Refine
Has Fulltext
- yes (11)
Is part of the Bibliography
- yes (11)
Year of publication
Document Type
- Doctoral Thesis (11)
Language
- English (11) (remove)
Keywords
- Dotierung (3)
- Einwandige Kohlenstoff-Nanoröhre (3)
- Kohlenstoff-Nanoröhre (3)
- Spektroskopie (2)
- Adsorption (1)
- Carbon Nanotube (1)
- Charge carrier characterization (1)
- Charges recombination (1)
- Chemisorption (1)
- Chiralität <Chemie> (1)
Investigation of Nanostructure-Induced Localized Light Phenomena Using Ultrafast Laser Spectroscopy
(2017)
In recent years, the interaction of light with subwavelength structures, i.e., structures that are smaller than the optical wavelength, became more and more interesting to scientific research, since it provides the opportunity to manipulate light-induced dynamics below the optical diffraction limit. Specifically designed nanomaterials can be utilized to tailor the temporal evolution of electromagnetic fields at the nanoscale. For the investigation of strongly localized processes, it is essential to resolve both their spatial and their temporal behavior. The aim of this thesis was to study and/or control the temporal evolution of three nanostructure-induced localized light phenomena by using ultrafast laser spectroscopy with high spatial resolution.
In Chapter 4, the absorption of near-infrared light in thin-film a-Si:H solar cells was investigated. Using nanotextured instead of smooth interfaces for such devices leads to an increase of absorption from < 20% to more than 50% in the near-infrared regime. Time-resolved experiments with femtosecond laser pulses were performed to clarify the reason for this enhancement. The coherent backscattered radiation from nanotextured solar cell devices was measured as a function of the sample position and evaluated via spectral interferometry. Spatially varying resonance peaks in the recorded spectra indicated the formation of localized photonic modes within the nanotextured absorber layers. In order to identify the modes separately from each other, coherent two-dimensional (2D) nanoscopy was utilized, providing a high spatial resolution < 40 nm. In a nanoscopy measurement on a modified device with an exposed nanotextured a-Si:H absorber layer, hot-spot electron emission was observed and confirmed the presence of localized modes. Fitting the local 2D nanospectra at the hot-spot positions enabled the determination of the resonance frequencies and coherence lifetimes of the modes. The obtained lifetime values varied between 50 fs and 130 fs. Using a thermionic emission model allowed the calculation of the locally absorbed energy density and, with this, an estimation of the localization length of the photonic modes (≈1 μm). The localization could be classified by means of the estimated localization length and additional data evaluation of the backscattered spectra as strong localization ─ the so-called Anderson localization.
Based on the experimental results, it was concluded that the enhanced absorption of near-infrared light in thin-film silicon solar cells with nanotextured interfaces is caused by the formation of strongly localized photonic modes within the disordered absorber layers. The incoming near-infrared light is trapped in these long-living modes until absorption occurs.
In Chapter 5, a novel hybridized plasmonic device was introduced and investigated in both theory and experiment. It consists of two widely separated whispering gallery mode (WGM) nanoantennas located in an elliptical plasmonic cavity. The goal was to realize a periodic long-range energy transfer between the nanoantennas. In finite-difference time-domain (FDTD) simulations, the device was first optimized with respect to strong coupling between the localized antenna modes and the spatially-extended cavity mode. The geometrical parameters of the antennas and the cavity were adjusted separately so that the m="0" antenna mode and the cavity mode were resonant at λ="800 nm" . A high spatial overlap of the modes was achieved by positioning the two antennas in the focal spots of the cavity, leading to a distance between the antenna centers of more than twice the resonant wavelength of the modes. The spectral response of the optimized device revealed an energy splitting of the antenna and the cavity mode into three separated hybridized eigenmodes within an energy range of about 90 meV due to strong coupling. It could be well reproduced by a simple model of three coupled Lorentzian oscillators. In the time domain, an oscillatory energy transfer between both antennas with a period of 86 fs and an energy transfer efficiency of about 7% was observed for single-pulse excitation. For the experiments, devices with cavities and antennas of varying size were fabricated by means of focused-ion-beam (FIB) milling. Time-resolved correlation measurements were performed with high spatial and temporal resolution by using sequences of two femtosecond laser pulses for excitation and photoemission electron microscopy (PEEM) for detection. Local correlation traces at antennas in resonant devices, i.e., devices with enhanced electron emission at both antenna positions, were investigated and reconstructed by means of the coupled-oscillator model. The corresponding spectral response revealed separated peaks, confirming the formation of hybridized eigenmodes due to strong coupling. In a subsequent simulation for single-pulse excitation, one back-and-forth energy transfer between both antennas with an energy transfer efficiency of about 10% was observed.
Based on the theoretical and experimental results, it was demonstrated that in the presented plasmonic device a periodic long-range energy transfer between the two nanoantennas is possible. Furthermore, the coupled-oscillator model enables one to study in depth how specific device properties impact the temporal electric-field dynamics within the device. This can be exploited to further optimize energy transfer efficiency of the device. Future applications are envisioned in ultrafast plasmonic nanocircuitry. Moreover, the presented device can be employed to realize efficient SPP-mediated strong coupling between widely separated quantum emitters.
In Chapter 6, it was investigated in theory how the local optical chirality enhancement in the near field of plasmonic nanostructures can be optimized by tuning the far-field polarization of the incident light. An analytic expression was derived that enables the calculation of the optimal far-field polarizations, i.e., the two far-field polarizations which lead to the highest positive and negative local optical chirality, for any given nanostructure geometry. The two optimal far-field polarizations depend on the local optical response of the respective nanostructure and thus are functions of both the frequency ω and the position r. Their ellipticities differ only in their sign, i.e., in their direction of rotation in the time domain, and the angle between their orientations, i.e., the angle between the principal axes of their ellipses, is ±π/"2" . The handedness of optimal local optical chirality can be switched by switching between the optimal far-field polarizations. In numerical simulations, it was exemplarily shown for two specific nanostructure assemblies that the optimal local optical chirality can significantly exceed the optical chirality values of circularly polarized light in free space ─ the highest possible values in free space. The corresponding optimal far-field polarizations were different from linear and circular and varied with frequency. Using femtosecond polarization pulse shaping provides the opportunity to coherently control local optical chirality over a continuous frequency range. Furthermore, symmetry properties of nanostructures can be exploited to determine which far-field polarization is optimal.
The theoretical findings can have impact on future experimental studies about local optical chirality enhancement. Tuning the far-field polarization of the incident light offers a promising tool to enhance chirally specific interactions of local electromagnetic fields with molecular and other quantum systems in the vicinity of plasmonic nanostructures. The presented approach can be utilized for applications in chiral sensing of adsorbed molecules, time-resolved chirality-sensitive spectroscopy, and chiral quantum control.
In conclusion, each of the localized light phenomena that were investigated in this thesis ─ the enhanced local absorption of near-infrared light due to the formation of localized photonic modes, the periodic long-range energy transfer between two nanoantennas within an elliptical plasmonic cavity, and the optimization of local optical chirality enhancement by tuning the far-field polarization of the incident light ─ can open up new perspectives for a variety of future applications.
.
This thesis will outline studies performed on the fluorescence dynamics of phenyl-benzo-
[c]-tetrazolo-cinnolium chloride (PTC) in alcoholic solutions with varying viscosity using
time-resolved fluoro-spectroscopic methods. Furthermore, the properties of femtosecond
Laguerre-Gaussian (LG) laser pulses will be investigated with respect to their temporal
and spatial features and an approach will be developed to measure and control the spatial
intensity distribution on the time scale of the pulse.
Tetrazolium salts are widely used in biological assays for their low oxidation and reduction
thresholds and spectroscopic properties. However, a neglected feature in these applications
is the advantage that detection of emitted light has over the determination of the
absorbance. To corroborate this, PTC as one of the few known fluorescent tetrazolium
salts was investigated with regard to its luminescent features. Steady-state spectroscopy
revealed how PTC can be formed by a photoreaction from 2,3,5-triphenyl-tetrazolium
chloride (TTC) and how the fluorescence quantum yield behaved in alcoholic solvents
with different viscosity. In the same array of solvents time correlated single photon counting
(TCSPC) measurements were performed and the fluorescence decay was investigated.
Global analysis of the results revealed different dynamics in the different solvents, but
although the main emission constant did change with the solvent, taking the fluorescence
quantum yield into consideration resulted in an independence of the radiative rate from
the solvent. The non-radiative rate, however, was highly solvent dependent and responsible
for the observed solvent-related changes in the fluorescence dynamics. Further studies
with the increased time resolution of femtosecond fluorescence upconversion revealed an
independence of the main emission constant from the excitation energy, however the dynamics
of the cooling processes prior to emission were prolonged for higher excitation
energy. This led to a conceivable photoreaction scheme with one emissive state with a
competing non-radiative relaxation channel, that may involve an intermediate state.
LG laser beams and their properties have seen a lot of scientific attention over the past two
decades. Also in the context of new techniques pushing the limit of technology further to
explore new phenomena, it is essential to understand the features of this beam class and
check the consistency of the findings with theoretical knowledge. The mode conversion
of a Hermite-Gaussian (HG) mode into a LG mode with the help of a spiral phase plate
(SPP) was investigated with respect to its space-time characteristics. It was found that
femtosecond LG and HG pulses of a given temporal duration share the same spectrum
and can be characterized using the same well-established methods. The mode conversion
proved to only produce the desired LG mode with its characteristic orbital angular momentum
(OAM), that is conserved after frequency doubling the pulse. Furthermore, it
was demonstrated that temporal shaping of the HG pulse does not alter the result of its
mode-conversion, as three completely different temporal pulse shapes produced the same
LG mode. Further attention was given to the sum frequency generation of fs LG beams
and dynamics of the interference of a HG and a LG pulse. It was found that if both are
chirped with inverse signs the spatial intensity distribution does rotate around the beam
axis on the time scale of the pulse. A strategy was found that would enable a measurement
of these dynamics by upconversion of the interference with a third gate pulse. The results
of which are discussed theoretically and an approach of an experimental realization had
been made. The simulated findings had only been reproduced to a limited extend due to
experimental limitations, especially the interferometric stability of the setup.
Doping plays a decisive role for the functionality of semiconductor-based (opto-)electronic
devices. Hence, the technological utilization of semiconductors necessitates control and a
fundamental understanding of the doping process. However, for low-dimensional systems like
carbon nanotubes, neither concentration nor distribution of charge carriers is currently well known.
The research presented in this thesis investigated the doping of semiconducting carbon nanotubes by spectroscopic methods. Samples of highly purified, intrinsic (6,5) single-wall carbon nanotubes were fabricated using polymer stabilization.
Chapter 4 showed that both electro- and redox chemical $p$-doping lead to identical bleaching,
blueshift, broadening and asymmetry of the S$_1$ exciton absorption band. The similar spectral changes induced by both doping schemes suggest that optical spectra can not be used to infer what process was used for doping. Perhaps more importantly, it also indicates that the distribution of charges and the character of the charge transfer states does not depend on the method by which doping was achieved.
The detailed analysis of the doping-induced spectral changes in chapter 5 suggests that surplus charges are distributed inhomogeneously. The hypothesis of carrier localization is consistent with the high sensitivity of the S$_1$ exciton photoluminescence to additional charge carriers and with the stretched-exponential decay of the exciton population following ultrafast excitation.
Both aspects are in good agreement with diffusion-limited contact quenching of excitons
at localized charges. Moreover, localized charges act – similar to structural defects – as
perturbations to the bandstructure as evidenced by a doping-induced increase of the D-band
antiresonance in the mid-infrared spectrum.
Quantum mechanical model calculations also suggest that counterions play a crucial role in
carrier localization. Counterion adsorption at the nanotube surface is thus believed to induce charge traps of more than 100 meV depth with a carrier localization length on the order of 3 - 4 nm. The doping-induced bleach of interband absorption is accompanied by an absorption increase in the IR region below 600 meV. The observed shift of the IR peak position indicates a continuous transition from localized to rather delocalized charge carriers. This transition is caused by the increase of the overlap of charge carrier wavefunctions at higher charge densities and was modeled by classical Monte-Carlo simulations of intraband absorption.
Chapter 6 discussed the spectroscopy of heavily (degenerately) doped nanotubes, which are
characterized by a Drude-response of free-carrier intraband absorption in the optical conductivity spectrum. In the NIR spectral region, the S$_1$ exciton and X$+^_1$ trion absorption is replaced by a nearly 1 eV broad and constant absorption signal, the so-called H-band. The linear and transient absorption spectra of heavily doped nanotubes suggest that the H-band can be attributed to free-carrier interband transitions.
Chapter 7 dealt with the quantification of charge carrier densities by linear absorption spectroscopy.
A particularly good measure of the carrier density is the S$_1$ exciton bleach. For a
bleach below about 50 %, the carrier density is proportional to the bleach. At higher doping
levels, deviations from the linear behavior were observed. For doping levels exceeding a
fully bleached S$_1$ band, the determination of the normalized oscillator strength f$\text{1st}$ over the
whole first subband region (trion, exciton, free e-h pairs) is recommended for quantification of carrier densities. Based on the nanotube density of states, the carrier density $n$ can be estimated using $n = 0.74\,\text{nm}^{−1} \cdot (1 − f_\text{1st})$.
In the last part of this thesis (chapter 8), the time-resolved spectroelectrochemistry was
extended to systems beyond photostable carbon nanotube films. The integration of a flowelectrolysis cell into the transient absorption spectrometer allows the investigation of in-situ electrochemically generated but photounstable molecules due to a continuous exchange of sample volume. First time-resolved experiments were successfully performed using the dye
methylene blue and its electrochemically reduced form leucomethylene blue.
In summary, we have prepared single-wall carbon nanotube (SWNT) thin films by the method of evaporation-induced self-assembly (EISA). Using the scalable two-plate or lens setups, sorts of different film types or patterns of SWNTs has been successfully fabricated directly from the evaporation of solvents and could be precisely controlled by the concentrations of SWNT in ambient conditions. The special geometry of meniscus as the capillary bridge has not only given rise to a much higher efficiency of fabrication than what previously reported but also allowed us to monitor the pinning and depinning process carefully and further investigate the mechanism underlying the formation of different film morphologies.
In contrast with the conventional "stick-slip" model, we have provided the new dynamical pinning and zipping model for the contact line (CL) behavior. By analyzing the motion of CL and varying deposited patterns, the traditionally so-called "stick" state should be treated as a dynamical pinning process due to the interfacial tension contrast between SWNT-covered and bare silicon surface. Besides, the plausible one-step "slip" motion could be dominated by the zipping-like kink propagation.
In addition, the experiments with heated substrates at higher temperatures between 30°C and 50 °C have shown that the striped pattern could be fabricated by both much lower SWNT and SDS concentrations than that in room temperature, which is consistent with our model of interfacial tension contrast. In this situation, the deposition rate was increased but the quality of SWNT alignment was undermined because the corresponding moving velocity of SWNT was also too fast for SWNTs to rotate when the evaporative rate was high.
The similar results were identified by the SWNT/polymer conjugates dispersed in chloroform under the similar setups and other identical conditions. The typical breathing motion of dynamical pinning and zipping-like propagation for depinning were confirmed by the new suspensions despite that some morphological parameters changed dramatically compared with that from the aqueous solution. For example, the spacing between stripes reached 100 µm ~ 200 µm because the large contact angle contrast between HDMS- and SWNT-covered surface accompanies with the high evaporation rate of chloroform in the pinning and depinning process. Likewise the average CL velocity for fabrication reached around 20 µm/s due to the much higher evaporation rate of chloroform than water.
Using alike suspensions, the modified EISA method called dose-controlled floating evaporative self-assembly (DFES) was employed to implement the self-assembly of SWNTs on the water/air interface and then deposit them on solid substrate by directed floating. Although the stripes were fabricated successfully by drops with certain doses and SWNT concentrations, there inevitably existed randomly oriented SWNTs from the water surface that built networks between the stripes containing well-aligned tubes. In order to slow down the evaporation rate and monitor the process detailedly, we used chlorobenzene as the solvent instead of chloroform and find the typical pinning/depinning movement of the CL. A preliminary analysis of the results in terms of chlorobenzene implied that the CL possibly followed the similar pinning/depinning process in consistence with our model with capillary bridge.
In the last part of the thesis, the primary research on the optical properties of these stripes of ultrahigh purity semiconducting nanotubes was conducted by fluorescence microscopy and photoluminescence excitation (PLE) spectroscopy. The energy transfer of the photogenerated excitons was confirmed between different tube species with controlled band gaps.
In short, the experiments performed in this thesis allowed to gain new insights about the fabrication of large-area SWNT thin films by the cost-effective solution-processed method and most importantly to uncover its intrinsic mechanism as well. Combined with the separation and selection technique like density gradient centrifugation or polyfluorene derivatives assisted method, highly monodisperse semiconducting nanotubes could be deposited into organized, controllable and functional arrays.
Beyond the ambient conditions, precise control for the evaporation under preset temperature and vapor pressure could possibly extend the technique to the industry level. Assisted by some other mature techniques such as roll-to-roll printing, the cost-effective method could be widely used in the manufacture of various thin film devices. More complex 2D or even 3D structures could be designed and accomplished by the method for the functional or stretchable requirements. Further research on the fundamental exciton transition and diffusion in different networks or structures of SWNTs will be the significant precondition for the real applications.
Looking ahead, from the individual carbon nanotube to its thin film, this promising material with outstanding properties had many challenges to overcome before the real-world applications. Thanks to the availability of pure and well-defined materials, the scalable solution-processed approaches for fabrication of thin films should be able to unlock the potential of carbon nanotubes and exploit them in (opto-)electronic devices in the foreseeing future.
In this work the successful synthesis, the linear and nonlinear spectroscopic properties as well as the electrochemical behaviour of some linear and star-shaped squaraine superchromophores that are based on indolenine derivatives were presented. The attempt to synthesise similar chromophores which contained only benzothiazole squaraines failed unfortunately. However, one trimer that contained mixed benzothiazole indolenine squaraines could be synthesised and investigated as well.
The linear spectroscopic properties, like red-shift and broadening of the absorption, of all superchromophores could be explained by exciton coupling theory. The heterochromophores (SQA)2(SQB)-N, (SQA)(SQB)2-N and (SQA)(SQB)-NH displayed additional to the typical squaraine fluorescence from the lowest excited state some properties that could be assigned to localised states. While the chromophores with N-core showed very small emission quantum yields, the chromophores with the other cores and the linear oligomers display an enhancement compared to the monomers.
Transient absorption spectroscopy experiments of the star-shaped superchromophores showed, that their formally degenerated S1 states are split due to a deviation of the ideal C3 symmetry. This is also the reason for the observation of an absorption band for the highest exciton state, which is derived from the S1-state of the monomers, as its transition-dipole moment would be zero in the symmetrical case.
The linear oligomers and the star-shaped superchromophores with a benzene or triarylamine core showed at least additive, sometimes even weak cooperative, behaviour in the two-photon absorption experiments. Additional to higher two-photon absorption cross sections the chromophores showed a pronounced broadening of the nonlinear absorption, due to symmetry breaking and a higher density of states.
Unfortunately it was not possible to solve the problem of the equilibrium of the cisoid and the transoid structure of donor substituted azulene squaraines, due to either instability of the squaraines or steric hindrance.
In the course of this dissertation, we have presented the interest of using spectroscopic methods to unravel the physics of polymer semiconductors in photovoltaic applications. Applying photoluminescence and photoinduced absorption spectroscopy to the reference system P3HT:PCBM has enabled us to study the major steps of photocurrent generation in organic bulk heterojunctions, from excitons generation to charges extraction and loss mechanisms and thus to improve the understanding of those mechanisms.
The exciton binding energy, is the first obstacle to overcome for photocurrent generation in organic solar cell and the reason for the use of two materials, whose heterojunction act as a driving force for charge separation. We developed an original photoluminescence-detected field-induced exciton quenching method to investigate this energy. Absorption and photoluminescence spectra of pure P3HT show that, while both amorphous and crystalline domains participate in
absorption, the energy is then transferred to the crystalline domains, from where the photoluminescence is exclusively originating. The field dependence of this photoluminescence showed that an energy of no less than 420 meV is necessary to split excitons into non photon-emitting species. Comparing those results with energy levels obtained by absorption and photoelectron spectroscopies, confirmed that the formation of those species is only a first step toward dissociation into free charges. Indeed, photoemission spectroscopy and the onset
of photocurrent upon increasing the photon energy in a pure P3HT solar cell, concomitantly show that the energy level of a pair of free polarons is located 0.7 eV above the one of the exciton. The comprehensive analysis of those results originating from those different method enable us to draw a global picture of the states and energies involved in free polarons generation in pure material. This work has been widely acknowledged by the scientific community, published in Physical Review B in 2010 [1] and presented in national [2] and international [3] conferences.
The spectroscopy of excited states is used to detect the presence of wanted species (charges) and potentially unwanted neutral species upon photoexcitation. As such, it offers us the possibility to qualify the efficiency of charge generation and, if any, identify the competing processes and the generation of unwanted species. In the frame of the European Marie Curie Research Network SolarNType,[4] this possibility was used - in combination with morphological,
charge transport and devices characterizationsn - to study a number of new donor:acceptor blends. Thanks to those techniques, we were able to not only quantify the potential of those blends, but also to provide the chemist laboratories with a precious and detailed feedback on the strengths and weakness of the molecules, regarding charge generation, transport and extraction. The detailed study of terrylene-3,4:11,12-bis(dicarboximide) as electron acceptor for
solar cells application was published in the peer review journal Synthetic Metals and was chosen to illustrate the cover page of the issue [5].
Finally, in the last chapter, we have used time resolved photoinduced absorption to improve the understanding of the charge carrier loss mechanisms in P3HT:PCBM active layers. This comprehension is of prime importance because, the fact that this recombination is far weaker than expected from the Langevin theory, enable polarons to travel further without recombining and thus to build thicker and more efficient devices. A comprehensive analysis of steady-state
PIA spectra of pure P3HT, indicates that probing at 980 nm at a temperature between 140 and 250 K enables to monitor specifically polaron densities in both neat P3HT and P3HT:PCBM. Applying this finding to transient absorption enabled us to monitor, for the first time, the bimolecular recombination in pure P3HT, and to discover that - in sharp contrast with the blend - this recombination was in agreement with the Langevin theory. Moreover, it enables us to pinpoint the important role played by the existence of two materials and of energetical traps in the slow recombination and high recombination orders observed in the blend. This work has been published in the Journal of Applied Physics.[6]
Those new insights in the photophysics of polymer:fullerene photoactive layers could have a strong impact on the future developement of those materials. Consistent measurements of the binding energy of excitons and intermediate species, would enable to clarify the role played by excess thermal energy in interfacial states dissociation. Better understanding of blends
morphology and its influence on solar cells parameters and in particular on recombination could enable to reproduce the conditions of limited recombination on material systems offering some promising performances but with only limited active layer thicknesses. However, due to the number of parameters involved, further experimentation is required, before we can reach a quantitative modeling of bimolecular recombination.
[1] Deibel et al., Phys. Rev. B, 81:085202, 2010
[2] Gorenflot et al., Deutsche Physikalische Gesellschaft Frühjahrstagung 2010, CPP20:10, Regensburg, Germany, 2010
[3] Gorenflot et al., International Conference of Synthetic Metals, 7Ax:05, Kyoto, Japan, 2010
[4] Marie-Curie RTN "SolarNTyp" Contract No. MRTN-CT-2006-035533
[5] Gorenflot et al., Synth. Met., 161(23{24):2669-2676, 2012
[6] Gorenflot et al., J. Appl. Phys., 115(14):144502, 2014
This thesis includes measurements that were recorded by cooperation partners. The EPR spec‐ trosa mentioned in section 5.2 were recorded by Michael Auth from the Dyakonov Group (Ex‐ perimental Physics VI, Julius‐Maximilians‐Universität, Würzburg). The TREFISH experiments and transient absorption in section 5.4 spectra were performed by Jašinskas et al. from the V. Gulbi‐ nas group (Center for Physical Sciences and Technology, Vilnius, Lithuania). This dissertation investigated the interactions of semiconducting single‐walled carbon nanotubes (SWNTs) of (6,5) chirality with their environment. Shear‐mixing provided high‐quality SWNT sus‐ pensions, which was complemented by various film preparation techniques. These techniques were in turn used to prepare heterostructures with MoS2 and hBN, which were examined with a newly constructed photoluminescence microscope specifically for this purpose. Finally, the change of spectral properties of SWNTs upon doping was investigated in more detail, as well as the behaviour of charge carriers in the tubes themselves. To optimise the SWNT sample preparation techniques that supplied the other experiments, the sample quality of shear‐mixed preparations was compared with that of sonicated samples. It was found that the quantum efficiency of sheared suspensions exceeds that of sonicated suspensions as soon as the sonication time exceeds 30 min. The higher PLQY is due to the lower defect concentration in shear‐mixed samples. Via transient absorption, a mean lifetime of 17.3 ps and a mean distance between defects of 192.1 nm could be determined. Furthermore, it was found that the increased efficiency of horn sonication is probably not only due to higher shear forces acting on the SWNT bundles but also that the shortening of PFO‐BPy strands plays a significant role. Sonication of very long polymer strands significantly increased their effectiveness in shear mixing. While previous approaches could only achieve very low concentrations of SWNTs in suspensions, pre‐sonicated polymer yielded results which were comparable with much shorter PFO‐BPy batches. Reference experiments also showed that different aggregation processes are relevant during production and further processing. Initial reprocessing of carbon nanotube raw material requires 7 h sonication time and over 24 h shear mixing before no increase in carbon nano concentration is detectable. However, only a few minutes of sonication or shear mixing are required when reprocessing the residue produced during the separation of the slurry. This discrepancy indicates that different aggregates are present, with markedly different aggregation properties. To study low‐dimensional heterostructures, a PL microscope was set up with the ability to ob‐ serve single SWNTs as well as monolayers of other low‐dimensional systems. Furthermore, sam‐ ples were prepared which bring single SWNTs into contact with 2D materials such as h‐BN andMoS2 layers and the changes in the photoluminescence spectrum were documented. For h‐BN, it was observed whether previous methods for depositing SWNTs could be transferred for photo‐ luminescence spectroscopy. SWNTs were successfully deposited on monolayers via a modified drip coating, with the limitation that SWNTs aggregate more at the edges of the monolayers. Upon contact of SWNTs with MoS2, significant changes in the emission properties of the mono‐ layers were observed. The fluorescence, which was mainly dominated by excitons, was shifted towards trion emission. Reference experiments excluded PFO‐BPy and toluene as potential causes. Based on the change in the emission behaviour of MoS2, the most plausible explanation is a photoinduced charge transfer leading to delocalised charge carriers on MoS2. In contrast, on SWNTs, the introduction of additional charges would constitute a quenching centre, which would quench their PL emission, making them undetectable in the PL image. In the last chapter, the electronic properties of doped SWNTs and the behaviour of charge carri‐ ers inside the tubes should be investigated. First, the change in the conductivity of SWNT films with increasing doping levels was docu‐ mented. The resistance of the films drops drastically at minimum doping. After the initial in‐ troduction of charges, the resistance drops with increasing dopant concentration according to a double logarithmic curve. The initial drop could be due to a reduction of contact resistances within the SWNT network film, but this could not be further investigated within the scope of this PhD thesis. In cooperation with Andreas Sperlich and Michael Auth, the spin concentration of SWNTs at different doping levels was determined. The obtained concentrations were compared with the carrier concentrations determined from PL and absorption spectra. At low spin densities, good agreement with previous models was found. Furthermore, the presence of isolated spins strongly suggests a localised charge carrier distribution at temperatures around 10 K. When the charge density is increased, the spin density deviates significantly from the charge carrier con‐ centration. This discrepancy is attributed to the increasing delocalisation of charge carriers at high charge densities and the interactions of neighbouring spins. These results strongly indicate the existence of localised charge carriers in SWNTs at low temperatures. Next, the effect of doping on the Raman spectra of SWNT suspensions was investigated. In gen‐ eral, doping is expected to reduce the intensity of the Raman bands, i.e. a consequence of the reduced resonance gain due to bleaching of the S2 transition. However, similar to the resistivity measurements, the oscillator strength of the G+ band drops sharply in the first doping steps. It was also found that the G+ band decreases more than would be expected due to loss of reso‐ nance condition. Furthermore, the G‐ is bleached faster than the G+ band. All these anomalies suggest that resonance enhancement is not the only relevant effect. Another much faster deac‐ tivation path for the excitons may be introduced by doping. This would leave less time for the scattering process to occur and reduce the oscillator strength of the Raman bands. In cooperation with Vidmantas et al., the photoinduced charge carrier behaviour of SWNT/PCBM films was investigated. The required films were prepared by drop coating. The SWNT suspen‐ sions required for this were obtained from sheared SWNT preparations. Using transient absorp‐ tion and TREFISH, a number of charge transfer effects were identified and their dynamics in‐ vestigated: the recombination of neutral excitons (< 50 ps), the electron transfer from carbon nanotubes to PCBM molecules (< 1 ps), the decay of charge‐transfer excitons (∼200 ps), the recombination of charge carriers between charge‐transfer excitons (1 ns to 4 ns) and finally the propagation through the SWNT network (∼20 ns)
Quantitative Electron Paramagnetic Resonance Studies of Charge Transfer in Organic Semiconductors
(2020)
In the present work we investigated various charge transfer processes, as they appear in the versatile world of organic semiconductors by probing the spin states of the corresponding charge carrier species via electron paramagnetic resonance (EPR) spectroscopy. All studied material systems are carbon-based compounds, either belonging to the group of polymers, fullerenes, or single-wall carbon nanotubes (SWNTs).
In the first instance, we addressed the change of the open circuit voltage (Voc) with the fullerene blend stoichiometry in fullerene-based solar cells for organic photovoltaics (OPV). The voltage depends strongly on the energy separation between the lowest unoccupied molecular orbital (LUMO) of the donor and the highest occupied molecular orbital (HOMO) of the acceptor. By exploiting the Gaussian distribution of the charge carriers in a two-level system, and thus also their spins in the EPR experiment, it could be shown that the LUMOs get closer by a few to a few hundred meV when going from pure fullerene materials to a fullerene mixture. The reason for this strong energetic effect is likely the formation of a fullerene alloy.
Further, we investigated the chemical doping mechanism of SWNTs with a (6,5)-chirality and their behaviour under optical excitation. In order to determine the unintentional (pre)-doping of SWNTs, EPR spectra of the raw material as well as after different purification steps were recorded. This facilitated the determination of nanotube defects and atmospheric p-doping as the causes of the measured EPR signals. In order to deliberately transfer additional charge carriers to the nanotubes, we added the redox-active substance AuCl3 where we determined an associated doping-yield of (1.5±0.2)%. In addition, a statistical occupation model was developed which can be used to simulate the distribution of EPR active, i.e. unpaired and localised charge carriers on the nanotubes.
Finally, we investigated the charge transfer behaviour of (6,5)-SWNTs together with the polymer P3HT and the fullerene PC60BM after optical excitation.
In order to shrink the size of semiconductor devices and improve their
efficiency at the same time, silicon-based semiconductor devices have
been engineered, until the material almost reaches its performance
limits. As the candidate to be used next in semiconducting devices,
single-wall carbon nanotubes show a great potential due to their
promise of increased device efficiency and their high charge carrier
mobilities in the nanometer size active areas. However, there are
material based problems to overcome in order to imply SWNTs in the
semiconductor devices. SWNTs tend to aggregate in bundles and it is
not trivial to obtain an electronically or chirally homogeneous SWNT
dispersion and when it is done, a homogeneous thin film needs to be
produced with a technique that is practical, easy and scalable. This
work was aimed to solve both of these problems.
In the first part of this study, six different polymers, containing
fluorene or carbazole as the rigid part and bipyridine, bithiophene or
biphenyl as the accompanying copolymer unit, were used to selectively
disperse semiconducting SWNTs. With the data obtained from
absorption and photoluminescence spectroscopy of the corresponding
dispersions, it was found out that the rigid part of the copolymer plays a
primary role in determining its dispersion efficiency and electronic
sorting ability. Within the two tested units, carbazole has a higher π
electron density. Due to increased π−π interactions, carbazole
containing copolymers have higher dispersion efficiency. However, the
electronic sorting ability of fluorene containing polymers is superior.
Chiral selection of the polymers in the dispersion is not directly
foreseeable from the selection of backbone units. At the end, obtaining a monochiral dispersion is found to be highly dependent on the used raw
material in combination to the preferred polymer.
Next, one of the best performing polymers due to high chirality
enrichment and electronic sorting ability was chosen in order to
disperse SWNTs. Thin films of varying thickness between 18 ± 5 to
755o±o5 nm were prepared using vacuum filtration wet transfer method
in order to analyze them optically and electronically.
The scalability and efficiency of the integrated thin film production
method were shown using optical, topographical and electronic
measurements. The relative photoluminescence quantum yield of the
radiative decay from the SWNT thin films was found to be constant for
the thickness scale. Constant roughness on the film surface and linearly
increasing concentration of SWNTs were also supporting the scalability
of this thin film production method. Electronic measurements on bottom
gate top contact transistors have shown an increasing charge carrier
mobility for linear and saturation regimes. This was caused by the
missing normalization of the mobility for the thickness of the active
layer. This emphasizes the importance of considering this dimension for
comparison of different field effect transistor mobilities.
Nanoelectronics is an essential technology for down-scaling beyond the limit of silicon-based electronics. Single-Wall Carbon Nanotubes (SWNT) are semiconducting components that exhibit a large variety of properties that make them usable for sensing, telecommunication, or computational tasks. Due to their high surface to volume ratio, carbon nanotubes are strongly affected by molecular adsorptions, and almost all properties depend on surface adsorption. SWNT with smaller diameters (0.7-0.9nm) show a stronger sensitivity to surface effects. An optimized synthesis route was developed to produce these nanotubes directly. They were produced with a clean surface, high quality, and large lengths of 2 μ m. The results complement previous studies on larger diameters (0.9-1.4nm). They allow performing statistically significant assumptions for a perfect nanotube, which is selected from a subset of nanotubes with good emission intensity, and high mechanical durability. The adsorption of molecules on the surface of carbon nanotubes influences the motion and binding strength of chargeseparated states in this system. To gain insight into the adsorption processes on the surface with a minimum of concurrent overlapping effects, a microscopic setup, and a measurement technique were developed. The system was estimated to exhibit excellent properties like long exciton diffusion lengths (>350nm), and big exciton sizes (8.5(5)nm), which was substantiated by a simulation. We studied the adsorption processes at the surface of Single-Wall Carbon Nanotubes for molecules in the gas phase, solvent molecules, and surfactant molecules. The experiments were all carried out on suspended individualized carbon nanotubes on a silicon wafer substrate. The experiments in the gas-phase showed that the excitonic emission energy and intensity experiences a rapid blue shift during observation. This shift was associated with the spontaneous desorption of large clusters of gaseous molecules caused by laser heat up. The measurement of this desorption was essential for creating a reference to an initially clean surface and allows us to perform a comparison with previous measurements on this topic. Furthermore, the adsorption of hydrogen on the nanotube surface at high temperatures was investigated. It was found that a new emission mode arises slightly red-shifted to the excitonic emission in these systems. The new signal is almost equally strong as the main excitonic peak and was associated with the brightening of dark excitons at sp3-defects through a K-phonon assisted pathway. The finding is useful for the direct synthesis of spintronic devices as these systems are known to act as single-photon emitters. The suspended nanotubes were further studied to estimate the effect of solvent adsorption on the excitonic states during nanotube dispersion for each nanotube individually. A significant quantum yield loss is observable for hexane and acetonitrile, while the emission intensity was found to be the strongest in toluene. The reference to a clean surface allowed us to estimate the exact influence of the dielectric environment of adsorbing solvents on the excitonic emission energy. Solvent adsorption was found to lead to an energy shift that is almost twice as high as suggested in previous studies. The amount of this energy shift, however, was comparably similar for all solvents, which suggests that the influence of the distinct dielectric constant in the outer environment less significantly influences the energy shift than previously thought. An interesting phenomenon was found when using acetonitrile as a solvent, which leads to greatly enhanced emission properties. The emission is more than twice as high as in the same air-suspended nanotubes, which suggests a process that depends on the laser intensity. In this study, it was reasonably explained how an energy down-conversion is possible through the coupling of the excitonic states with solvent vibrations. The strength of this coupling, however, also suggests adsorptions to the inside of the tubular nanotube structure leading to a coupled vibration of linear acetonitrile molecules that are adsorbed to the inner surface. The findings are important for the field of nanofluidics and provide an excellent system for efficient energy down-conversion in the transmission window of biological tissue. Having separated the pure effect of solvent adsorption allowed us to study the undisturbed molecular adsorption of polymers in these systems. The addition of polyfluorene polymer leads to a slow but stepwise intensity increase. The intensity increase is overlapping with a concurrent process that leads to an intensity decrease. Unfortunately, observing the stepwise process has a low spacial resolution of only 100-250nm, which is in the range of the exciton diffusion length in these systems and hinders detailed analysis. The two competing and overlapping processes processes are considered to originate from slow π-stacking and fast side-chain binding. Insights into this process are essential for selecting suitably formed polymers. However, the findings also emphasize the importance of solvent selection during nanotube dispersion since solvent effects were proven to be far more critical on the quantum yield in these systems. These measurements can shed light on the ongoing debate on polymers adsorption during nanotube individualization and allow us to direct the discussion more towards the selection of suitable solvents. This work provides fundamental insights into the adsorption of various molecules on the surface of individually observed suspended Single-Wall Carbon Nanotubes. It allows observing the adsorption of individual molecules below the optical limit in the solid, liquid, and gas phases. Nanotubes are able to act as sensing material for detecting changes in their direct surrounding. These fundamental findings are also crucial for increasing the quantum yield of solvent-dispersed nanotubes. They can provide better light-harvesting systems for microscopy in biological tissue and set the base for a more efficient telecommunication infrastructure with nano-scale spintronics devices and lasing components. The newly discovered solvent alignment in the nanotube surrounding can potentially also be used for supercapacitors that are needed for caching the calculation results in computational devices that use polymer wrapped nanotubes as transistors. Although fundamental, these studies develop a strategy to enlighten this room that is barely only visible at the bottom of the nano-scale.