X-ray Photoemission, Absorption and Scattering Studies of Metal Oxide and Organic Thin-Films and Interfaces for Photovoltaic Applications

Organic photovoltaics (OPV) have attracted a lot of attention during the last couple ofdecades, due to them having several advantageous properties such as light-weight, abundantand non-toxic materials, mechanical flexibility, semi-transparency, low energy pay back timesand low cost manufacturing, which favours them over conventional Si-based PV, when itcomes to price, pay-back times, accessibility, production speed, environmental impact andease of installation. Currently, some OPV devices have reached power conversion efficiency(PCE) values that exceed 19%, which is getting close to Si-based PV, however OPV are stillchallenged by shorter lifetimes and poor long-term stability, which makes it hard for them tocompete with their Si-based counterparts. In order for large scale commerialization of OPVwith significant PV market share to become a reality, it is crucial to improve the device performance and stability of OPV.

This can be achieved by device engineering and by introducingnew efficient and stable materials (thin-films), but to do that, we need a deeper understandingof the structure and optoelectronic properties of such materials. In addition, it is also important to understand the energy level alignment and electronic interactions at the interfacesbetween materials in an OPV device. The structure, crystallinity, composition and electronicstructure of thin-films can be studied using X-ray based characterization techniques, sinceX-rays have wavelengths comparable to inter-atomic distances and photon energies comparable to energy differences between electronic levels in materials. In this thesis, three X-raysbased characterization techniques; X-ray photoelectron spectroscopy (XPS), Near-edge X-rayabsorption Fine Structure (NEXAFS) and X-ray scattering (XRD) have been used to studyfunctional materials, namely metal oxide and organic thin films for OPV applications. Thisincludes their interfaces and their electronic interactions at their interfaces, to gain a deeperinsight of the structure and properties of these thin-films, and their interfaces and thereby tounveil their contribution to the performance and stability of OPV.A former study from our group has shown that using the organic molecule, 4P-NPD, as an exciton blocking layer (EBL) in DBP:C70 based OPV devices resulted in an improvement of thePCE of up to 24% compared to reference devices. This PCE enhancement was only observedwhen ultra-thin 4P-NPD layers of maximum 1 nm were used, however when the thickness of4P-NPD was increased, the resulting devices performed worse than reference devices, whichsuggests that the interface plays an important part in the process. The 4P-NPD EBL wasimplemented between the DBP donor and the MoOx hole transport layer (HTL). In order tounderstand, why only ultra-thin 4P-NPD layers improve device performance, while thickerlayers result in diminished device performance, we studied the in-situ prepared DBP/4PNPD and 4P-NPD/MoOx interfaces using synchrotron based XPS and NEXAFS. We observea small HOMO off-set at the DBP/4P-NPD interface, which enables hole transport betweenthe two materials, however for thicker layers, this HOMO off-set increases and creates a holeextraction barrier. We also observe an upwards shift of all the electronic levels of 4P-NPDat the 4P-NP/MoOx interface, which results in a large LUMO off-set between DBP and 4PNPD that blocks electrons and thereby makes the interface hole-selective. Additionally, an upwards shift of the 4P-NPD HOMO level removes any extraction barrier towards DBP. Afterincreasing the 4P-NPD thickness, the electronic levels of 4P-NPD relax to their initial positions, which results in a small LUMO off-set that does not block electrons and the recoveryof a HOMO extraction barrier. These findings explain why only ultra-thin layers of 4P-NPDimprove the device performance.

In another study, we show that using sputter deposited TiOx films, deposited at 150oC andcooled in the presence of oxygen, as electron transport layers (ETLs) in PBDB-T:ITIC basedOPV devices, results in a significant improvement of the long-term device stability compared to conventionally used solution processed ZnO ETLs. The devices with the sputteredTiOx ETLs saturated at 60% of their initial PCE after 14 days of degradation, while thedevices with ZnO saturated at 35% of their initial PCE after 14 days of degradation. Optical spectroscopy showed that this improved stability results from a reduced UV-inducedphoto-catalytic degradation of the ITIC acceptor by the sputtered TiOx compared to ZnO.This is a surprising finding, since TiOx is also known to have a strong photo-catalytic activity. To understand the mechanisms behind the decreased photo-catalytic degradation ofITIC by the sputtered TiOx, we studied the defect states and electronic state interplay at thesemi in-situ prepared sputtered TiOx/ITIC interface using synchrotron based XPS, NEXAFSand resonant photoemission spectroscopy (RESPES). These results were compared to similar measurements on an anatase TiOx single crystal, since anatase TiOx is known to have astrong photo-catalytic activity and therefore acts as a good reference for our sputtered oxide.We observe a defect feature in the valence band of both sputtered and anatase TiOx, whichis related to oxygen vacancies. RESPES measurements showed, that all the oxygen vacancies in the anatase crystal lie in the sub-surface region, while the sputtered TiOx has bothsurface and sub-surface vacancies, however the surface vacancies strongly dominate. Thismajor difference could explain the reduced photo-catalytic activity of the sputtered TiOx,since surface oxygen vacancies can be quenched by surface adsorbates, which prevents themfrom participating in the photo-catalysis process.

In another stability study of PBDB-T:ITIC devices, we replaced conventional thermally deposited MoOx HTLs with crystalline sputter deposited MoOx films. Seven different sputteredMoOx films were tested, post-annealed in vacuum at 200oC and 350oC with different heatingrates. Stability testing showed, that all the devices with the 200oC annealed films had an improved long-term stability compared to devices with thermally deposited MoOx. The heatingrate did affect device stability and the champion devices, which saturated at 47% of theirinitial PCE after 215 h of degradation, were the devices with the 200oC annealed sputteredMoOx annealed with a heating rate of 45 K/min. In comparison, the devices with the thermalMoOx saturated at 25% of their initial PCE after 215 h of degradation. The 350oC annealedfilms resulted in poor long-term device stability comparable to thermal MoOx, which can beexplained by a larger concentration of adsorbed OH defects and a larger surface roughness.

The degradation behaviour of the PM6 polymer donor and the Y7 NFA was studied usingsynchrotron based XPS. PM6 was degraded in-situ by X-rays and light/heat in vacuum,while Y7 was degraded in-situ by X-rays in vacuum. Degradation of PM6 by X-rays showedevidence of polymerization, while degradation of PM6 by light and heat in show evidence ofbonds being broken. Degradation of Y7 by X-rays resulted in large shifts of the C1s core leveland HOMO on-set towards lower binding energies. The shifts observed in the Y7 core levelsand HOMO on-set after degradation by X-rays are larger than the shifts observed for PM6after degradation by X-rays, which suggests that Y7 is more sensitive to X-rays than PM6.