Refine
Has Fulltext
- yes (15)
Is part of the Bibliography
- yes (15)
Document Type
- Doctoral Thesis (15)
Keywords
- Organische Solarzelle (15) (remove)
Institute
ResearcherID
- D-1250-2010 (1)
Besides established, conventional inorganic photovoltaics—mainly based on silicon—organic photovoltaics (OPV) are well on the way to represent a lowcost, environment friendly, complementary technology in near future. Production costs, solar cell lifetime and performance are the relevant factors which need to be optimized to enable a market launch of OPV. In this work, the efficiency of organic solar cells and their limitation due to charge carrier recombination are investigated. To analyze solar cells under operating conditions, time-resolved techniques such as transient photovoltage (TPV), transient photocurrent (TPC) and charge extraction (CE) are applied in combination with time delayed collection field (TDCF) measurements. Solution processed and evaporated samples of different material composition and varying device architectures are studied. The standard OPV reference system, P3HT:PC61BM, is analyzed for various temperatures in terms of charge carrier lifetime and charge carrier density for a range of illumination intensities. The applicability of the Shockley Equation for organic solar cells is validated in case of field-independent charge photogeneration. In addition, a consistent model is presented, directly relating the ideality factor to the recombination of free with trapped charge carriers in an exponential density of states. An approach known as j=V reconstruction enables to identify the performance limiting loss mechanism of as-prepared and thermally treated P3HT:PC61BM solar cells. This procedure, involving TPV, CE and TDCF measurements, is extended to samples based on the rather new, low-band gap polymer PTB7 in combination with PC71BM. While in the devices processed from pure chlorobenzene solution considerable geminate and nongeminate losses are observed, the use of a solvent additive facilitates efficient polaron pair dissociation minimizing geminate recombination. Finally, in collaboration with the IMEC institute in Leuven, the two main organic solar cell device architectures, planar and bulk heterojunction—both based on CuPc and C60—are directly compared in terms of nongeminate recombination and charge carrier distribution. Two experimental techniques, TPV and CE, as well as a macroscopic device simulation are applied to reveal the origin of different Voc vs. light intensity dependence.
In the field of organic photovoltaics, one of the most intensely researched topics to date is the charge carrier photogeneration in organic bulk heterojunction solar cells whose thorough understanding is crucial for achieving higher power conversion efficiencies. In particular, the mechanism of singlet exciton dissociation at the polymer–fullerene interface is still controversially debated.
This work addresses the dissociation pathway via relaxed charge transfer states (CTS) by investigating its field dependence for reference material systems consisting of MDMO-PPV and one of the fullerene derivatives PC61BM, bisPCBM and PC71BM. Field dependent photoluminescence (PL(F)) and transient absorption (TA(F)) measurements give insight into the recombination of charge transfer excitons (CTE) and the generation of polarons, respectively. Optically detected magnetic resonance and atomic force microscopy are used to characterize the morphology of the samples.
The comparison of the experimental field dependent exciton recombination recorded by PL(F) and the theoretical exciton dissociation probability given by the Onsager–Braun model yields the exciton binding energy as one of the key parameters determining the dissociation efficiency. The binding energies of both the singlet exciton in neat MDMO-PPV and the CTE in MDMO-PPV:PC61BM 1:1 are extracted, the latter turning out to be significantly reduced with respect to the one of the singlet exciton.
Based on these results, the field dependence of CTE dissociation is evaluated for MDMO-PPV:PC61BM blends with varying fullerene loads by PL(F) and TA(F). For higher PC61BM contents, the CTE binding energies decrease notably. This behavior is ascribed to a larger effective dielectric constant for well-intermixed blends and to an interplay between dielectric constant and CTE delocalization length for phase separated morphologies, emphasizing the importance of high dielectric constants for the charge carrier photogeneration process.
Finally, the CTE binding energies are determined for MDMO-PPV blends with different fullerene derivatives, focusing on the influence of the acceptor LUMO energy. Here, the experimental results suggest the latter having no or at least no significant impact on the binding energy of the CTE. Variations of this binding energy are rather related to different trap levels in the acceptors which seem to be involved in CTS formation.
The focus of this work is studying recombination mechanisms occurring in organic solar cells, as well as their impact on one of their most important parameters — the open circuit voltage Voc.
Firstly, the relationship between Voc and the respective charge carrier density n in the active layer under open circuit conditions is analyzed. Therefor, a model after Shockley for the open circuit voltage is used, whose validity is proven with the aid of fits to the measured data. Thereby, it is emphasized that the equation is only valid under special conditions. In the used reference system P3HT:PC61BM the fits are in agreement with the measurement data only in the range of high temperatures (150 - 300 K), where Voc increases linearly with decreasing temperature. At lower temperatures (50 – 150 K), the experiment shows a saturation of Voc. This saturation cannot be explained with the model by the measured falling charge carrier density with decreasing temperatures. In this temperature range Voc is not directly related to the intrinsic properties of the active layer. Voc saturation is due to injection energy barriers at the contacts, which is ascertained by macroscopic simulations. Furthermore, it is observed that Voc in the case of saturation is equivalent to the so-called built-in potential. The difference between the built-in potential and the energy gap corresponds thereby to the sum of the energy barriers at both contacts.
With the knowledge of the Voc(n) dependency for not contact limited solar cells, it is possible to investigate the recombination mechanisms of charge carriers in the active layer. For Langevin recombination the recombination rate is Rn2 (recombination order RO = 2), for Shockley-Read-Hall (SRH) Rn1 (RO=1); in various publications RO higher than two is reported with two main explanations.
1: Trap states for charge carriers exist in the respective separated phases, i.e. electrons in the acceptor phase and holes in the donor phase, which leads to a delayed recombination of the charge carriers at the interface of both phases and finally to an apparent recombination order higher than 2.
2: The enhanced R(n) dependency is attributed to the so called recombination prefactor, which again is dependent from n dependent mobility µ.
It is shown that for the system P3HT:PC61BM at room temperature the µ(n) dependency does nearly completely explain the higher RO but not at lower temperatures which in this case supports the first explanation. In the material system PTB7:PC71BM the increased RO cannot be explained by the µ(n) dependency even at room temperature.
To support the importance of trap states in combination with a phase separation for the explanation of the enhanced RO, additional trap states were incorporated in the solar cells to investigate their influence on the recombination mechanisms. To achieve this, P3HT:PC61BM solar cells were exposed to synthetic air (in the dark and under illumination) or TCNQ was added in small concentrations to the active layer which act as electron traps. For the oxygen degraded solar cell the recombination order is determined by a combination of open Voc-transients and Voc(n) measurements. Thereby, a continuous increase of the recombination order from 2.4 to more than 5 is observed with higher degradation times. By the evaluation of the ideality factor it can be shown that the impact of SRH recombination is increasing with higher trap concentration in relation to Langevin recombination. A similar picture is revealed for solar cells with TCNQ as extrinsic trap states.
Finally, a phenomenon called s-shaped IV-curves is investigated, which can sometimes occur for solar cells under illumination. As course of this a reduced surface recombination velocity can be found. Experimentally, the solar cells were fabricated using a special plasma treatment of the ITO contact. The measured IV-curves of such solar cells are reproduced by macroscopic simulations, where the surface recombination velocity is reduced. Hereby, it has to be distinguished between the surface recombination of majority and minority charge carriers at the respective contacts. The theory can be experimentally confirmed by illumination level dependent IV-curves as well as short circuit current density and open circuit voltage transients.
A quite new approach to low-cost mass production of flexible solar cells are organic photovoltaics. Even though the device efficiencies increased rapidly during the last years, further imporvements are essential for a successful market launch. One important factor influencing the device efficiency is the photocurrent of a solar cell, which is defined as the difference between the current under illumination and in the dark. In case of organic bulk heterojunction (BHJ) solar cells it is — in contrast to inorganic devices — dependent on the applied bias voltage. The voltage dependence results in a reduced fill factor and thus an even more pronounced influence of the photocurrent on the device efficiency. It is therefore crucial to understand the underlying processes determining the photocurrent in order to be able to further improve the solar cell performance. In a first step the photocurrent of P3HT:PC61BM devices was investigated by a pulsed measurement technique in order to prevent disturbing influences due to device heating under continous illumination. The resulting photocurrent was hyperbolic tangent like and featured a point symmetry, whose origin and meaning were discussed. In addition, the photocurrent was described by a combined model of Braun–Onsager and Sokel–Hughes theory for field dependent polaron pair dissociation and charge extraction, respectively. After this macroscopic view on the photocurrent, the focus of this work moves to the more basic processes determining the photocurrent: charge photogeneration and recombination. In a comparative study the field-dependence of these was investigated by time-delayed collection field (TDCF) measurements for two well-known reference systems, namely P3HT:PC61BM and MDMO-PPV:PC61BM. It was possible to identify two different dominating scenarios for the generation of free charge carriers. The first one — via a thermalized charge transfer state (CTS) — is clearly influenced by geminate recombination and therefore less efficient. In the second scenario, the free charge carriers are either generated directly or via an excited, “hot” CTS. In addition, clear differences in the nongeminate recombination dynamics of both material systems were found. Similar studies were also be presented with two modern low bandgap polymers which only differ by the bridging atom in the cyclopentadithiophene (PCPDTBT:PC71BM vs. Si-PCPDTBT:PC71BM). Such small changes in the chemical structure were already sufficient to affect the charge photogeneration as well as the morphology of the blend. These findings were set into relation to current–voltage characteristics in order to discuss the origin of the clear differences in the solar cell performance of both materials. Another crucial parameter limiting the solar cell efficiency is the builtin potential of a device. Within the range of semiconducting pn-junctions, Mott–Schottky analysis is an established method to determine the built-in potential. As it was originally derived for abrupt pn-junctions, its validity for organic BHJ solar cells — a bipolar, effective medium — was discussed. Experimental findings as well as the contradictions to Mott–Schottky theory indicated, that a direct transfer of this method to organic photovoltaics is not appropriate. Finally, the results obtained in the framework of the MOPS-project (Massengedruckte Organische Papier-Solarzellen) will be presented, in which the first completely roll-to-roll printed paper solar cells were realized.
In this work, the trap states in the conjugated polymer P3HT, often used as electron donor in organic bulk heterojunction solar cells, three commonly used fullerene based electron acceptors and P3HT:PC61BM blends were investigated. Furthermore, the trap states in the blend were compared with these of the pure materials. Concerning the lifetime of organic solar cells the influence of oxygen on P3HT and P3HT:PC61BM blends was studied. The experimental techniques used to investigate the trap states in the organic semiconductors were (fractional) thermally stimulated current (TSC) and current based deep level transient spectroscopy (Q-DLTS). Fractional TSC measurements on P3HT diodes revealed a quasi-continuous trap distribution. The distribution suggested two different traps in P3HT with approximately Gaussian energy distributions and maxima at about 50 meV and 105 meV. Thereby, the former was attributed to the tail states within the regular Gaussian density of states due to the low activation energy. The latter, deeper traps, however, exhibited a strong dependence on oxygen. Exposure of the P3HT diodes to oxygen, ambient air and synthetic (dry) air all revealed an increase of the deeper traps density with exposure time in the same manner. While the lower limit of the trap density in non aged P3HT samples was in the range of (1.0 − 1.2)×10^22 m^−3, it was more than doubled after an exposure of 50 h to air. An increase of the trap density with oxygen exposure time was also seen in the Q-DLTS measurements accompanied with an increase of the temperature dependence of the emission rates, indicating an enhanced formation of deeper traps. Due to the raise in density of the deeper traps, the charge carrier mobility in P3HT significantly decreased, as revealed by photo-CELIV measurements, resulting in a loss in mobility of about two orders of magnitude after 100 h exposure to synthetic air. The increased trap density was attributed to p-doping of P3HT by the transfer of an electron to adsorbed oxygen. This effect was partially reversible by applying vacuum to the sample for several hours or, more significantly, by a thermal treatment of the devices in nitrogen atmosphere. The trap states in the methanofullerenes PC61BM, bisPC61BM and PC71BM were investigated by TSC measurements. PC61BM yielded a broad quasi-continuous trap distribution with the maximum of the distribution at about 75 meV. The comparison of the TSC spectra of the three methanofullerenes exhibited significant differences in the trap states with higher activation energies of the most prominent traps in bisPC61BM and PC71BM compared to PC61BM. This probably originates from the different isomers bisPC61BM and PC71BM consist of. Each of the isomers yields different LUMO energies, where the lower ones can act as traps. The lower limit of the trap density of all of the three investigated fullerene derivatives exhibited values in the order of 10^22 m^−3, with the highest for bisPC61BM and the lowest for PC61BM. By applying fractional TSC measurements on P3HT:PC61BM solar cells, it was shown that the trap distribution in the blend is a superposition of the traps in pure P3HT and PC61BM and additional deeper traps in the range of about 250 meV to 400 meV. The origin of these additional traps, which can not be related to the pure materials, was attributed to a higher disorder in the blend and P3HT/PC61BM interfaces. This conclusion was supported by standard TSC and Q-DLTS measurements performed on pristine and annealed P3HT:PC61BM blends, exhibiting a higher ratio of the deep traps in the pristine samples. The lower limit of the trap density of the investigated annealed solar cells was in the range of (6−8)×10^22 m^−3, which was considerably higher than in the pure materials. The influence of oxygen on P3HT:PC61BM solar cells was investigated by exposure of the devices to synthetic air under specific conditions. Exposure of the solar cells to oxygen in the dark resulted in a strong decrease in the power conversion efficiency of 60 % within 120 h, which was only caused by a loss in short-circuit current. Simultaneous illumination of the solar cells during oxygen exposure strongly accelerated the degradation, resulting in an efficiency loss of 30 % within only 3 h. Thereby, short-circuit current, open-circuit voltage and fill factor all decreased in the same manner. TSC measurements revealed an increase of the density of deeper traps for both degradation conditions, which resulted in a decrease of the mobility, as investigated by CELIV measurements. However, these effects were less pronounced than in pure P3HT. Furthermore, an increase of the equilibrium charge carrier density with degradation time was observed, which was attributed to oxygen doping of P3HT. With the aid of macroscopic simulations, it was shown that the doping of the solar cells is the origin of the loss in short-circuit current for both degradation conditions.