two key intermediates of these reactions are the [HO-Fe-R]+ insertion intermediate and Fe+(ROH) (R=CH3 or C6H5) exit channel complex. This work describes measurements of the vibrational spectra of these intermediates and electronic structure theory calculations of the potential energy surfaces for the reactions. They help to characterize the mechanism for these reactions. Chapter 1 describes previous studies of methane-to-methanol and benzene-to-phenol conversion by gas-phase transition metal oxide cations. Spectra of gas-phase reaction intermediates are obtained using photofragment spectroscopy, in which absorption of a photon leads to bond breaking. Utuilizing this technique to measure vibrational spectra is challenging, due to the low photon energies involved. Techniques used to measure vibrational spectra of ions - argon tagging, infrared multiple photon dissociation (IRMPD), vibrationally mediated photodissociation (VMP) and infrared laser assisted photodissociation spectroscopy (IRLAPS) are also detailed in chapter 1. The photofragment spectrometer and laser systems used in these studies are described in chapter 2, as is a multi-pass mirror arrangement which I implemented. This greatly improved the quality of vibrational spectra, particularly those measured using IRMPD. Chapter 3 describes studies of the O-H and C-H stretching vibrations of two intermediates of the FeO+ + CH4 reaction. These intermediates are selectively formed by reaction of laser ablated Fe+ with specific organic precursors and are cooled in a supersonic expansion. Vibrations of the sextet and quartet states of the [HO-Fe-CH3]+ insertion intermediate and Fe+(CH3OH) exit channel complex are measured by IRMPD and Ar-tagging. Studies of the O-H stretching vibrations of the [HO-Fe-C6H5] + and Fe+(C6H5OH) intermediates of the FeO+ + C6H6 reaction are discussed in chapter 4. For Fe+(C6H5OH), the O-H stretch is observed at 3598 cm-1. Photodissociation primarily produces Fe+ + C6H5OH. IRMPD of [HO-Fe-C6H5] + mainly produces FeOH+ + C6H5 and the O-H stretch spectrum consists of a peak at ~3700 cm-1 with a shoulder at ~3670 cm-1. Chapter 5 compares three techniques - IRMPD, argon-tagging, and IRLAPS - in the quality of the measured vibrational spectra of Ag+(CH3OH) ions produced under identical conditions. The sharpest spectrum is obtained using IRLAPS. The O-H stretch is observed at 3660 cm-1. Monitoring loss of argon from Ag+(CH3OH)(Ar) gives a slightly broader peak, with no significant shift. The vibrational spectrum obtained using IRMPD is shifted to 3635 cm-1, is substantially broader, and is asymmetrical, tailing to the red.

No description available.

Wetting and phase transitions play a very important role our daily life. Molecularly thin films of long-chain alkanes at solid/vapour interfaces (e.g. C30H62 on silicon wafers) are very good model systems for studying the relation between wetting behaviour and (bulk) phase transitions. Immediately above the bulk melting temperature the alkanes wet partially the surface (drops). In this temperature range the substrate surface is covered with a molecularly thin ordered, solid-like alkane film ("surface freezing"). Thus, the alkane melt wets its own solid only partially which is a quite rare phenomenon in nature. The thesis treats about how the alkane melt wets its own solid surface above and below the bulk melting temperature and about the corresponding melting and solidification processes.<br> Liquid alkane drops can be undercooled to few degrees below the bulk melting temperature without immediate solidification. This undercooling behaviour is quite frequent and theoretical quite well understood. In some cases, slightly undercooled drops start to build two-dimensional solid terraces without bulk solidification. The terraces grow radially from the liquid drops on the substrate surface. They consist of few molecular layers with the thickness multiple of all-trans length of the molecule. By analyzing the terrace growth process one can find that, both below and above the melting point, the entire substrate surface is covered with a thin film of mobile alkane molecules. The presence of this film explains how the solid terrace growth is feeded: the alkane molecules flow through it from the undercooled drops to the periphery of the terrace.<br> The study shows for the first time the coexistence of a molecularly thin film ("precursor") with partially wetting bulk phase. The formation and growth of the terraces is observed only in a small temperature interval in which the 2D nucleation of terraces is more likely than the bulk solidification. The nucleation mechanisms for 2D solidification are also analyzed in this work. More surprising is the terrace behaviour above bulk the melting temperature. The terraces can be slightly overheated before they melt. The melting does not occur all over the surface as a single event; instead small drops form at the terrace edge. Subsequently these drops move on the surface "eating" the solid terraces on their way. By this they grow in size leaving behind paths from were the material was collected. Both overheating and droplet movement can be explained by the fact that the alkane melt wets only partially its own solid. For the first time, these results explicitly confirm the supposed connection between the absence of overheating in solid and "surface melting": the solids usually start to melt without an energetic barrier from the surface at temperatures below the bulk melting point. Accordingly, the surface freezing of alkanes give rise of an energetic barrier which leads to overheating. / Sowohl Benetzung als auch Phasenübergänge spielen eine sehr wichtige Rolle im täglichen Leben. Molekular dünne Filme langkettiger Alkane an Festkörper/Gas-Grenzflächen (z. B. C30H62 an Silizium-Waferoberflächen) sind sehr gute Modellsysteme um die Wechselbeziehung zwischen Benetzungsverhalten und (Volumen-)Phasenübergängen zu untersuchen. In einem Temperaturbereich knapp oberhalb der Volumenschmelztemperatur benetzt die Alkanschmelze die Substratoberfläche nur partiell (Alkantropfen). In diesem Temperaturbereich ist die Substratoberfläche mit einer molekular dünnen, festkörperartig geordneten Alkanschicht bedeckt ("Oberflächengefrieren" ). Die Alkanschmelze benetzt also die eigene Festkörperoberfläche nur partiell, ein in der Natur ziemlich seltenes Phänomen. Die Dissertation beschäftigt sich damit wie die Alkanschmelze ihre eigene Festkörperoberfläche über und unter dem Volumenschmelzpunkt benetzt und mit den entsprechenden Vorgängen beim Schmelzen bzw. Erstarren. Flüssige Alkantropfen lassen sich einige Grad unter ihren Schmelzpunkt unterkühlen ohne sich sofort zu verfestigen. Dieses "Unterkühlungsverhalten" ist üblich und es ist theoretisch qualitativ gut verstanden. Allerdings beobachtet man bei den Alkanen bei leichter Unterkühlung statt einer eventuellen Volumenverfestigung oft die Ausbildung von zweidimensionalen Terrassen aus erstarrtem Alkanen. Die Terrassen wachsen auf der Substratoberfläche radial aus den flüssigen Tropfen. Sie bestehen aus wenigen Alkanlagen mit jeweils der Dicke einer Moleküllänge. Die Analyse der Terrassen-Wachstumsprozesse zeigt, dass die gesamte Substratoberfläche einschliesslich der Terrassen sowohl oberhalb als auch unterhalb der Volumenschmelztemperatur mit einer dünnen Schicht mobiler Alkanmoleküle bedeckt ist. Durch diese Schicht fliessen bei Unterkühlung die Alkane vom unterkühlten Tropfen zur Terrassenkante und liefern den Nachschub für deren Wachstum. Die Untersuchungen zeigen damit erstmalig die Koexistenz eines molekular dünnen Films ("Precursor") mit einer partiell benetzenden Volumenphase. Die Entstehung und das Wachstum der Terrassen wird nur in einem engen Temperaturfenster beobachtet in dem die Keimbildung zweidimensionaler Terrassen wahrscheinlicher ist als die dreidimensionale Volumenverfestigung. Auch dieses Keimbildungsverhalten wird in der Dissertation genauer analysiert. Noch erstaunlicher als das Terrassenwachstum, d. h. das Verfestigungsverhalten ist das Schmelzverhalten der Terrassen. Sie lassen sich bis zu einer gewissen Temperatur überhitzen bevor sie schmelzen! Weiterhin findet bei genügender Überhitzung das Schmelzen nicht gleichzeitig überall statt sondern es entstehen zuerst kleine Alkantropfen an den Terrassenrändern. Diese bewegen sich dann über die Substratoberfläche und "fressen" sich durch die festen Terrassen. Dabei wachsen sie weil sie das geschmolzene Material aufnehmen und hinterlassen eine alkanfreie Spur. Sowohl die Überhitzung als auch die Tropfenbewegung lassen sich damit erklären dass die flüssige Alkanschmelze ihren eigenen Festkörper nur partiell benetzt. Die Ergebnisse bestätigen erstmals explizit den seit vielen Jahrzehnten vermuteten Zusammenhang zwischen der üblicherweise nicht beobachtbaren Überhitzung von Festkörpern und Oberflächenschmelzen: Festkörper beginnen normalerweise ohne Energiebarriere von der Oberfläche an zu schmelzen. Entsprechend bildet das Oberflächengefrieren der Alkane eine Energiebarriere und erlaubt damit deren Überhitzen.

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal.

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal

Master of Science / Department of Physics / Christopher M. Sorensen / Gold Nanoparticles (AuNP) 5nm in diameter, ligated with n-dodecanethiol, were dissolved in various hydrocarbon solvents including normal alkanes from n-hexane to n-hexadecane as well as two aromatics, toluene and para-xylene. These solutions were centrifuged at room temperature under 12000g acceleration for one hour to separate larger clusters from AuNP monomers dissolved in the supernatants. UV-Vis absorbance data were taken on the supernatants and were then converted to concentrations in moles of Au atoms/L. These concentrations correspond to the saturated concentration of dissolved AuNP monomers in equilibrium with a precipitate at room temperature. For the alkanes, we discovered a non-monotonic functionality of saturated concentration vs. solvent chain length with a maximum corresponding to n-dodecane. This agreed with predictions made of the ligands’ interactions with the solvents based on comparisons of solubility parameters where the n-dodecanethiol ligands were approximated as n-dodecane. The concentrations of AuNPs when dissolved in the aromatics did not follow the trend predicted by solubility parameters.

Thesis (MScEng (Process Engineering))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: Hydrocarbons provide excellent feed stocks for bioconversion processes to produce value added products using various micro-organisms. However, hydrocarbon-based aerobic bioprocesses may exhibit transport problems where the bioconversion is limited by oxygen supply rather than reaction kinetics. Consequently, the overall volumetric oxygen transfer coefficient (KLa) becomes critical in designing, operating and scaling up of these processes. In view of KLa importance in hydrocarbon-based processes, this work evaluated KLa measurement methodologies as well as quantification of KLa behavior in aerated agitated alkane-solid-aqueous dispersions. A widely used KLa measurement methodology, the gassing out procedure (GOP) was improved. This improvement was done to account for the dissolved oxygen (DO) transfer resistances associated with probe. These resistances result in a lag in DO response during KLa measurement. The DO probe response lag time, was incorporated into the GOP resulting in the GOP (lag) methodology. The GOP (lag) compared well with the pressure step procedure (PSP), as documented in literature, which also incorporated the probe response lag time. Using the GOP (lag), KLa was quantified in alkane-solid-aqueous dispersions, using either inert compounds (corn flour and CaCO3) or inactive yeast cells as solids to represent the micro-organisms in a hydrocarbon bioprocess. Influences of agitation, alkane concentration, solids loading and solids particle sizes and their interactions on KLa behavior in these systems were quantified. In the application of an accurate KLa measurement methodology, the DO probe response lag time was investigated. Factors affecting the lag, which included process conditions such as agitation (600-1200rpm), alkane concentration (2.5-20% (v/v), alkane chain length (n-C10-13 and n-C14-20), inert solids loading (1-10g/L) and solids particle sizes (3-14μm) as well as probe characteristics such as membrane age and electrolyte age (5 day usage), were investigated. Kp, the oxygen transfer coefficient of the probe, was determined experimentally as the inverse of the time taken for the DO to reach 63.2% of saturation after a step change in DO concentration. Kp dependence on these factors was defined using 22 factorial design experiments. Kp decreased on increased membrane age with an effect double that of Kp decrease due to electrolyte age. Additionally, increased alkane concentration decreased Kp with an effect 7 times higher compared to that of Kp decrease due to increased alkane chain length. This was in accordance to Pareto charts quantification. KLa was then calculated, using the GOP (lag), according to equation [1] which incorporates the influence of Kp. Equation 1 is derived from the simultaneous solution of the models which describe the response of the system and of the probe to a step change in DO. 1 1 * L p p p K at K t L p p La C K e K ae C K K = -  - - -  -   [1] The KLa values documented in literature from the PSP and KLa calculated by the GOP (lag) showed only a 1.6% difference. However KLa values calculated by the GOP (lag) were more accurate than KLa calculated by the GOP, with up to >40% error observed in the latter according to t-tests analyses. These results demonstrated that incorporating Kp markedly improved KLa accuracy. Consequently, the GOP (lag) was chosen as the preferred KLa measurement methodology. KLa was determined in n-C14-20-inert solid-aqueous dispersions. Experiments were conducted in a stirred tank reactor with a 5L working volume at constant aeration of 0.8vvm, 22ºC and 101.3kPa. KLa behavior across a range of agitations (600- 1200rpm), alkane concentrations (2.5-20% (v/v)), inert solids loadings (1-10g/L) and solids particle sizes (3-14μm) was defined using a 24 factorial design experiment. In these dispersions, KLa increased significantly on increased agitation with an effect 5 times higher compared to that of KLa increase due to interaction of increased alkane concentration and inert solids loading. Additionally, KLa decreased significantly on increased alkane concentration with an effect 4 times higher compared to both that of increased solids particle sizes and the interaction of increased agitation and solids particle size. In n-C14-20-yeast-aqueous dispersions, KLa was determined under narrowed process conditions better representing typical bioprocess conditions. KLa behavior across a range of agitations (600-900rpm), alkane concentrations (2.5-11.25% (v/v)) and yeast loadings (1-5.5g/L) using a 5μm-yeast cell was defined using a 23 factorial design experiment. In these dispersions, KLa increased significantly on increased agitation. Additionally, KLa decreased significantly on increased yeast loading with an effect 1.2 times higher compared to that of KLa decrease due to interaction of increased alkane concentration and yeast loading. In this study, the importance of Kp for accurate KLa measurement in alkane based systems has been quantified and an accurate and less complex methodology for its measurement applied. Further, KLa behavior in aerated alkane-solid-aqueous dispersions was quantified, demonstrating KLa enhancement on increased agitation and KLa depression on increased alkane concentration, solids loading and solids particle sizes. / AFRIKAANSE OPSOMMING: Koolwaterstowwe dien as uitstekende voervoorraad vir ´n verskeidenheid van mikroorganismes wat aangewend word in biologiese omsettingsprosesse ter vervaardiging van waardetoevoegende produkte. Hierdie biologiese omsettingsprosesse word egter vertraag weens die gebrek aan suurstoftoevoer onder aerobiese toestande. Die tempo van omsetting word dus beheer deur die volumetriese suurstofoordragkoeffisiënt (KLa) eerder as die toepaslike reaksiekinetika. Die bepaling van ´n akkurate KLa word dus krities tydens die ontwerp en opskalering van hierdie prosesse. Met dit in gedagte het hierdie studie die huidige metodes om KLa te bepaal geëvalueer en die gedrag van KLa in goed vermengde en belugde waterige alkaanmengsels met inerte vastestowwe, soos gisselle, in suspensie ondersoek. ´n Deesdae populêre metode om KLa te bepaal, die sogenaamde gasvrylatingsprosedure (GOP) is in hierdie studie verbeter. Die verbetering berus op die ontwikkeling van ´n prosedure om die suurstofoordragsweerstand van die pobe wat die hoeveelheid opgeloste suurstof (DO) meet, in berekening te bring. Hierdie weerstand veroorsaak ´n vertragin in the responstyd van die probe. Die verbeterde metode, GOP (lag), vergelyk goed met die gepubliseerde resultate van die drukstaptegniek (PSP) wat ook die responstyd in ag neem. GOP (lag) is ingespan om KLa te gekwantifiseer vir waterige alkaan-vastestof suspensies. Inerte componente soos mieliemeel, kalsiumkarbonaat en onaktiewe gisselle het gedien as die vastestof in suspensie verteenwoordigend van die mikroörganismes in ´n koolwaterstof bio-proses. Die invloed van vermengingstempo, alkaan konsentrasie, vastestof konsentrasie en partikelgrootte asook die interaksie van al die bogenoemde op KLa is kwatitatief bepaal in hierdie studie. Faktore wat die responstyd van die DO probe beïnvloed is ondersoek. Hierdie faktore is onder meer vermengingstempo (600-1200opm), alkaankonsentrasie (2.5-20% (v/v)), alkaankettinglengte (n-C10-13 en n-C14-20), vastestofkonsentrasie (1-10g/L) en partikelgrootte (3-14 μm). Faktore wat die eienskappe van die probe beïnvoed, naamlik membraan-en elektrolietouderdom (5 dae verbruik), is ook ondersoek. Kp, die suurstofoordragskoeffisiënt, is bepaal deur ´n inkrementele verandering in die suurstofkonsentrasie van die mengsel te maak en die tyd vir 63.2% versadiging van die probelesing te noteer. Die genoteerde tyd is die response tyd van die probe en Kp, die inverse van hierdie tyd. Die afhanklikheid van Kp op die bogenoemde faktore is ondersoek in ´n 22 faktorieël ontwerpte reeks eksperimente. Kp toon ´n afname met ´n toename in membraanouderdom. Hierdie afname is dubbel in grootte as dit vergelyk word met die afname relatief tot die toename in elektrolietouderdom. Verder toon Kp ´n afname met ´n toename in alkaankonsentrasie. Hierdie afname is 7 keer groter relatief tot die afname gesien met die toename in alkaan kettinglengte. Hierdie is in goeie ooreenstemming met Pareto kaarte as kwantifiseringsmetode. KLa is bereken met die inagname van Kp volgens vergelyking [1]: 1 1 * L p p p K at K t L p p La C K e K ae C K K = -  - - -  -   [1] Vergelyking [1] is afgelei vanaf die gelyktydige oplossing van die bestaande modelle wat die responstyd van die pobe vir ´n stapverandering in DO bereken. Die KLa waardes van die PSP metode uit literatuur verskil in die orde van 1.6% van dié bereken deur vergelyking [2]. Hierdie verskil is weglaatbaar. Die KLa waardes verkry uit die GOP metode wat nie Kp in berekening bring nie, verskil met meer as 40% van die huidige, verbeterde metode volgens die statistiese t-test analiese. Dit bewys dat die inagname van Kp ´n merkwaardige verbetering in die akuraatheid van KLa teweeg bring. GOP (lag) kry dus voorkeur vir die berekening van KLa verder aan in hierdie studie. KLa is bereken vir n-C14-20-water mengsels met inerte vastestofsuspensies. Die eksperimente is uitgevoer in ´n 5L geroerde reaktor met ´n konstante belugting van 0.8vvm (volume lug per volume supensie per minuut), 22ºC en 101.3kPa. Die gedrag van KLa met betrekking tot vermengingstempo (600-1200opm), alkaankonsentrasie (2.5-20% (v/v)), vastestofkonsentrasie (1-10g/L) en partikelgrootte (3-14μm) is ondersoek in ´n 24 faktorieël ontwerpte reeks eksperimente. Verder is die invloed van vloeistofviskositeit en oppervlakspanning op KLa ondersoek in ´n 23 faktorieël ontwerpte reeks eksperimente. KLa het ´n beduidende toename getoon met ´n toename in vermengingstempo. Hierdie toename was 5 keer groter as die toename relatief tot die interaksie van alkaan-en vastestofkonsentrasie. KLa het ook beduidend afgeneem met ´n toename in alkaankonsentrasie. Die toename was 4 keer groter as die toename relatief tot die toename in partikelgrootte en die interaksie van vermengingstempo en partikelgrootte. In n-C14-20-water mengsels met gisselsuspensies is KLa bepaal onder kondisies verteenwoordigend van tipiesie biologiese omsettingsprosesse. Die gedrag van KLa met betrekking tot vermengingstempo (600-900opm), alkaankonsentrasie (2.5-11.25% (v/v)) en giskonsentrasie (1-5.5g/L) met ´n partikelgroote van 5μm is ondersoek in ´n 23 faktorieël ontwerpte reeks eksperimente. Hierdie eksperimente het ´n beduidende toename in KLa met ´n toename in vermengingstempo getoon sowel as ´n beduidende afname met ´n toename in giskonsentrasie. Hierdie afname is in die orde van 1.2 keer groter in vergelyking met die interaksie van alkeen- en giskonsentrasie. Hierdie studie bring die kritieke rol wat Kp speel in die akkurate bepaling van KLa in waterige alkaansisteme met inerte vastestofsuspensies na vore. Dit stel verder ´n metodiek voor vir die akurate meting van en kwantifisering van beide Kp en KLa onder aerobiese toestande met betrekking tot vermengingstempo, alkaankonsentrasie, vastestofkonsentrasie en partikelgrootte.