No description available.

Interest exists in the development of transparent conducting oxide materials, which have diverse applications in areas such as transparent coatings for display technologies, solar cells, and optoelectronics. Since many of the applications require the use of thin film forms, the need is to establish useful experimental approaches to the fabrication of such structures. One relatively new method in this area is oxygen-plasma assisted atomic beam epitaxy (OPABE) in which oxide layers are grown under normal molecular beam epitaxy (MBE) conditions with the addition of an oxygen atom beam to ensure full oxidation of the depositing metallic species. Work in this area has to date mainly focussed on the growth of relatively stable oxides such as ZnO, MgO and In₂O₃ which are the strongly thermodynamically favoured reaction products, across a broad range of reaction conditions. In contrast, the present work is concerned with the growth of Cu2O and a range of delafossite materials, namely CuInO₂, CuCrO₂ and CuGaO₂, which are expected to require much more sensitive control to achieve the desired reaction product. Studies of the OPABE growth of Cu₂O on MgO (100) and MgO (110) substrates have been carried out, using a broad range of physical techniques to characterise the grown Cu₂O deposits. It is demonstrated that CuO is the favoured reaction product at low growth temperatures, although Cu₂O becomes increasingly favoured as the growth temperature increases. Alternatively, it is also shown that a novel bilayer growth method, whereby some pure Cu is deposited prior to oxide growth, can be used to form the desired Cu (I) phase. Varying crystal orientations are seen, depending on the exact growth conditions; core level and valence band X-ray photoelectron spectroscopy (XPS), optical band gap and atomic force microscopy (AFM) measurements are used to characterise the deposits. Further growth investigations of the delafossite compounds CuInO₂, CuCrO₂ and CuGaO₂ using OPABE are also recorded, and for the case of CuInO₂, comparison is also made with the pulsed laser deposition approach. For all three materials systems, oriented crystal growth on basal planes sapphire substrates is seen, with either the (001) plane or the (015) plane orienting parallel to the substrate depending on the growth temperature, provided approximately correct metal fluxes are used as set by the Knudsen-cell temperatures. The typical valence band electronic structure of delafossite materials is observed in all three cases, and XPS peak shifts suggest that the layers can be electrically doped by adding appropriate metal fluxes during growth. AFM measurements show the grown films are relatively rough and it is suggested that the growth mode follows an island growth mechanism in which oriented three dimensional islands formed at the start of growth gradually enlarge and coalesce as the film thickens. Optical absorption measurements are consistent with the generally accepted optical band gaps of the materials concerned.

Submitted by DANILO FRANCO null (franco.danilo@gmail.com) on 2018-01-08T23:48:13Z No. of bitstreams: 1 Tese_Danilo_Franco.pdf: 1663923 bytes, checksum: 2d8e6e2bc505a1017e531d2fb608942f (MD5) / Approved for entry into archive by Alexandra Maria Donadon Lusser Segali null (alexmar@fcav.unesp.br) on 2018-01-09T09:37:21Z (GMT) No. of bitstreams: 1 franco_d_dr_jabo.pdf: 1663923 bytes, checksum: 2d8e6e2bc505a1017e531d2fb608942f (MD5) / Made available in DSpace on 2018-01-09T09:37:21Z (GMT). No. of bitstreams: 1 franco_d_dr_jabo.pdf: 1663923 bytes, checksum: 2d8e6e2bc505a1017e531d2fb608942f (MD5) Previous issue date: 2017-11-13 / Neste estudo, foram realizados três experimentos (E1, E2 e E3). Os dois primeiros experimentos (E1 e E2) referem-se a avaliações da eficiência dose-resposta dos fungicidas hidróxido de cobre (HC), oxicloreto de cobre (OC) e óxido cuproso (OCp), aplicados nos estádios F1 e F2, no controle de Phyllosticta citricarpa, agente causal da mancha preta dos citros (MPC). E, no terceiro experimento (E3), foi avaliada a influência da aplicação de piraclostrobina em consonância com eventos chuvosos, no controle da MPC. Em E1 e E2 os fungicidas e dosagens avaliadas foram (g de cobre metálico 100 L-1 de calda): HC (43,7; 35,0; e 26,2; OC (100,8; 90,7; e 68,0) e; OCp (90,0; 56,3; e 42,2). No experimento E1 foram utilizadas plantas de laranjeira ‘Pera’, sob condições naturais de infecção, em Olímpia/SP. Foi adotado o delineamento experimental de blocos ao acaso, com quatro repetições. Cada parcela foi constituída de 33 plantas. Os fungicidas foram aplicados mediante turbopulverizador e vazão de 135 mL de calda m-3 de copa. Posteriormente, em todos os tratamentos, a partir do estádio F4 foram realizadas quatro aplicações de azoxistrobina a 75 g de i.a. 2.000 L-1, em intervalo de 42 dias. Foram realizadas quatro avaliações mensais, de agosto a novembro, amostrando-se casualmente 100 frutos nas seis plantas centrais da parcela, determinando-se a incidência (I) e severidade dos sintomas mediante escala de notas. A partir dos dados de severidade dos sintomas foi determinado o índice de doença (ID). Com os dados de ID, ao longo do tempo, calculou-se a área abaixo da curva de progresso da doença (AACPD). Em E2 foram empregadas plantas de laranjeira ‘Pera‘, também em Olímpia/SP. Os tratamentos avaliados foram semelhantes aos anteriormente citados, sendo, porém, aplicados com pulverizador manual do tipo “De Vilbs”, com volume suficiente para cobertura uniforme dos frutos. Foi adotado o delineamento de blocos ao acaso, em arranjo fatorial 10 tratamentos, sendo nove combinações de fungicidas cúpricos e uma testemunha, sem fungicida) x 5 (épocas de inoculação) e 4 (repetições). Cada parcela foi constituída por 20 frutos previamente ensacados com papel semi-kraft, no estádio F1. Posteriormente, no estádio F4, procedeu-se remoção dos sacos de papel, seguido do tratamento dos frutos exclusivamente com fungicidas cúpricos. O tratamento dos frutos deu-se mediante sua imersão, durante x segundos, na calda fungicida acrescida de óleo mineral a x%. Posteriormente, após secagem da calda procedeu-se o reensacamento dos frutos. A seguir, em intervalo semanal, entre 0 a 28 dias após o tratamento, foi realizada inoculação mediante aspersão de suspensão contendo 104 conídios mL-1. Na avaliação, realizada quando da colheita da maturação dos frutos, foram determinados os valores de I e ID. O experimento E3 foi realizado em pomar comercial de laranjeira ‘Folha Murcha’, no município de Paraíso/SP. Inicialmente, nos estádios F1 e F2 foram realizadas duas aplicações de oxicloreto de cobre (90 g de cobre metálico 100 L-1 de calda). Os tratamentos avaliados foram: aplicação do fungicida piraclostrobina em intervalo de 14 dias entre si, na dose de 25 g 2000 L-1; aplicação de piraclostrobina, na dose de 75 g 2000L-1 após o primeiro evento de chuva ocorrido 14 dias após a última aplicação do tratamento; aplicação de piraclostrobina, na dose de 75 g 2000L-1 após o primeiro evento de chuva, ocorrido 35 dias após a última aplicação do tratamento; aplicação de piraclostrobina, com intervalo de 42 dias entre si nas doses de 50 e 75 g 2000L-1. Esses tratamentos foram comparados com o tratamento padrão de azoxistrobina, na dose de 75 g 2000L-1, aplicado em intervalo de 42 dias. Foi utilizado o delineamento de blocos casualizados, e cada parcela foi constituída por 66 plantas. As condições operacionais adotadas para aplicação dos fungicidas foram semelhantes às adotadas em E1, porém com 150 mL de calda m-3 de copa. Os critérios de avaliações foram semelhantes aos citados em E1, em intervalos mensais, de setembro a janeiro. Concluiu-se que os fungicidas HC, OC e OCp nas doses de 43,7, 90,7 e 56,3 g de cobre metálico 100 L-1, respectivamente, aplicados nos estádios F1 e F2, complementado por aplicações de azoxistrobina, foram igualmente eficientes no controle de MPC. A eficiência do efeito protetor dos fungicidas cúpricos foi constatada por pelo menos 28 dias após a sua aplicação. Porém, a maior eficiência dos fungicidas cúpricos deu-se quando da sua aplicação previamente à inoculação de P. citricarpa. Aplicações de piraclostrobina a 75 g de i.a. 2000L-1 em intervalo de no mínimo 14 dias entre eventos chuvosos resultaram em maior controle da doença, quando comparado aos demais intervalos e doses de piraclostrobina e a azoxistrobina, na dose de 75 g 2.000 L-1, em intervalo fixo de 42 dias, com consequente redução da incidência e severidade de sintomas de MPC. / In this study, three experiments were performed (E1, E2 and E3). The first two experiments refer to the rate-response efficiency of the fungicides copper hydroxide (HC), copper oxychloride (OC) and cuprous oxide (OCp) in the control of Phyllosticta citricarpa, causal agent of citrus black spot (CBS). And the third (E3) the influence of pyraclostrobin application was evaluated in consonance with rainy events. In E1 and E2 the evaluated rates were: HC (43.7; 35.0; and 26.2 g of metallic copper 100L-1); OC (100.8, 90.7, and 68.0 g of metallic copper 100L-1); OCp (90.0, 56.3, and 42.2 g of metallic copper 100L-1). The experiment (E1) was conducted under natural infection by P. citricarpa in 'Pera' sweet orange planted in Olímpia/SP. A randomized blocks design was used, with four replications. Each plot was represented by 33 plants. The fungicides were applied in stages F1 and F2 by turbosprayer and 135 ml m-3 of the tree canopy. Subsequently, in F4 stage, all treatments were treated with azoxystrobin at 75 g 2000 L-1, four times in a 42-day interval. Four monthly evaluations were performed, from August to November, evaluating 100 fruits to determining the incidence and symptoms severity. Data were used to determined area under disease progress curve (AUDPC). In E2, a 'Pera' sweet orange plants were used, also located in Olímpia/SP. The treatments evaluated were similar those previously in E1, but were applied by a manual "De Vilbs" sprayer, with sufficient volum for uniform fruit coverage. A randomized blocks design were used in a factorial arrangement of 10 treatments (9 copper fungicides treatments and one untreated check) x 5 (inoculation times) and 4 (replicates). Each plot was formed by 20 pre-bagged fruits with semi-kraft paper in the F1 stage. In the F4 stage the bags were removed and the fruits sprayed exclusively with copper fungicides. Subsequently, in a weekly interval from 0 to 28 days after application at inoculation with suspension spray containing 104 conidia mL-1 was performed. In the harvest time an evaluation was performed and the values of I and DI were determined. E3 was carried out in a commercial 'Folha Murcha' sweet orange orchard, in Paraíso/SP. Initially, in stages F1 and F2, two applications of copper oxychloride based on 90 g of metallic copper 100 L-1 were perfomed. The treatments evaluated were: application with 14 days fixed interval in the pyraclostrobin rate of 25 g 2000L-1; application of 75 g 2000L-1 after the first rain event occurred 14 days after the last fungicide application; application of 75 g 2000L-1 after the first rain event occurred 35 days after the last fungicide application; application with 42 days fixed interval of 50 and 75 g 2000L-1. These treatments were compared with the standard treatment with 42 days fixed interval of azoxystrobin at rate 75 g 2000L-1. A randomized blocks design was used with 66 plants per plot. The fungicides applications were similar that in E1, but with 150 mL m-3 of canopy volume. The evaluations were similar that in E1, in monthly intervals, from September to January, in the ten central plants per plots. It was concluded that the fungicides HC, OC and OCp in the rates of 43.7, 90.7 and 56.3 g of metallic copper 100 L-1, respectively, applied in stages F1 and F2, complemented by applications of azoxystrobin, were equally efficient to control CBS. The protective effect of the copper fungicides was maintained until 28 days after treatment. However, the best response was obtained when inoculating P. citricarpa immediately after copper fungicides treatment. The 75 g 2000L-1 of pyraclostrobin application after the first rainy event occurred at least 14 days from the last fungicide application resulted in a greater reduction in the incidence and severity of CBS symptoms.

Une étude de l’oxydation d’une poudre de nanoparticules de cuivre a été menée à 120 - 145°C sous des pressions partielles d’oxygène allant de 1 à 40 kPa. La réaction a été suivie par thermogravimétrie afin d’obtenir les données cinétiques. Des caractérisations chimiques, texturales et morphologiques de la poudre ont été réalisées à différents moments de la transformation. La cuprite (Cu2O) (produit unique) de la réaction croît d’une manière anisotrope et par développement externe autour de la particule initiale qui devient creuse. Une diminution de la surface spécifique et de la porosité de la poudre au cours de la transformation a été mise en évidence.Des tests cinétiques ont montré l’existence d’une étape limitante de croissance jusqu’à un taux de conversion de 0,7 à 140°C. Il a également été montré que pour P(O2) ≤ 4 kPa, les processus de germination et de croissance de l’oxyde interviennent simultanément pendant la réaction et que l’adsorption de l’oxygène est l’étape limitante. Pour P(O2) ≥ 20 kPa, la germination se fait instantanément au début de la transformation dont la vitesse est contrôlée par le processus de croissance, la diffusion du cuivre étant alors l’étape limitante. Deux modèles ont été construits et testés avec succès pour décrire la cinétique dans les deux gammes de P(O2) jusqu’à un taux de conversion donné. Pour expliquer le ralentissement observé au-delà de ce taux de conversion et pour P(O2) ≤ 4 kPa, le modèle a été couplé aux phénomènes de transfert de chaleur et de matière au sein des agglomérats. Ce couplage permet d’évaluer l’hypothèse d’un ralentissement de la réaction par la diffusion des molécules d’oxygène dans les pores de l’agglomérat. / The oxidation of copper nanoparticles at 120 – 145°C was investigated using TGA, DSC, XRD, SEM, TEM and nitrogen adsorption techniques (BET, BJH,..). Isothermal and isobaric studies of the oxidation reaction were carried out under various oxygen partial pressures 1 kPa – 40 kPa. The cuprous oxide (Cu2O) (the unique product of the reaction) grows in an anisotropic manner by outward diffusion of the copper. A considerable decrease of the specific surface area and the porosity of the powder during the transformation was highlighted.It was found that working under P(O2) ≤ 4 kPa leads to reaction where nucleation of Cu2O is in competition with its growth. The study of the dependency of the growth rate on the oxygen partial pressure has shown the adsorption of oxygen to be the rate-determining step. However, when the reaction is conducted under P(O2) ≥ 20 kPa, the nucleation occurs instantaneously in the early beginning of the reaction which will be governed by the growth process. Under these latter conditions of oxygen partial pressure the diffusion of copper ionized vacancies becomes the rate determining step. Hence, two kinetic models have been established to interpret the experimental curves in the two different oxygen partial pressure ranges. The expression of the extent of conversion was successfully confronted to the kinetic data up to an extent of conversion corresponding to the slowdown of the reaction. The kinetic model for P(O2) ≤ 4 kPa was coupled with mass and heat transfer within the porous agglomerate to verify if the diffusion of oxygen molecules in pores is in the origin of the reaction slowdown, this latter hypothesis was found not satisfactory.

<p>Self-oscillating behaviour in alkaline Cu(II)-lactate and -tartrate systems has been investigated by in-situ pH and confocal Raman spectroscopy measurements. Formation of Cu(II)-lactate and -tartrate complexes is a key factor underlying the self-oscillations. Dynamic processes in the diffusion layer have been probed to give a better understanding of the self-oscillating process.</p><p>The self-oscillating behaviour is found to be an effect of pH variations in the diffusion layer. Mainly copper is deposited at lower pH values and potentials; at the same time, the pH increases. This is an effect of the dissociation of the Cu(II)-complex during electrochemical reduction. The absence of a buffer within a given pH region is crucial to the fast and sudden pH increase and thereby to the positive potential shift, where cuprous oxide is deposited. A precipitation reaction probably decreases the pH again, leading to a negative potential shift, and copper again begins to deposit. The concentration and strength of the buffer in the electrolyte affect the appearance of the oscillation pattern. The pH and temperature of the bulk electrolyte also influence the self-oscillations. The deposit consists of copper and cuprous oxide, where the composition of the phases deposited is a function of the working-electrode potential. Cuprous oxide is deposited at the higher potentials and mainly copper at the lower potentials.</p><p>Finally, two-dimensional arrays of Cu/Cu₂O microcylinders have been deposited using the Cu(II)-lactate system through the application of a template method.</p>

The foci of this thesis are the metal oxides Cu2O, ZnO and Fe3O4 and their interaction with water and sulfur dioxide (SO2). The intention is to study SO2-induced atmospheric corrosion on a molecular level. All studies are based on photoelectron spectroscopy (PES) and scanning tunneling microscopy (STM) measurements. The band structure of Cu2O in the Γ-M direction has been probed by angle-resolved PES (ARPES). It reveals a more detailed picture of the bulk band structure than earlier data and gives the first experimental evidence of a dispersive hybridized Cu 3d-Cu 4s state. The experimental data is compared to band structure calculations. The structure of clean metal oxide surfaces and impact of sample preparation have been studied. Oxygen vacancies can form a (√3x√3)R30° reconstruction on Cu2O(111). Oxygen atoms adjacent to copper vacancies, steps or kinks are shown to be adsorption sites for both water and SO2. Annealing temperature influences the defect density and hydrogen content in ZnO, which can have large impact on the surface properties of ZnO(0001). Water is shown to adsorb dissociatively on ZnO(0001) and partly dissociatively on Cu2O(111). The dissociation occurs at undercoordinated oxygen sites on both surfaces. Water stays adsorbed on ZnO(0001) at room temperature but on Cu2O(111), all water has desorbed at 210 K. SO2 interacts with one or two undercoordinated O-sites on all studied oxide surfaces forming SO3 or SO4 species respectively. SO4 on Fe3O4(100) follows the (√2x√2)R45° reconstruction. On Cu2O(111) and ZnO(0001), SO2 adsorbs on defect sites. An SO3 to SO4 transition is observed on Cu2O(111) when heating an SO3 adsorbate layer from 150 K to 280K. Coadsorption of water and SO2 on ZnO(0001) and Fe3O4(100) has been studied briefly. Water blocks SO2 adsorption sites on ZnO(0001). On Fe3O4(100) and on one type of reduced ZnO(0001) sample, SO2 dissociation to atomic sulfur or sulfide occurs to a higher extent on water exposed surfaces than on clean surfaces. Water thus appears to increase the charge density on some surfaces. Further studies are needed to reveal the cause of this unexpected effect. / <p>QC 20110516</p>

Self-oscillating behaviour in alkaline Cu(II)-lactate and -tartrate systems has been investigated by in-situ pH and confocal Raman spectroscopy measurements. Formation of Cu(II)-lactate and -tartrate complexes is a key factor underlying the self-oscillations. Dynamic processes in the diffusion layer have been probed to give a better understanding of the self-oscillating process. The self-oscillating behaviour is found to be an effect of pH variations in the diffusion layer. Mainly copper is deposited at lower pH values and potentials; at the same time, the pH increases. This is an effect of the dissociation of the Cu(II)-complex during electrochemical reduction. The absence of a buffer within a given pH region is crucial to the fast and sudden pH increase and thereby to the positive potential shift, where cuprous oxide is deposited. A precipitation reaction probably decreases the pH again, leading to a negative potential shift, and copper again begins to deposit. The concentration and strength of the buffer in the electrolyte affect the appearance of the oscillation pattern. The pH and temperature of the bulk electrolyte also influence the self-oscillations. The deposit consists of copper and cuprous oxide, where the composition of the phases deposited is a function of the working-electrode potential. Cuprous oxide is deposited at the higher potentials and mainly copper at the lower potentials. Finally, two-dimensional arrays of Cu/Cu2O microcylinders have been deposited using the Cu(II)-lactate system through the application of a template method.

This talk was presented in the 14th International Conference on Atomic Layer Deposition (ALD 2014) in Kyoto, Japan on 18th June 2014. Abstract Atomic Layer Deposition (ALD) is emerging as a ubiquitous method for the deposition of conformal and homogeneous ultra-thin films on complex topographies and large substrates in microelectronics. Electrochemical deposition (ECD) is the first choice for the deposition of copper (Cu) into the trenches and vias of the interconnect system for ULSI circuits. The ECD of Cu necessitates an electrically conductive seed layer for filling the interconnect structures. ALD is now considered as a solution for conformal deposition of Cu seed layers on very high aspect ratio (AR) structures also for technology nodes below 20 nm, since physical vapor deposition is not applicable for structures with high AR. Cu seed layer deposition by the reduction of Cu2O, which has been deposited from the Cu(I) β-diketonate [(nBu3P)2Cu(acac)] (1) used as Cu precursor, has been successfully carried out on different substrates like Ta, TaN, SiO2, and Ru [1, 2]. It was found that the subsequent gas-phase reduction of the Cu2O films can be aided by introducing catalytic amounts of a Ru precursor into the Cu precursor, so that metallic copper films could potentially obtained also on non-catalytic substrates [3, 4]. In this work, in situ X-ray photoelectron spectroscopy (XPS) investigation of the surface chemistry during Cu2O ALD from the mixture of 99 mol % of 1 and 1 mol % of [Ru(η5 C5H4SiMe3)(η5-C7H11)] (2) as ruthenium precursor, and the reduction of Cu2O to metallic Cu by formic acid carried out on SiO2 substrate are demonstrated. Oxidation states of the Cu in the film are identified by comparing the Cu Auger parameter (α) [5] with literature data. α calculated after ALD equals 362.2 eV and after reduction equals 363.8 eV, comparable to the Cu2O and metallic Cu in thin-films [6] respectively. In addition, <10 % of Cu(I), Cu(II), and Cu(OH)2 species are identified from the Cu 2p3/2 and Cu L3VV Auger spectrum after reduction. Consequently, the ALD Cu2O is successfully reduced to metallic copper by in-situ thermal reduction using formic acid. [1] T. Waechtler et al., J. Electrochem. Soc., 156 (6), H453 (2009). [2] T. Waechtler et al., Microelectron. Eng., 88, 684 (2011). [3] S. Mueller et al., Conference Proceedings SCD 2011, Semiconductor Conference Dresden, pp. 1-4. [4] T. Waechtler et al., US Patent Application Publication, US 2013/0062768. [5] C. D. Wagner, Faraday Discuss. Chem. Soc., 60, 291 (1975). [6] J. P. Espinós et al., J. Phys. Chem. B, 106, 6921 (2002).

京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第22775号 / 工博第4774号 / 新制||工||1747(附属図書館) / 京都大学大学院工学研究科材料工学専攻 / (主査)教授 邑瀬 邦明, 教授 杉村 博之, 教授 宇田 哲也 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

Acid/Base characterizations of metal oxide surfaces are often used to explain their catalytic behavior. However, the vast majority of these studies have been performed on powders or supported oxides, and there is very little information available in the literature on the interaction of acid/base probe molecules with well-defined oxide surfaces of known coordination geometry and oxidation state. The well-defined, single crystal surfaces of Cu₂O (111), SnO₂ (110), and Cr₂O₃ (101̲2) were investigated for their acid/base properties by the interactions between the probe molecules and the well-defined surface features. The adsorption of NH₃ at cation sites was used to characterize the Lewis acidity of SnO₂ (110) and Cu₂O (111) surfaces. The adsorption of CO₂, a standard acidic probe molecule, was used to characterize the Lewis basicity of the oxygen anions on SnO₂ (110), Cu₂O (111) , and Cr₂O₃ (101̲2) surfaces. BF₃, while not a standard probe molecule, has been tested as a probe of the Lewis basicity of the oxygen anions on SnO₂ (110) and Cr₂O₃ (101̲2). By studying probe molecules on well-defined metal oxide surfaces with known coordination geometry and oxidation state, an overall evaluation of NH₃, CO₂, and BF₃ as probe molecules can be made using the surfaces studied. NH₃ probed differences in Lewis acidity of Sn cations on SnO₂ (110), which had differences in coordination environments and oxidation states. But, NH₃ adsorption failed to provide any direct information on differences in Lewis acidity of Cu cations in different local coordination geometries on Cu₂O (111). CO₂ is a poor probe of the Lewis basicity of oxygen anions on the metal oxide surfaces studied here. CO₂ does not strongly adsorb to either SnO₂ (110) or Cu₂O (111). On Cr₂O₃ (101̲2), CO₂ does interact with oxygen sites but in two different coordinations, which vary with surface condition, making a comparison of basicity difficult. In the cases studied here, CO₂ either does not adsorb, or it does not provide a clear set of results that can be related simply to Lewis basicity. BF₃ seems to be a much better probe of the Lewis basicity than CO₂ for the well-defined metal oxide surfaces studied here. On SnO₂ (110) and Cr₂O₃ (101̲2), the boron atom of BF₃ directly interacts with oxygen sites by accepting their electrons. BF₃ thermal desorption seems to provide a direct measure of the Lewis basicity of different surface oxygen species as long as they are thermally-stable in vacuum. / Ph. D.