This thesis deals with the characterization and implementation of transparent electron transport materials (ETM) in vacuum deposited p-i-n type organic solar cells (OSC) for substituting the parasitically absorbing standard ETM composed of n-doped C60. In addition to transparency in the visible range of the sun spectrum, the desired material properties include high electron mobility and conductivity, thermal and morphological stability, as well as good energy level alignment relative to the adjacent acceptor layer which is commonly composed of intrinsic C60. In this work, representatives of three different material classes are evaluated with regard to the above mentioned criteria.
HATCN (hexaazatriphenylene hexacarbonitrile) is a small discoid molecule with six electron withdrawing nitrile groups at its periphery. It forms smooth thin films with an optical energy gap of 3.3eV, thus being transparent in the visible range of the sun spectrum. Doping with either 5wt% of the cationic n-dopant AOB or 7wt% of the proprietary material NDN1 effectively increases the conductivity to 7.6*10^-6 S/cm or 2.2*10^-4 S/cm, respectively. However, the fabrication of efficient OSC is impeded by the exceptionally high electron affinity (EA ) of approximately 4.8eV that causes the formation of an electron injection barrier between n-HATCN and intrinsic C60 (EA=4.0eV). This work presents a strategy to remove the barrier by introducing doped and undoped C60 intermediate layers, thus demonstrating the importance of energy level matching in a multi-layer structure and the advantages of Fermi level control by doping.
Next, a series of six Bis-Fl-NTCDI (N,N-bis(fluorene-2-yl)-naphthalenetetracarboxylic diimide) compounds, which only differ by the length of the alkyl chains attached to the C9 positions of the fluorene side groups, is examined. When increasing the chain length from 0 to 6 carbon atoms, the energy levels remain nearly unchanged: We find EA=3.5eV as estimated from cyclic voltammetry, an ionization potential (IP ) in the range between 6.45eV and 6.63eV, and Eg,opt=3.1eV which means that all compounds form transparent thin films. Concerning thin film morphology, the addition of side chains results in the formation of amorphous layers with a surface roughness <1nm on room temperature glass substrates, and (1.5+/-0.5)nm for deposition onto glass substrates heated to 100°C. In contrast, films composed of the side chain free compound Bis-HFl-NTCDI exhibit a larger surface roughness of (2.5+/-0.5)nm and 9nm, respectively, and are nanocrystalline already at room temperature. Moreover, the conductivity achievable by n-doping is very sensitive to the side chain length: Whereas doping of Bis-HFl-NTCDI with 7wt% NDN1 results in a conductivity in the range of 10^-4 S/cm, the attachment of alkyl chains causes a conductivity which is more than three orders of magnitude smaller despite equal or slightly higher doping concentrations. The insufficient transport properties of the alkylated derivatives lead to the formation of pronounced s-kinks in the jV -characteristics of p-i-n type OSC while the use of n-Bis-HFl-NTCDI results in well performing devices.
The last material, HATNA-Cl6 (2,3,8,9,14,15- hexachloro-5,6,11,12,17,18-hexaazatrinaphthylene), exhibits Eg,opt=2.7eV and is therefore not completely transparent in the visible range of the sun spectrum. However, its energy level positions of EA=4.1eV and IP=7.3eV are well suited for the application as ETM in combination with i-C60 as acceptor. The compound is dopable with all available n-dopants, resulting in maximum conductivities of sigma=1.6*10^-6, 3.5*10^-3, and 7.5*10^-3 S/cm at 7.5wt% AOB, Cr2(hpp)4, and NDN1, respectively. Applying n-HATNA-Cl6 instead of the reference ETM n-C60 results in a comparable or improved photocurrent density at an ETM thickness d(ETM)=40nm or 120nm, respectively. At d(ETM)=120nm, the efficiency eta is more than doubled as it increases from eta(n-C60)=0.4% to eta(n-HATNA-Cl6)=0.9% .
Optical simulations show that the replacement of n-C60 by n-Bis-HFl-NTCDI, n-HATNA-Cl6, or the previously studied n-NTCDA (naphthalenetretracarboxylic dianhydride) in p-i-n or n-i-p type device architectures is expected to result in an increased photocurrent due to reduced parasitic absorption. For quantifying the gain, the performance of p-i-n type OSC with varying ETM type and thickness is evaluated. Special care has to be taken when analyzing devices comprising the reference ETM n-C60 as its conductivity is sufficiently large to extend the area of the aluminum cathode and thus the effective device area which may lead to distorted results. Overall, the experiment is able to confirm the trends predicted by the optical simulation. At large ETM thickness in the range between 60 and 120nm, the window layer effect of the ETM is most pronounced. For instance, at d(ETM)=120nm, eta(C60) is more than doubled using n-HATNA-Cl6 and even more than tripled using n-Bis-HFl-NTCDI or n-NTCDA. At optimized device geometry the photocurrent gain is slightly less than expected but nonetheless, the efficiency is improved from eta(max)=2.1% for n-C60 and n-HATNA-Cl6 solar cells to eta(max)=2.3, and 2.4% for n-Bis-HFl-NTCDI and n-NTCDA devices, respectively. This development is supported by generally higher Voc and FF in solar cells with transparent ETM.
Finally, p-i-n type solar cells with varying ETM are aged at a temperature of 50°C and an illumination intensity of approximately 2 suns. Having extrapolated lifetimes t(80) of 36, 500, and 14000h and nearly unchanged jV-characteristics after 2000h, n-C60 and n-Bis-HFl-NTCDI devices exhibit the best stability. In contrast, n-NTCDA devices suffer from a constant decrease in Isc while n-HATNA-Cl6 solar cells show a rapid dscegradation of both Isc and FF associated with a decomposition of the material or a complete de-doping of the ETM. Here, lifetimes of only 4500h and 445hare achieved.
| Identifer | oai:union.ndltd.org:DRESDEN/oai:qucosa.de:bsz:14-qucosa-89214 |
| Date | 15 October 2012 |
| Creators | Falkenberg, Christiane |
| Contributors | Technische Universität Dresden, Fakultät Mathematik und Naturwissenschaften, Prof. Dr. Karl Leo, Prof. Dr. Karl Leo, Prof. Dr. Klaus Müllen |
| Publisher | Saechsische Landesbibliothek- Staats- und Universitaetsbibliothek Dresden |
| Source Sets | Hochschulschriftenserver (HSSS) der SLUB Dresden |
| Language | English |
| Detected Language | English |
| Type | doc-type:doctoralThesis |
| Format | application/pdf |
Page generated in 0.0032 seconds