Fakultät für Physik - Digitale Hochschulschriften der LMU - Teil 02/05

The scope of the work presented is the investigation of photochemical reactions by means of ultrafast spectroscopy. Naturally these reactions start off in an optically bright excited state. Femtosecond time-resolved fluorescence spectroscopy is thus the method of choice to track the spectral and temporal dynamics of these emissive states. Here, an ultrafast fluorescence spectrometer based on the optical Kerr-effect serves as the appropriate tool to pursue this task. Additional information on dark states and ground states is provided by Uv-Vis transient absorption experiments.

The first part of the thesis deals with a fundamental concept of mechanistic chemistry

– the pericyclic reactions. The spectroscopic consequences implied within this theoretical framework are investigated by means of emission and absorption spectroscopy. The molecular probe is an indolyl-substituted fulgimide which undergoes a light-induced cyclization or cycloreversion, respectively. Both reactions feature a bi-phasic emission decay (cyclization: 0.06 ps, 0.4 ps, cycloreversion: 0.09 ps, 2.4 ps) whereas the slower component goes along with the product formation. The large difference in the slower time constants as well as the spectral properties of the corresponding emission point to the existence of different excited state pathways for both reactions. These results challenge the basic one-dimensional reaction scheme commonly used to describe pericyclic reactions. Referring to theoretical investigations, a two-dimensional reactive space is proposed to hold responsible for the different

behaviour of the two isomers.

The second part of the studies focuses on the dynamics of a certain type of photolabile protecting groups. These molecules are intramolecularly sensitised by a triplet energy donor, namely thioxanthone, and feature an ortho-substituted nitroaromatic as the reactive core. Investigations on the closely related energy donor xanthone reveal that photo-excitation is followed by a rapid (~ 1 ps) equilibration between the emissive singlet and a triplet state. This equilibrium holds responsible for a delayed fluorescence with a lifetime of ~ 0.1 − 1 ns and is ”switched off” by an internal conversion within the triplet manifold. These results can be directly transferred to thioxanthone and the sensitised protecting groups. The energy transfer in the latter molecules features a fast component from the initially populated triplet state (~ 100 ps) and a further slower contribution from the relaxed triplet state. Finally, the photo-reactive ortho-nitrobenzaldehyde (o-NBA) is compared with its

non-reactive isomers m- and p-NBA as model systems to obtain information on the reactive core of the protecting groups. These first fluorescence experiments on monocyclic

nitrated aromatics feature bi-phasic emission decays in all three cases – each with time constants of <100 fs and ~ 1 ps. For o-NBA the slower process matches the formation time of a ketene intermediate, whereas for the other isomers intersystem crossing to a triplet state is observed. The sub 100 fs process common in all three molecules is associated with a prepending 1pipi* -> 1npi* relaxation.

Near the end of their evolution, moderate-mass stars (approximatively between 1 and 8 solar masses) go through the asymptotic giant branch (AGB) phase. The AGB and the transition to the ensuing post-AGB are very time (both computational and human) consuming to simulate. Hence, large grids of models originating from detailed

calculations are rare; only few stars are generally computed. Some stellar libraries have nonetheless been constituted in the past. These are very sought-after by population synthesis modelers, as they try to

reproduce observed properties of larger stellar systems by assembling together many individual models. However such AGB and post-AGB grids can rapidly become outdated because of constant input-physics ameliorations. So-called synthetic AGB models are a solution as they are computationally less time demanding: nonetheless their major drawback is that they remain approximations.

In this context, new detailed models for stars of 1 to 6 solar masses, and compositions with a total heavy element content that corresponds to values between a fortieth and a double of the solar one (for two different relative scalings), have been computed. They evolve through

all stellar phases from the main sequence to the white dwarf stage; here focus is on the AGB and post-AGB parts.

The parameter space coverage is comparable only with the latest synthetic grids, as none of the previous detailed stellar libraries extend over such wide mass/metallicity intervals. Furthermore, the most recent physics, relevant for the AGB and post-AGB, are included. These updates, at the level of sophistication brought by this work,

are present neither in synthetic nor in any detailed computations. Because they are of paramount importance, their inclusion is the major asset of this work. Specifically, three updates have been implemented:

a) new chemistry-dependent low-temperature opacities b) up-to-date chemistry-dependent mass-loss prescriptions c) convective overshooting. With overshooting, envelope mixing will be more efficient and dredge-up additional material from deeper layers. The stellar surface will thus be converted from an oxygen- to a carbon-rich one, and changes will directly affect the opacities and in turn the mass-loss rates. While ignored in previous models, these

transformations are now taken into account.

We show, how such amendments were a necessity by comparing the results with older grids, and how models are in agreement, without any prior calibration, with observations.

The new stellar library contains 100 AGB and 60 post-AGB sequences. The AGB tracks cover the entire parameter space. Post-AGB models are also available for the full metallicity range but, because of convergence issues, only for star between 1 and 2 solar masses.

To our knowledge this work provides the scientific community with the most up-to-date, self-consistent and extended grid, originating from detailed AGB and post-AGB models. It is expected to be a valuable input-tool for population synthesis work and aims to supersede the

more than 10-year old outdated tracks that are currently still being used

In this thesis, gravitational lensing in the concordance LambdaCDM cosmology is investigated by carrying out ray-tracing along past light cones through the Millennium Simulation, a very large N-body simulation of cosmic structure formation. The method used for tracing light rays substantially extends previous ray-tracing methods that are based on the Multiple-Lens-Plane approximation.

Strong lensing is investigated by shooting random light rays through the Millennium Simulation. The probability is evaluated that an image of a small distant light source will be highly magnified, will be highly elongated or will be one of a set of multiple images. It is found that these probabilities increase strongly with increasing source redshift. It is shown that strong-lensing events can almost always be traced to a single dominant lensing object, and the mass and redshift distribution of these primary lenses is studied. The observed lens-mass range extends to lower masses than those found in earlier studies using simulations with lower spatial and mass resolution. Furthermore, effects of additional material along the line-of-sight are investigated. Although strong-lensing lines-of-sight are indeed biased towards higher than average mean densities, this additional matter typically contributes only a few percent of the total surface density.

The influence of stellar mass in galaxies on strong lensing is investigated by comparing the results obtained for lensing by dark matter alone to those obtained by also including the luminous matter. The dark-matter component of the lensing matter is constructed directly from the dark-matter particle distribution of the Millennium Simulation, while the luminous component is inferred from semi-analytic galaxy-formation models implemented within the evolving dark-matter distribution of the simulation. It is found that the inclusion of the stellar mass strongly enhances the probability for strong lensing compared to a 'dark-matter only' universe. The identification of the lenses associated with strong-lensing events reveals that the stellar mass of galaxies (i) significantly enhances the strong-lensing cross-sections of group and cluster halos, and (ii) gives rise to strong lensing in smaller halos, which would not produce noticeable effects in the absence of the stars. Even if only image splittings >10 arcsec are considered, the luminous matter can still enhance the strong-lensing optical depths by up to a factor of two.

Finally, the potential capabilities of future radio telescopes for imaging the cosmic matter distribution are discussed. The Millennium Simulation is used to simulate large-area maps of the lensing convergence with the noise, resolution and redshift-weighting achievable with a variety of idealised surveys. It is shown that by observing lensing of 21-cm emission during reionization with a sufficiently large radio telescope, an image of the matter distribution could be obtained whose signal-to-noise far exceeds that of any map made using galaxy lensing. These mass images would allow the dark-matter halos of individual galaxies to be viewed directly, giving a wealth of statistical and morphological information about the relative distributions of mass and light. For telescopes like the planned Square Kilometre Array, mass imaging may be possible near the resolution limit of the core array of the telescope.