In this thesis, the photophysics and spin chemistry of donor-photosensitizer-acceptor triads were investigated. While all investigated triads comprised a TAA as an electron donor and a NDI as an electron acceptor, the central photosensitizers (PS) were different chromophores based on the dipyrrin-motif. The purity and identity of all target compounds could be confirmed by NMR spectroscopy, mass spectrometry and elemental analysis. The first part of the work dealt with dipyrrinato-complexes of cyclometalated heavy transition metals. The successful synthesis of novel triads based on Ir(III), Pt(II) and Pd(II) was presented. The optical and electrochemical properties indicated charge separation (CS), which was confirmed by transient absorption (TA) spectroscopy. TA-spectroscopy also revealed that the process of CS is significantly slower and less efficient for the triads based on Pt(II) and Pd(II) than for the analogous Ir(III) triads. This is mostly due to a much more convoluted reaction pathway, comprising several intermediate states before the formation of the final charge separated state (CSS2). On the other hand, CSS2 exhibits long lifetimes which are dependent on the central metal ion. While the Ir(III) triads show lifetimes of about 0.5 µs in MeCN, the Pt(II) and Pd(II) analogues show lifetimes of 1.5 µs. The magnetic field effect on the charge recombination (CR) kinetics of CSS2 was investigated by magnetic field dependent ns-TA spectroscopy and could be rationalized based on a classical kinetic scheme comprising only one magnetic field dependent rate constant k±. The behavior of k± shows a clear separation of the coherent and incoherent spin interconversion mechanisms. While the coherent spin evolution is due to the isotropic hyperfine coupling with the magnetic nuclei of the radical centers, the incoherent spin relaxation is due to a rotational modulation of the anisotropic hyperfine coupling tensor and is strongly dependent on the viscosity of the solvent. This dependence could be used to measure the nanoviscosity of the oligomeric solvent pTHF, which was found to be distinctly different from its macroviscosity. The second part of the work dealt with bisdipyrrinato complexes and their bridged porphodimethenato (PDM) analogues. Initially, the suitability of the different chromophores for the use as PS in donor-acceptor substituted triads was tested by a systematic investigation of their steady state and transient properties. While the PDM-complex of Zn(II) and Pd(II) exhibited promising characteristics such as a high exited state lifetime and relatively intense emission, the purely organic parent PDM and the non-bridged bisdipyrrinato-Pd(II) complex were less suitable. The difference between the two Pd(II) complexes could be explained by a structural rearrangement of the non-bridged complex which results in a non-emissive metal centered triplet state with disphenoidal geometry. This rearrangement is prevented by the dimethylmethylene-bridges in the bridged analogue resulting in higher phosphorescence quantum yields and excited state lifetimes. With the exception of the Zn(II)PDM-complex, the synthesis of novel donor acceptor substituted triads could be realized for all desired central chromophores. They were investigated equivalently to the cyclometalated triads described in the first part. The steady state properties indicate a stronger electronic coupling between the subunits due to the lack of unsaturated bridges between the donor and the central chromophore. Photoinduced CS occurs in all investigated triads. Due to the low exited state lifetimes of the central chromophores, CSS is formed less efficiently for the triads based on the unbridged Pd(II)-complex as well as the purely organic PDM. In the triad based on the bridged Pd(II) complex, the CR of CSS2 is faster than its formation resulting in low intermediate concentrations. For its elongated analogue, this is not the case and CSS2 can be observed clearly. Although the spin-chemistry of the triads based on bisdipyrrinato-Pd(II) and porphodimethenato-Pd(II) is less well understood, first interpretations of the magnetic field dependent decay kinetics gave results approximately equivalent to those obtained for the cyclometalated triads. Furthermore, the MFE was shown to be useful for the investigation of the quantum yield of CS and the identity of the observed CSSs. In both parts of this work, the influence of the central photosensitizer on the photophysics and the spin chemistry of the triads could be shown. While the process of CS is directly dependent on the PS, the PS usually is not directly involved in the final CSSs. None the less, it can still indirectly affect the CR and spin chemistry of the CSS since it influences the electronic coupling between donor and acceptor, as well as the geometry of the triads.

In this work the synthesis of dendritic macromolecules and small redox cascades was reported and studies of their energy and electron transfer properties discussed. The chromophores in the dendrimers and the redox cascades are linked via triazoles, which were built up by CuAAC. Thereby, a synthetic concept based on building blocks was implemented, which allowed the exchange of all basic components. Resulting structures include dendrimers composed exclusively of TAAs (G1–G3), dendrimers with an incorporated spirobifluorene core (spiro-G1 and spiro-G2) and the donor-acceptor dendrimer D-A-G1, in which the terminal groups are exchanged by NDIs. Furthermore, a series of model compounds was synthesised in order to achieve a better understanding of the photophysical processes in the dendrimers. A modification of the synthetic concept for dendrimers enabled the synthesis of a series of donor-acceptor triads (T-Me, T-Cl and T-CN) consisting of two TAA donors and one NDI acceptor unit. The intermediate TAA chromophore ensured a downhill redox gradient from the NDI to the terminal TAA, which was proved by cyclic voltammetry measurements. The redox potential of the intermediate TAA was adjusted by different redox determining substituents in the “free” p-position of the TAA. Additionally, two dyads (Da and Db) were synthesised which differ in the junction of the triazole to the TAA or the NDI, respectively. In these cascades a nodal-plane along the N-N-axes in the NDI and a large twist angle between the NDI and the N-aryl substituent guaranteed a small electronic coupling. The photophysical investigations of the dendrimers focused on the homo-energy transfer properties in the TAA dendrimers G1–G3. Steady-state emission spectroscopy revealed that the emission takes place from a charge transfer state. The polar excited state resulted in a strong Stokes shift of the emission, which in turn led to a small spectral overlap integral between the absorption of the acceptor and the emission of the donor in the solvent relaxed state. According to the Förster theory, the overlap integral strongly determines the energy transfer rate. Fluorescence up-conversion measurements showed a strong and rapid initial fluorescence anisotropy decay and a much slower decrease on the longer time scale. The experiment revealed a fast energy transfer in the first 2 ps followed by a much slower energy hopping. Time resolved emission spectra (TRES) of the model compound M indicated a solvent relaxation on the same time scale as the fast energy transfer. The Förster estimation of energy transfer rates in G1 explains fast energy transfer in the vibrotionally relaxed state before solvent relaxation starts. Thereby, the emission spectrum of G1 in cyclohexane served as the time zero spectrum. Thus, solvent relaxation and fast energy transfer compete in the first two ps after excitation and it is crucial to discriminate between energy transfer in the Franck-Condon and in the solvent relaxed state. Furthermore, this finding demonstrates that fast energy transfer occurs even in charge transfer systems where a large Stokes shift prevents an effective spectral overlap integral if there is a sufficient overlap integral in before solvent relaxation. Energy transfer upon excitation was also observed in the spiro dendrimers spiro-G1 and spiro-G2 and identified by steady-state emission anisotropy measurements. It was assumed that the energy in spiro-G1 is completely distributed over the entire molecule while the energy in spiro-G2 is probably distributed over only one individual branch. This finding was based on a more polarised emission of spiro-G2 compared to spiro-G1. This issue has to be ascertained by e.g. time resolved emission anisotropy measurements in further energy transfer studies. Concerning the electron transfer properties of TAA-triazole systems the radical cations of G1–G2, spiro-G1 and spiro-G2 and of the model compound M were investigated by steady-state absorption spectroscopy. Experiments showed that the triazole bridge exhibits small electronic communication between the adjacent chromophores but still possesses sufficient electronic coupling to allow an effective electron transfer from one chromophore to the other. Due to the high density of chromophores, their D-A-D structure and their superficial centrosymmetry, the presented dendrimers are prospective candidates for two-photon absorption applications. The dyads, triads and the donor-acceptor dendrimer D-A-G1 were investigated regarding their photoinduced electron transfer properties and the effects that dominate charge separation and charge recombination in these systems. The steady-state absorption spectra of all cascades elucidated a superposition of the absorption characteristics of the individual subunits and spectra indicated that the chromophores do not interact in the electronic ground state. Time resolved transient absorption spectroscopy of the cascades was performed in the fs- and ns-time regime in MeCN and toluene as solvent. Measurements revealed that upon with 28200 cm-1 (355) nm and 26300 cm-1 (380 nm), respectively, an electron is transferred from the TAA towards the NDI unit yielding a CS state. In the triads at first a CS1 state is populated, in which the NDI is reduced and the intermediate TAA1 is oxidised. Subsequently, an additional electron transfer from the terminal TAA2 to TAA1 led to the fully CS2 state. Fully CS states of the dyads and triads exhibit lifetimes in the ns-time regime. In contrast for Db in MeCN, a lifetime of 43 ps was observed for the CS state together with the population of a 3NDI state. The signals of the other CS states decay biexponentially, which is a result of the presence of the 1CS and the 3CS states. While magnetic field dependent measurements of Db did not show an effect due to the large singlet-triplet splitting, T-CN exhibited a strong magnetic field dependence which is an evidence for the 1CS/3CS assignment. Further analysis of the singlet-triplet dynamics are required and are currently in progress. Charge recombination occurred in the Marcus inverted region for compounds solved in toluene and in the Marcus normal region for MeCN as solvent. However, a significant inverted region effect was observed only for Db. Triads are probably characterised by charge recombination rates in the inverted and in the normal region near to the vertex of the Marcus parabola. Hence the inverted region effect is not pronounced and the rate charge recombination rates are all in the same magnitude. However, compared to the charge recombination rate of Db the enlarged spatial distance between the terminal TAA and the NDI in the fully CS2 states in the triads resulted in reduced charge recombination rates by ca. one order of magnitude. More important than a small charge recombination rate is an overall lifetime of the CS states and this lifetime can significantly be enhanced by the population of the 3CS state. The reported results reveal that a larger singlet-triplet splitting in the dyads led to a CS state lifetime in the us time regime while a lifetime in the ns-time regime was observed in cases of the triads. Moreover, the singlet-triplet splitting was found to be solvent dependent in the triads, which is a promising starting point for further investigations concerning singlet-triplet splitting. The donor-acceptor dendrimer D-A-G1 showed similar characteristics to the dyads. The generation of a CS state is assumed due to a clear NDI radical anion band in the transient absorption spectrum. Noteworthy, the typical transient absorption band of the TAA radical cation is absent for D A-G1 in toluene. Bixon-Jortner analysis yielded a similar electronic coupling in D-A-G1 compared to the dyads. However, the charge recombination rate is smaller than of Db due to a more energetic CS state, which in the inverted region slows down charge recombination. In combination a singlet-triplet splitting similar to the dyads prolongs the CS state lifetime up to 14 us in diluted solution. Both effects result in an even better performance of D-A-G1 concerning energy conversion. D A-G1 is therefore a promising key structure for further studies on light harvesting applications. In a prospective study a second generation donor-acceptor dendrimer D-A-G2 might be an attractive structure accessible by “click reaction” of 13 and 8. D-A-G2 is expected to exhibit a downhill oriented gradient of CS states as assumed from the CV studies on G1–G3.

Im Rahmen der vorliegenden Arbeit wurde durch einzelmolekülspektroskopischer bzw. -mikroskopischer Methoden in Kombination mit einer mikrofluischen Zel- le unter Potenzialkontrolle die Elektrochemie von einzelnen einwandigen (6,5)- Kohlenstoffnanoröhren untersucht. Hierfür wurde ein Nahinfrarot-Photolumineszenz- Mikroskop aufgebaut und eine speziell an die experimentellen Vorgaben angepasste elektrochemische Zelle entwickelt, insofern als drei Elektroden (Arbeits-, Gegen- und Referenzelektrode) in einen mikrofluidischen Chip integriert wurden. Darüber hinaus war für die Durchführung der Experimente unter Wasser- und Sauerstoffaus- schluss die Konstruktion eines Handschuhkastens notwendig, sowie eine allgemeine Vorbehandlung der Elektrolytlösungen zur Entfernung gelöster Gase und Wasserreste. Ein weiteres Projekt umfasste den Aufbau einer chemischen Gasphasenabschei- dungsapparatur zur Synthese von Kohlenstoffnanoröhren. Die hierbei durchgeführten Experimente erbrachten Klarheit über den Einfluss der Prozessparameter Druck, Temperatur und Durchflussrate an Edukten. Aus den PL-Intensitätsänderungen bei Potenzialvariation konnten Reduktions- und Oxidationspotenziale (ERed = 0.15 V; EOx = 1.34 V) einzelner (6,5)-SWNTs gegen- über einer Platin Referenzelektrode und einem daraus resultierenden Redoxpotenzial von ∆ERedOx = 1.19 V ermittelt werden. Durch diese einzelmolekülspektroskopische Methode konnte zum einen gewährleistet werden, dass nur dieser spezielle Chira- litätstyp untersucht wurde und zum anderen eine Verfälschung der Resultate durch einen Potenzialabfall wie er typischerweise in CNT-Filmen auftritt aussgeschlossen werden. Eine Kombination der PL-Daten mit der Ramanintensitätsabhängigkeit des (6,5)-SWNT-S2-Übergangs bei Potenzialvariation erlaubte eine genauere Analyse des Löschmechanismus der PL von Kohlenstoffnanoröhren. Mithilfe eines von Her- tel et al. entwickelten diffusionslimitierten Stoßdesaktivierungsmodells konnte eine invers-quadratische Proportionalität zwischen der (6,5)-SWNT-Emission und den spannungsinduzierten Ladungsträgern ausgemacht werden. Auf Grundlage dieses Ergebnisses folgt, dass die über Photolumineszenzänderungen ermittelten Reduktions-und Oxidationswerte nicht mit den Bandkanten der CNTs übereinstimmen müssen, und dass für deren Bestimmung vielmehr auf Raman- bzw. Absorptionsspektroskopi- sche Techniken zurückgegriffen werden muss. Die einzelmolekülspektroskopische Herangehensweise ermöglichte ferner eine statis- tische Analyse der Verteilung der Reduktions- und Oxidationspotenziale im Vergleich zu den jeweiligen Erwartungswerten. Hierdurch konnte eine Einteilung der Modifika- tionseinflüsse auf das SWNT-Redoxverhalten in zwei Grenzfälle erfolgen. Es wurde angenommen, dass diese als “Dispergiermitteleffekte” und “CNT-Strukturdefekte” be- zeichneten Auswirkungen entweder das Resultat einer heterodispersen Verteilung an DOC auf der CNT-Oberfläche oder eine Folge von Defekten in der CNT-Gitterstruktur waren. In diesem Zusammenhang ergab sich aus der interpartikulären Analyse der Reduktions- und Oxidationswerte eine Korrelation, die einem dominierenden Einfluss der “CNT-Strukturdefekte” zugeordnet werden konnte. Dieser Beobachtung entgegen- gesetzt konnten aber auch über Untersuchungen der Redoxpotenziale innerhalb einer (6,5)-SWNT lokale Bereiche ausgemacht werden, die eine signifikante Abhängigkeit von “Dispergiermitteleffekte” aufwiesen. Abgesehen von diesen Einflüssen auf den Emissionsverlauf wurde auch eine Be- trachtung der Breite des spannungsgesteuerten Emissionsabfall durchgeführt. Da- raus konnte ermittelt werden, dass diese Ausdehnung eine Konsequenz aus der PL- Löschungseffizienz der Ladungsträger ist und, dass bei einer Verteilung von 0.32 Löschzentren pro Nanometer eine vollständige Abnahme der Photolumineszenzinten- sität induziert wird. Darüber hinaus wurde im Rahmen dieser Arbeit das redoxchemische Verhalten in- dividueller (6,5)-SWNTs in Wechselwirkung mit Ferrocenmolekülen untersucht. Die erhaltenen Ergebnisse ließen annehmen, dass die sich ausbildende Verbindung nicht-kovalenter Natur ist. Zwei verschiedene Gründe führten zu dieser Erkennt- nis: einerseits ließen sich die Ferrocenmoleküle von der CNT-Oberfläche durch ein Durchspülen des mikrofluidischen Kanals mit einer reinen DMF-Lösung entfernen und andererseits war keine dauerhafte Emissionsminderung durch die Ausbildung kovalenter Bindungen zu beobachten. Aus der potenzialabhängigen PL wurde zudem ein Elektronentransfer der Ferrocenmoleküle in die optisch generierten Löcher des CNT-Valenzbandes festgestellt und über eine anregungsintensitätsabhängige Messung die Zunahme dieses Ladungstransfers bei steigendem Photonenfluss nachgewiesen. Hinsichtlich der Anwendung von Kohlenstoffnanoröhren zur Elektrolyse bzw. Photo- lyse von Wasser wurde auch die Redoxchemie von (6,5)-SWNTs in diesem Solvens untersucht. Bezüglich der Emissionsintensität konnte gezeigt werden, dass diese im Vergleich zu organischen Lösungsmitteln reduziert vorliegt. Außerdem wurde eine irreversible Reaktion nach anodischer Polarisation über eine dauerhafte Löschung der PL beobachtet. Die Bestimmung der hierfür notwendigen Reaktionsumstände erbrachte, dass Wasser, Exzitonen (erzeugt durch optische Anregung) und spannungs- induzierte Löcher im Valenzband zur Bildung einer [SWNT(Q)]-Spezies führen, welche die irreversible Minderung der CNT-Emission verursacht. Darüber hinaus konnte die Reaktionsgeschwindigkeit über eine Kinetik pseudo-nullter-Ordnung be- schrieben werden, unter der Voraussetzung, dass die soeben genannten Parameter konstant verblieben. Desweiteren zeigte sich in einer ferrocenhaltigen Lösung, dass der Löscheffekt der [SWNT(Q)]-Spezies im anodischen Potenzialbereich teilweise reduziert wird. Es wurde angenommen, dass diese Beobachtung auf eine Oxidation der Löschzentren durch die Fc+-Kationen gründet. Mit Hilfe der CVD-Apparatur gelang es Kohlenstoffnanoröhren zu synthetisieren, wobei Ethanol als Kohlenstoffquelle und ein Eisen-Kobalt-Zeolith-Gemenge als Ka- talysator diente. Die Analyse der verschiedenen Prozessparameter zeigte, dass bei T = 750 °C das beste Verteilungsverhältnis zwischen den gewünschten (6,5)-SWNTs und anderen CNT-Chiralitäten bzw. dem amorphen Kohlenstoff vorliegt. Hierfür war, dass bei T < 750 °C die Verbrennung unerwünschter amorpher Kohlenstoffreste nur geringfügig stattfindet, und dass bei T > 750 °C die Bildung anderer Chiralitäten mit größerem Durchmesser als die (6,5)-SWNT bevorzugt wurde. Die Variation der Durchflussrate hingegen wirkte sich nur in einer absoluten Zunahme aller Chirali- täten aus. Die Steigerung des (6,5)-SWNT-Anteils für höhere Durchflüsse gelang trotzdem durch die geschickte Auswahl geeigneter Druck- und Temperaturwerte. Die Experimente zur Untersuchung der Druckabhängigkeit wiesen auf eine Relation mit dem Gesetz von Le Chatelier hin, insofern als bei einer Druckverringerung eine Verschiebung der Ethanol-Crackreaktion auf Produktseite stattfand. In diesem Zusam- menhang wurde angenommen, dass die damit verstärkt gebildeten Moleküle Ethan, Ethen und Methan den CNT-Anteil zwar erhöhen, jedoch auch eine Steigerung der amorphen Kohlenstoffkonzentration verursachen. Dementsprechend ergab ein Druck von p = 9 mbar das beste (6,5)-SWNT zu dem amorphen Kohlenstoffverhältnis. Anhand der Arbeiten in dieser Dissertation sind neue Erkenntnisse zwischen der PL-Sensitivität von (6,5)-SWNTs und deren Ladungszustand erhalten worden. Insbe- sondere die genaue Bestimmung der Korrelation zwischen der Photolumineszenz und den induzierten Ladungsträgern ermöglicht einen gezielteren Einsatz von Kohlenstoff- nanoröhren – so zum Beispiel im Bereich der Sensorik. In diesem Zusammenhang zeigen auch die interpartikulären Analysen der Redoxpotenzialverteilung die genau- en Auswirkungen vom Lösungsmittel und der Defektdichte auf die elektronische Struktur der CNTs auf. Darüber hinaus kann aus der Ursachenbestimmung für die Varianz der literaturbekannten Reduktions- bzw. Oxidationspotenziale fortan die ge- eignete spektroskopische Methode zur Evaluierung der Position von Leitungs- und Valenzband in Kohlenstoffnanoröhren besser eingegrenzt werden. Die spektroelektro- chemischen Analysen von (6,5)-SWNTs im Lösungsmittel Wasser und speziell die Bestimmung der Kinetik für die auftretende Reaktion liefern einen tieferen Einblick in die Wechselwirkung (6,5)-SWNT-H2O. Diese Ergebnisse sind insbesondere bei der Verwendung von Kohlenstoffnanoröhren als Elektrodenmaterial für die photolytische bzw. elektrolytische Spaltung von Wasser in Wasserstoff und Sauerstoff von Bedeu- tung. Neben der Untersuchung der SWNT-Wasser Interaktion unter andoischer und optischer Anregung, die zu einer kovalenten Bindung führte, wurde mit Hilfe der (6,5)- SWNT-Ferrocen Wechselwirkung ein Beispiel für eine nichtkovalente Redoxreaktion dargestellt, womit ein Vergleich dieser beiden Spezies und ihrer unterschiedlichen Auswirkungen auf die elektronische Struktur aufgezeigt werden konnte.

Within this study, the influence of the energetics of the bridge unit on electron transfer (ET) in an electrode-bridge-donor system was investigated in a monolayer environment. This was realized by specifically designing molecules containing ferrocene carboxylic ester donors and hydroquinone derivatives as bridge units and by using a gold electrode as acceptor. The energetics of the hydroquinone derivatives was adjusted by synthetically varying its substituents with the intention of changing the ET speed and mechanisms. Thereby the choice of the substituents was based on the literature known half-wave potentials of similar solvated hydroquinone derivatives and successively confirming them by conducting cyclic voltammetry on the actual bridge units synthesized. Then, a synthetic pathway, which accommodated the limited stability of the integrated terminal ferrocene carbon acid ester, was developed and successfully employed. This was followed by developing a procedure for preparing very dense and highly ordered monolayers from the target molecules on self-made gold microelectrodes. For the electrochemical investigations, several electrolyte solutions were tested until one, which ensured low susceptibility of the characterization setup towards slight changes of the electrode arrangement and measurement parameters while ensuring sufficient stability of the monolayers, was found. Furthermore, a new, commercially available potentiostat was established for the impedance measurements, which reduced the stress on the monolayers during the electrochemical characterizations in comparison to the equipment used in many former studies. Regarding the determination of the ET rates, the data analysis protocol for the impedance measurements developed by Creager et al. was slightly adapted to allow analysis of the investigated monolayers despite their non-ideal behavior. In addition, the influence of changes to the electrical parameters of the impedance scans was investigated to minimize the error in the acquired data. The electrochemical analysis of the monolayers by conducting cyclic voltammetry on MA, MB and MC prepared from A, B and C confirmed the accomplishment of near ideal surface coverage and exceptionally high order. The surface coverages of MB and MC were, probably due to the space filled by the substituents on their bridge units, slightly lower than those of MA. Furthermore, the shape of the redox waves of the ferrocene carboxylic acid redox center in the voltammogram of MA showed a broadening and a shift towards higher potentials, which was assigned to electrostatic interference of oxidized terminal redox centers due to the especially dense packing. However, in the voltammogram of MB, no sharp redox waves of the bridge units, as predicted by the analysis of preliminary monolayers of the same type with low surface coverage, were present. This was attributed to the different and varying microenvironment of the bridge units deeply embedded within high-density monolayers. In detail, the different degree of shielding of each individual bridge unit from counter ions and solvent molecules probably resulted in the half wave potential being shifted to varying higher potentials, thus preventing the formation of sharp redox waves. In addition, electrostatic effects of oxidized bridge units could have enhanced this effect. This leads to the conclusion that the half-wave potentials of fully solvated bridge units determined by the cyclic voltammetry are not suited to predict the energetics of the oxidized bridge states embedded within the prepared high density monolayers. Finally, the monolayers were successfully analyzed by impedance spectroscopy, which showed that the ET rate of MA is slightly higher than that of MB, and both are higher than that of MC. All of the values were, according to literature, in the expected region considering the length and degree of conjugation of the backbone. However, this picture is relativized when considering the targeted energetic alignment of the bridge units. According to the predicted very small energy gap between the oxidized states of the donor and the bridge unit in MB, a domination of the hopping mechanism should have led to a several orders of magnitude higher ET rate than in MA and MC. That this was not the case was attributed to the underestimation of the energy of the oxidized bridge states by utilizing cyclic voltammetry of the fully solvated bridge units (see above). According to the small differences of the ET rates the superexchange process was assumed to be the dominating mechanism not only in MA and MC but also in MB. However, even when shifted, the predicted energetic order of the oxidized bridge states should have led to a moderately decreasing ET rate from MB over MA to MC. The reason for the actual ET rate in MA being slightly higher than in MB might be found in the electrostatic interference of the terminal redox centers in MA (see above). In conclusion, the targeted model systems were prepared and the ET rates were successfully determined. However, the problems concerning the relative energetic positioning of the involved states within the dense monolayers prevented the specific alteration of the speed and mechanism of the ET. The reason for this can be probably found in the high density and order of the monolayers prepared within this work, which hamper the intrusion of the components of the electrolyte solutions. This various degree of stabilization for the individual bridge units by counter ions and solvent molecules leads to the energy of the oxidized bridge states being splitted and shifted towards higher potentials with respect to fully solvated bridge units. This effect might be further enhanced by electrostatics of neighboring already oxidized bridge states. All this makes the predetermination of the energetics of the embedded bridge units extremely difficult. On one hand, this behavior can be considered an obstacle and could probably be circumvented by designing molecules with bulky anchor groups and rigid molecular backbones, which would ensure perpendicular arrangement to the surface and full exposure of the bridge and terminal redox centers to the solvent molecules and counter ions. On the other hand, monolayers which completely embed integral redox centers might open up the opportunity to study the effects of microenvironments similar to those in solid state materials. Regarding mixed valence compounds, the present study focuses on bistriarylamine radical cation F∙+, which contains the [3.3]paracyclophane bridge unit. The results were compared to the, except for the bridge units, identical literature known compounds G∙+ and N∙+ with [2.2]paracyclophane and p-xylene bridges respectively. This led to the conclusion that slightly different bridge units can induce substantial changes to the internal reorganization energy. This is especially noteworthy since it is usually believed that structural adaption limited to the redox centers taking part in the charge transfer dominates the internal reorganization energy. Furthermore, the application of the two-state Mulliken-Hush approach shows that compounds F∙+ and G∙+ have near identical couplings and similar thermal barriers. Confirmation of the latter finding as well as near identical thermal electron transfer rates for both compounds were provided via a cooperation project by Grampp et al. in which these values were directly extracted from temperature dependent electron paramagnetic resonance measurements. These results are quite unexpected since the “through-space” distances of the stacked pi-systems in the paracyclophane bridges differ significantly. They are well within the sum of the van der Waals radii in G∙+ and barely within them in compound F∙+. In addition, these findings weaken the common assumption of the ethylene bridges in G∙+ substantially adding to the electronic coupling, since then, in F∙+, due to its propylene linkers, the coupling should be substantially reduced. Finally, relying on the fact that the electronic couplings are only three times higher and the thermal electron transfer rates are only one order of magnitude higher for N∙+ than for compounds F∙+ and G∙+ shows that intermolecular electron transfer in solid state materials can remain efficient, if the interacting pi-systems stay within the sum of van der Waals radii of their carbons. Concerning the donor-acceptor dyads, the current investigation centers on triarylamine-cyclophane-naphtalene diimide (TAA-CP-NDI) compounds which display almost complete photoinduced charge separation. Furthermore, their singlet charge separated states show lifetimes of hundreds of nanoseconds, which is rarely found in such simple dyads. In the present case they can be attributed to the particular amount of electronic coupling V (on the order of 100 cm^–1), which is brought about by incorporation of the smallest model systems for pi-stacks, the CPs, together with the nodes on the NDI lowest unoccupied molecular orbital, which electronically decouples the central NDI from its nitrogen substituents. In agreement with studies of [2.2]- and [3.3]paracyclophane bridged mixed valence compounds (see above), the cycolphane bridged dyads show very similar electronic coupling when dealing with ground state processes like charge recombination. However, when investigating excited state processes, like charge separation in the TAA-CP-NDI dyads, one has to bear in mind that the CP orbitals are involved in the formation of intermediate states that likely possess charge transfer character. In this case, the [2.2]paracyclophane bridge obviously induces a stronger coupling than the [3.3]paracyclophane. Another interesting property of the dyads studied here is the substantial population of the triplet charge separated (CS) state of ca. one third regarding both CS states, which is brought about by singlet-triplet interconversion from the singlet CS state. Thus, the triplet CS state with a lifetime of several microseconds acts as a kind of buffer for the CS state before recombining to the ground state and, thus, leads to distinctly prolonged overall lifetimes of the charge separated states. Thus it can be concluded that the intersystem crossing and charge recombination (CR) processes of the CS states are governed by a delicate balance of a large electronic coupling V and a large exchange interaction 2J (both with regard to systems containing a through-space pathway). The latter appears to be induced by second order interaction with a local triplet state lying close in energy to the CS state. This balance results in slow CR- and singlet-triplet- interconversion rates, which differ only by one order of magnitude. Compared to the many NDI containing dyads studied so far, these features of the dyads studied here are, to the best of our knowledge, unique. Especially the combination of high quantum yield of charge separation, long lifetimes and high energy of the charge separated state make the investigated systems interesting for practical applications. Furthermore, the presented unraveling of the underlying mechanisms is of substantial value for the future design of dyads for practical applications regarding the implementation and adjustment of these favorable properties.

In the first part of this thesis, the synthesis of a series of bistriarylamine (bisTAA) compounds was presented. On the one hand, the substitution pattern of the TAA at the benzene bridging unit was varied from meta- to para-position (pX and mX), on the other hand, the energetic position of the bridging unit was tuned by use of two electron-donating or electron-accepting substituents X (with X = OMe, Me, Cl, CN, NO2) in 2,5-position. In case of the meta-series, compounds with X in 4,6-position were synthesized (mX46). The photophysical and electrochemical properties of the neutral compounds were investigated. The cationic mixed valence (MV) bisTAA compounds could be generated by oxidation. Thermally induced hole transfer (HT) in the groud state was investigated by temperature depending ESR spectroscopy. While the HT rate k and HT barrier ΔG in mX are unaffected by the substituents X, k and ΔG in the pX series increase simultaneously with increasing electron-donating strength of X. This, at first contradictory observation can be explained by an increasingly important solvent dynamic effect and an additional, effective barrier. The optically induced HT was examined by UV/Vis/NIR spectroscopy. The pX-series revealed an increase of the electronic coupling V, and correspondingly a decrease of ΔG, with an increase of the electron donating character of X. For mX, a spectroscopic determination of these parameters was not possible. mX46 showed an intermediate behavior, MV compounds with strong electron-donating X, obtained coupling of similar magnitude as pX, which could be explained by means of DFT calculations, with regard to the molecular orbitals. In the second part of this work, the synthesis of a series of dyads with triarylamine (TAA) as a donor and naphthalene diimide (NDI) as an acceptor was presented. Again, the substitution pattern of the redox centers at the benzene bridging unit was varied in the form of a meta- or para-position (pXNDI or mXNDI) and the energetic position of the bridging unit was varied by X (with X = OMe, Me, Cl, CN, NO2) attached in the 2,5-position. Additionally, compound mOMe46NDI with methoxy substitution in 4,6-position was synthesized. The photophysical and electrochemical properties of these compounds were investigated. The electron transfer (ET) processes of charge separation (CS) and charge recombination (CR) of these were investigated by means of transient absorption (TA) spectroscopy in toluene. This was not possible for the nitro-compounds p-/mNO2NDI, since they decomposed under irradiation. In addition to that, the CR of pXNDI was not detectable by ns-setup, which is why the focus was given to the mXNDI series (with X = OMe–CN).The CS was examined by fs-TA spectroscopy, where the formation of a CS state could be detected. The rise time of the CS states decreases with increasing electron-withdrawing substituents X. CR was examined with ns-TA spectroscopy and shows a biexponential decay behavior, which is caused by singlet-triplet equilibrium in the CS state. By applying an external magnetic field, the decay behavior was decisively changed and the singlet-triplet splitting could be determined. This finding could also be confirmed by simulating the decay curves. In both parts of this work, the decisive influence of the benzene bridging unit on the appearing ET processes became obvious. For the HT in the ground state of the MV compound, as well as for the ET in the exited states of the DA compounds, the highest transfer rates were found for the para-series pX and pXNDI, and much smaller rates for the meta-series mX and mXNDI. The meta46-compounds mX46 and mOMeNDI46 showed an intermediate behavior in both parts of this work.