Quantitative Spectroscopic Investigations of Organic Photovoltaic Material Degradation

Abstract

As global average temperatures continue to rise despite mitigation efforts, the need to transition from traditional fossil fuels to renewable energy sources is more critical than ever. Although the adoption of photovoltaic (PV) devices as an alternative energy source have steadily increased over the past two decades, traditional Si solar panels have several drawbacks, such as inflexible module designs, high manufacturing costs, and the incorporation of toxic materials. Organic photovoltaics (OPVs) are a potential solution to the deficiencies of the Si PV module. However, OPVs suffer from inherent chemical instabilities in ambient environmental conditions. Further, there are several deficiencies in the traditional spectroscopic techniques and studies used to investigate chemical degradation of OPVs. Namely, these are the lack of a universal substrate for spectroscopic studies, qualitative and time consuming interpretation of collected spectral data, and experimental design that does not incorporate the complexities of operating devices. Therefore, there exists a need for data-driven analyses that provide actionable information about the chemical degradation in working OPVs To address these shortcomings, this dissertation applies chemometrics and two-dimensional correlation spectroscopy (2D-COS) to FTIR spectral data of two model OPV materials: 4,7-bis(9,9-dimethyl-9H-fluoren-2-yl)benzo[c][1,2,5]thiadiazole (FBTF) and poly[[4,8-bis[(2-ethylhexyl)oxy]benzo[1,2-b:4,5-b']dithiophene-2,6-diyl][3-fluoro-2-[(2- ethylhexyl)carbonyl]thieno[3,4-b]thiophenediyl ]] (PTB7). To investigate the effect of substrate on photodegradation mechanisms, FBTF was drop cast onto either indium tin oxide (ITO)-coated glass slides or Ag substrates before cumulative exposure to an Xe lamp simulating early stage photodegradation. Differences in the collected infrared reflectance-absorption (IRRA) spectra strongly suggest substrate-dependent pathways. Principal component analysis (PCA) and linear discriminant analysis (LDA) of the IRRA spectral data of FBTF revealed the ability of multivariate techniques to discriminate spectra based on substrate type and degradation extent despite the visual similarities during early-stage degradation. To enhance the understanding of the different degradation mechanisms observed in FBTF on ITO and Ag substrates, 2D-COS was applied to both spectral datasets. The use of 2D-COS techniques allowed for more specific determination about when the degradation mechanisms on each substrate diverged and the order of spectral intensity changes of heavily overlapped degradation products. Given the success of PCA-LDA for discrimination of FBTF spectral data, PTB7 was chosen as a second system to test the ability of chemometric techniques to discriminate spectral data by degradation extent. A dataset of 132 ATR-FTIR spectra of PTB7 thin films subjected to cumulative 30-min exposure intervals to a solar simulator was collected and used to train a PCA-LDA model. To simulate real-world data, the PCA-LDA model was used on test datasets collected and/or preprocessed by different analysts. Over the selective frequency region of 1300-1900 cm-1, the model was able to classify 67.4% of the samples by their exact number of 30-min exposure intervals and >90% of the samples to within ±30 min of their exact exposure interval. Expanding the range to 1000-3400 cm-1 resulted in similarly satisfactory outcomes, with 62.1% accuracy for the exact 30-min interval and >90% of the samples to within ±60 min. When tasked with classifying the spectra as either pristine or degraded, the model accurately classified 100% of the spectra over 1300-1900 cm-1 and 89.4% of the spectra over 1000-3400 cm-1. Besides photochemical degradation, working OPVs are also vulnerable to voltage-induced degradation arising from the electric field of the devices. The moving charges in OPV devices result in polaron forms of the organic material that could have different chemical reactivities than the neutral species. PTB7 films were exposed to either application of voltage in dark conditions, photodegradation, and the simultaneous combination of (light+voltage). Based on differences in IRRA spectra and 2D-COS analysis, it was concluded that different reactive oxygen species were responsible for degradation under each experimental condition. The application of voltage in dark conditions combined with doping of atmospheric O2 led to superoxide anion generation and film degradation which is a pathway that differs from purely photodegradative processes where singlet oxygen is known to be the main driver of degradation. In the simultaneous condition of (light+voltage), both reactive oxygen species are present although the greater spectral and 2D-COS similarity to the voltage-only condition led to the conclusion that the addition of voltage directed degradation to the more reactive polaron sites. To further study the effects of different perturbations, a more device-like sample configuration was designed to better simulate a working OPV. For ATR-FTIR measurements, PTB7 thin films were sandwiched between a Au layer deposited onto a ZnSe internal reflection element (IRE) and a transparent anode made from sputtered ITO. As with the previous experiments, the individual degradative conditions of applied voltage and photodegradation were tested along with the simultaneous application of (light+voltage). 2D-COS plots clearly showed that the addition of the Au and ITO layers changed the chemical degradation observed in PTB7 and provided further evidence of the limitation of studying degradation in isolated thin films. Overall, the data suggests that PTB7 films in this device configuration undergo more similar chemical processes upon application of each degradative condition than PTB7 films experiencing greater atmospheric exposure. In each case, spectral evidence of a common degradative pathway leading to thioester formation was observed. However, the significant difference in the timescales necessary for thioester formation provides more evidence that different reactive oxygen species are dominant for each perturbation.