Nous avons étudié la dynamique de relaxation de spin des électrons dans des puits quantiques GaAs/AlGaAs élaborés sur substrat d'orientation <111> par spectroscopie de photoluminescence résolue en temps. En appliquant un champ électrique externe d'environ 50 kV/cm le long de l'axe de croissance, nous avons observé une augmentation spectaculaire du temps de relaxation de spin de l'électron qui peut atteindre des valeurs plus grandes que 30 ns. Ceci est le résultat du contrôle électrique du décalage en énergie "spin-orbite" de la bande de conduction qui peut s'annuler lorsque le terme de Rashba compense exactement celui de Dresselhaus. Ceci entraîne une suppression du mécanisme de relaxation de spin de type D'Yakonov-Perel, mécanisme dominant dans les puits quantiques non dopés à des températures supérieures à 50 K. Les mesures effectuées sous champ magnétique externe transverse (configuration Voigt) montrent que les temps de relaxation de spin pour les trois directions de l'espace peuvent être contrôlés simultanément par le champ électrique. Ce contrôle "total" de la relaxation de spin ne peut se produire que pour des puits quantiques élaborés sur une orientation <111>. Nous avons finalement développé un modèle permettant d'interpréter les mesures expérimentales de la dépendance en champ électrique de l'anisotropie de la relaxation de spin dans ces puits quantiques <111>. / We have studied the electron spin dynamics in <111>-oriented GaAs/AlGaAs quantum wells grown on <111>-substrate by time-resolved photoluminescence spectroscopy. By applying an external electric field about 50 kV/cm along growth direction, we observed the spectacular increase of electron spin which can attain values greater than 30 ns. This phenomenon comes from the electrical control of spin-orbit interaction in conduction band that make the Rashba term compensate exactly with the Dresselhaus term. The cancellation effect of these two terms results in the suppression of electron spin relaxation induced by D'yakonov-Perelmechanism which is dominant in undoped quantum wells and at the temperatures greater than 50K. The measurement under an external transverse magnetic field (Voigt configuration) demonstrates that the spin relaxation times in three spatial directions are also controlled simultaneously by electric field. The "total" control of electron spin relaxation can only be observed in <111>-oriented quantum wells. Finally, we also develop the model to interpret the experimental measurement of spin relaxation anisotropy depending on electric field in <111>-oriented quantum wells.

The growing fields of spintronics and nanotechnology have created increased interest in developing the means to manipulate the spin of electrons. One such method arises from the combination of the spin-orbit interaction and the broken inversion symmetry that arises at surfaces and interfaces, and has prompted many recent investigations on metallic surfaces. A method by which surface states, in the absence of spin orbit effects, have been successfully investigated is the Green function embedding scheme of Inglesfield. This has been integrated into a self consistent FLAPW density functional framework based on the scalar relativistic K¨olling Harmon equation. Since the spin of the electron is a direct effect of special relativity, calculations involving the spin orbit interaction are best performed using solutions of the Dirac equation. This work describes the extension of Green’s function embedding to include the Dirac equation and how fully relativistic FLAPW surface electronic structure calculations are implemented. The general procedure used in performing a surface calculation in the scalar relativistic case is closely followed. A bulk transfer matrix is defined and used to generate the complex band structure and an embedding potential. This embedding potential is then used to produce a self consistent surface potential, leading to a Green’s function from which surface state dispersions and splittings are calculated. The bulk embedding potential can also be employed in defining channel functions and these provide a natural framework in which to explore transport properties. A relativistic version of a well known expression for the ballistic conductance across a device is derived in this context. Differences between the relativistic and nonrelativistic methods are discussed in detail. To test the validity of the scheme, a fully relativistic calculation of the extensively studied spin orbit split L-gap surface state on Au(111) is performed, which agrees well with experiment and previous calculations. Contributions to the splitting from different angular momentum channels are also provided. The main advantages of the relativistic embedding method are the full inclusion of the spin orbit interaction to all orders, the true semi infinite nature of the technique, allowing the full complex bands of the bulk crystal to be represented and the fact that a only small number of surface layers is needed in comparison to other existing methods.

Nitrenes are reactive intermediates for many organic reactions, such as Curtius rearrangement. The thermo- or photochemical- decomposition of azides or isocyanates was known to generate nitrenes. We investigated the thermal chemistry of nitrene adsorbed on Cu(111) using benzyl azide (Bz-N=N=N), benzyl isocyanate (Bz-N=C=O) and phenyl isocyanate (ph-N=C=O) as precursors under ultrahigh vacuum conditions using temperature-programmed reaction/desorption (TPR/D), reflectionabsorption infrared spectroscopy (RAIRS) and X-ray photoelectron spectroscopy (XPS). Our study shows that despite of the isoelectronic functionalities (-N=N=N vs. -N=C=O) these molecules undergo different reaction pathways. For benzyl azide (Bz-N=N=N), the azido group losses N2 ,and the phenyl group migrates from nitrogen to carbon, forming surface bound H2C=N-Ph at 210 K. Eventually, H2 elimination and a carbon-to-nitrogen phenyl shift give the thermally stable ph-CN final product. XPS reveals that benzyl isocyanate (Bz-N=C=O) rearranges to form amide intermediate on the surface, which breaks into CO2, HCN and toluene at 410 K. RAIRs suggests that phenyl isocyanate (ph-N=C=O) undergoes cyclodimerization, cyclotrimerization and condensation to remove CO2 at 170 K, and phenyl group shifts from nitrogen to carbon to produce a metal bound acyl nitrene species (Ph-(C=O)-N---Cu) at 410 K.

In our earlier research, we found the MOKE signal incident from 45 degree would cause perpendicular signal to couple with the longitudinal one. To distinguish the signal from one to another, we arranged 45 degree and 0 degree optical setup for the measurement of longitudinal and perpendicular respectively. However, hysteresis loops are only observed in the longitudinal direction. To measure smaller signal in our experiment, we reduced the noise level by one order and thus improved the signal to noise ratio. It¡¦s ten times better then previous result. Under smaller signal or worse S/N ratio condition, we still measured the MOKE signal. In order to quantitatively analyze the current-induced field, we made a metal coil and try to measure the bias produced by the field. And we injected the reverse current-induced field, , try to counteract the bias of hysteresis loop which is induced by sample current. Besides, we also dosed oxygen on the sample and measured the MOKE signal. There was no change comparing with which is measured in gas-free condition. But we found the coercivity became larger after annealing to 300K, and the current-induced bias in oxygen become smaller.

This dissertation employs the density functional theory (DFT) to investigate the oxidation of carbon monoxide (CO) on the W(111) surface and on the surface of Wn (n=10¡V15) nanoparticles. Since the properties of materials are significantly dependent on material size, we look into the influence of both the size and surface structure of tungsten catalysts on the CO oxidation process. The work contains two parts. Part 1: The adsorption and dissociation of O2 and CO on W(111) surface and Wn (n=10¡V15) nanoparticles. The chemical adsorption of O2 and CO on solid catalysts plays a very important role in heterogeneous catalysis for the CO oxidation reaction. The configurations, adsorption energies, vibration frequencies and electronic structures of adsorbates on W(111) and Wn (n=10¡V15) nanoparticles have been calculated to investigate their surface activity. The results indicate that adsorption of O2 and CO on Wn (n=10¡V15) nanoparticles are more stable compared to on the W(111) surface. The minimum energy pathways and transition states of chemical reaction processes on metal surfaces were also studied by the nudged elastic band (NEB) method. The dissociation barriers of O2 chemisorbed on Wn (n=10¡V15) nanoparticles are smaller those for the W(111) surface. Our results demonstrate that both the surface structure and size of metal significantly influence the adsorption and dissociation properties of adsorbates. Density functional theory-molecular dynamics (DFT-MD) simulation was also adapted to clarify the mechanism of O2 deposition on the W(111) surface. Observations of the variations of energy and bond lengths as a function of time show that the interaction between O2 and W atoms weakens the O¡VO bond, giving rise to the dissociation process. We conclude that the dissociation probability of an O2 molecule is affected by chemisorbed O2 coverage in the vicinity. Part 2: The mechanism of CO oxidation on W(111) and Wn nanoparticles. The oxidation of the CO molecule on transition metals usually follows two reaction pathways, either the Eley-Rideal (ER) mechanism or the Langmuir-Hinshelwood (LH) mechanism. In the ER mechanism, the CO molecule in the gas phase reacts directly with activated O2. The LH mechanism generally involves a few elementary steps, namely the co-adsorption of the O2 and CO molecules, O2 dissociation to form atomic oxygen, diffusion of atomic oxygen, and desorption of CO2. The oxidation of CO on a W10 nanoparticle surface and the W(111) surface are investigated by DFT calculations. Three pathways were studied in this dissertation: (i) CO + O2¡÷CO2 + O, (ii) CO + O2¡÷CO + O + O¡÷CO2 + O and (iii) CO + O¡÷CO2 via both LH and ER mechanisms. The calculated results show that CO oxidation on both the W10 nanoparticle and W(111) surfaces follow the ER rather than the LH mechanism. The CO oxidation on the W10 nanoparticle and W(111) surfaces occurs most easily via pathway (i) as compared to other two.

Nitromethane is the simplest organic-nitro compound as well as the archetype of an important class of high explosive. Homogeneous nitromethane reactions have been the subject of extensive studies. Particularly the unimolecular isomerization of nitromethane to methyl nitrite is proven to be competitive with simple C-N bond (bond energy 60 kcal/mol) rupture. The activation energy for the rearrangement was measured to be 55.5 kcal/mol and methyl nitrite has a very weak CH3O-NO bond energy 42 kcal/mol lower than that for homolysis. The thermal chemistry of nitromethane on Cu(111) was studied by a combination of temperature-programmed desorption (TPD) and reflection absorption infrared spectroscopy (RAIRS) techniques. TPD spectra show that the desorption features include the physisorbed multilayer and monolayer of CH3NO2 at 150 and 190 K, respectively. The major decomposition pathway is via cleavage of O-N bond to yield a major product NO, which is characterized by m/z 30(NO+). A possible contribution from isomerization of nitromethane to methyl nitrite (CH3NO2 CH3ONO) on the surface cannot be ruled out at 278 K. In addition to isomerization, the dehydrogenation products CO and CO2 are also unveiled as part of the desorption features at 314 and 455 K, respectively. We can further prove the reactivity of nitromethane on Cu(111) at 367 K by using the deuterated form of nitromethane which reveals the corresponding desorption TPR/D signals of D2, D2O and CD4. However, we find that nitromethane also reacts by dissociating the C-H bond and the O-N bond, however, leaving the C-N bond intact. Along this reaction channel, HCN desorbs as a product above 360 K, as evidenced by a broad desorption feature of m/z 27. Dimerization of CN to C2N2 occurs at 815 K. The RAIR spectroscopy demonstrates that nitromethane is indeed adsorbed on Cu(111) at 100 K. The formation of methoxy and formyl are supported by the observation of desorption of NO at 278 K with the characteristic NO stretching mode found at 1535 cm-1. Moreover, we assign side-bonded CN and aminomethylene (HC-NH2) present on Cu(111). After the surface is annealed to 330 K, a signature band at 2173 cm-1 is assigned to terminal-bounded CN stretching mode. This band eventually fades out above 900 K consistent with the evolution of cyanogen at 815 K.

In organometallic study, activation of C-F bond is an interesting subject, especially in fluoro-substituted propyl groups, because of their different reactivityn from fluoro-substituted methyl and ethyl groups. In this thesis, fluorinated propyl groups were studied on a Cu(111) surface under ultrahigh vacuum (UHV) conditions. We have examined the kinetics of the £]-elimination reaction in CF3CF2CH2-Cu, CHF2CF2CH2-Cu, and CF3CH2CH2-Cu. These species all decompose via £]-elimination to give CF3CF=CH2, CHF2CF=CH2, and CF3CH=CH2. The first two species undergo £]-fluoride elimination and the third one undergoes £]-hydride elimination. The difference in activation energies between the first two accounts for the charge separation (R-C£]+£_¡KF−£_¡KM+£_) in the transition state proposed by Gellman. The activation energies for £]-hydride elimination (CF3CH2CH2-Cu) and £]-fluoride elimination (CF3CF2CH2-Cu) was also compared. The activation energy for £]-fluoride elimination is found to be lower than that of £]-hydride elimination. In the studies of reaction pathways for perfluoropropyl groups (n-C3F7-Cu and i-C3F7-Cu) on Cu(111), we discovered novel chemistry in TPD. n-C3F7-Cu undergoes Cu-C homolytic cleavage (radical desorption) at 340 K, whereas i-C3F7-Cu eliminates the £]-fluorine at 365 K. By changing the Cu-C bond length in the i-C3F7-5Cu models their IR spectra was calculated. We discover that the IR of i-C3F7-5Cu with shorter Cu-C bond (1.728&#x00C5;) is more similar to the experimental IR spectra. That demonstrates the bond strength of Cu-C bond of i-C3F7-Cu is too strong to undergo Cu-C homolytic cleavage at 340 K. Hence, £]-F decomposition becomes the favorite pathway to i-C3F7-Cu because there are more £]-F atoms available in this moiety.

none

Two most useful families of copper CVD precursors that have been utilized widely are the Cu(I) and Cu(II) £]-diketone complexes. The Cu(II)precursors require the use of an external reducing agent such as hydrogen to deposit copper films, i.e. CuII(£]-diketonate)2 + H2 ¡÷ Cu0+2 £]-diketonate. The Cu(I) precursors deposit pure copper films without the use of an external agent via a disproportionation reaction that produces a Cu(II)£]-diketonate in conjunction with other organic byproducts, i.e. 2CuI(£]-diketonate)L ¡÷ Cu0+ CuII(£]-diketonate)2+2L where L is a typical Lewis base neutral ligand. However, Do those ligands resulting from the dissociation of the precursors simply desorb intact from the substrate or the growing films, or react further on the surface? To understand the surface chemistry of these ligands may provide better knowledge for designing more superior precursors and improvement of fabrication processes. Cu(hfac)(VTMS) and Cu(hfac)(MHY) are the most promising Cu(I) precursors, as shown in Scheme 1.1. Here we report studies on the chemistry of VTMS, MHY and hfacH on a Cu(111) surface. It should be noted that the hfacH is the simplest molecule containing the hfac, so we use it as a reference for £]-diketonate ligand. The Cu(111) single crystal was used to mimic the reactivity of these ligands on a growing Cu film during copper CVD. In situ analysis of ligand surface chemistry is carried out by TPD/R (temperature-programmed desorption/reaction) and RAIRS (reflection adsorption infrared spectroscopy) to elucidate plausible reaction mechanisms by which ligands decompose and eventually lead to impurity incorporation into the growing films, and to suggest means of minimizing such reactions.

We investigated the reactivity and bonding of perfluoroethyl groups (C2F5) on Cu(111) under ultra high vacuum conditions. Perfluoroethyl moieties bonded to the surface were generated by the dissociative adsorption of perfluoroethyl iodide. Temperature-programmed reaction/desorption (TPR/D) and reflection- adsorption infrared spectroscopy (RAIRS) revealed abounding reaction pathways, and a variety of intermediates were either identified or inferred. The major desorption products, hexafluoro-2-butyne and hexafluorocyclobutene, were detected at 360K and 440K, and some octafluorobutene was observed at 320K at higher coverages, implicating that two fluorine atoms were abstracted step-by-step from the C2F5 on Cu(111). Two sets of signature IR bands were recognized. One set (2054cm-1, 1409cm-1, 1210cm-1) was found to correlate with the surface-bound trifluorovinyl moieties which were also confirmed by directly generating this species from trifluorovinyl iodide. The other set of vibrational features (1322cm-1, 1224cm-1, 950cm-1) presumably implied the trifluoro- ethylidyne intermediate on the surface. Hence, C2F5(ad) underwent the £\-F and £]-F elimination reactions in sequence to yield trifluorovinyl which eventually led to hexafluoro-2-butyne. The alternative route was that C2F5(ad) proceeded via the £\-F elimination reaction twice to render trifluoroethylidyne which ultimately resulted in hexafluorocyclobutene. To our knowledge, the occurrence of the sequential £\-F and£]-F elimination pathway, or the double £\-F elimination reaction has never been observed in any single system.