Photo-degradation in bulk heterojunction organic solar cells using a fullerene or a non-fullerene derivative electron acceptor

Highlights

• An electron-donor polymer (PF2) blended with PC₇₁BM leads to a high power conversion efficiency with low stability.

• PF2 blended with a non-fullerene acceptor (EH-IDTBR) leads to a high stability with a lower efficiency.

• The higher stability is linked to an improved morphology despite a change in the main charge-carrier recombination mechanism.

• The lower stability is linked to a slight modification of the nano-morphology.

Abstract

The use of Non-Fullerene Acceptors (NFAs) in the active layer of organic solar cells (OSCs) has made it possible to exceed 18% conversion efficiency. However, OSCs still present stability issues under operational conditions that need to be surpassed for their industrialization. In this work, we investigated binary and ternary blends to examine their efficiency and their stability as active layers of OSCs. We used a fluorinated polymer (PF2) as an electron donor and two different electron acceptors, a fullerene derivative (PC₇₁BM) and a NFA (EH-IDTBR). We demonstrated that using EH-IDTBR instead of PC₇₁BM leads to a decrease in efficiency attributed to the low out-of-plane electron mobility measured in the blend. However, using EH-IDTBR as single electron-acceptor significantly enhanced the OSCs stability under continuous illumination. Ternary blends were tested to reach simultaneously a high efficiency and a long-term stability. The best efficiency/stability compromise appeared to be when using EH-IDTBR only as electron-acceptor. We identified changes in the main charge-carrier recombination mechanism in photo-degraded devices from bimolecular in low EH-IDTBR content blends to trap-assisted in high EH-IDTBR ones. Finally, the blend morphology at a nanometer scale appeared as stable in high EH-IDTBR content blends while photo-degradation impacted significantly the morphology of the low EH-IDTBR content blend.

Introduction

Organic solar cells (OSCs) have received a lot of attention in the photovoltaic field due to their interesting properties compared to their inorganic counterparts such as flexibility, lightweight, semi-transparency, low cost and large-scale manufacturing in roll-to-roll processes. Starting from the bulk heterojunction (BHJ) concept discovery [1], the Power Conversion Efficiency (PCE) of OSCs has increased from roughly 1% up to more than 18% nowadays [2]. In the BHJ configuration, the electron donor (D) and the electron acceptor (A) materials are intimately intermixed, forming nano-scaled continuous interpenetrating networks. This BHJ configuration enhances the free charge carriers' generation and transport. Fullerene derivatives have been widely used as electron acceptors in OSCs [[3], [4], [5]], thanks to their large electron affinity and high electron mobility with an isotropic charge transport. However, fullerene derivatives suffer from a weak absorption in the visible spectrum and from energy levels that are difficult to adjust. A strong bottleneck to reach PCEs higher than 10% with OSCs has long been the availability of alternative electron acceptors.

Researchers have made great efforts to design novel π-conjugated small molecules with adjusted frontier energy levels, high electron mobility and broad light absorption range. In the last two years, progress was made in non-fullerene acceptors (NFAs) design in order to overcome the 10% PCE barrier. Yuan et al. introduced DAD fused core in the push-pull structure of Y5 (BTP) [6]. Its fluorinated version (BTP-4F), also known as Y6, has emerged as a promising NFA giving a certified PCE approaching 15% in binary blends [7]. In order to promote the BHJ supramolecular organization to enhance the exciton dissociation and charge carriers transport and reduce the nonradiative energy losses, researchers adopted the ternary strategy by adding a third component to the main D/A binary blend [8,9]. The third component can be either a donor [[10], [11], [12]] or an acceptor [[13], [14], [15]], a polymer, a small molecule, or even a nanoparticle [16] and is selected according to its compatibility with the D and/or A of the main blend [17,18]. Recently, Sun et al. have made chemical structure adjustments on Y6 by modifying the branched alkyl chain [19], leading to a high efficiency in single junctions, around 18.60% (certified at 18.20%) in both binary [2] and ternary [20] blends. With the increasing interest in the ternary blends very recently, the F-BTA3 NFA has been added as the third component to the PBQx-TF:eC9-2Cl binary blend, which improved the photovoltaic performances and led to the record PCE of 19% in single-junction OSCs [21]. Moreover, tandem OSCs using NFAs have reached PCEs exceeding 20% [22].

A high PCE is not the only important property when considering a photovoltaic system. PV systems must operate effectively for long periods and in different environmental conditions. The Levelized Cost Of Energy (LCOE) [23,24] is defined as the total lifetime cost divided by the energy produced by the PV system. The LCOE is then strongly affected by the operating stability, which represents today a major challenge for OSCs. During their entire life, PV systems do not operate with the same efficiency. Jordan et al. [25] called the “degradation process” the progressive deterioration of the appearance, the photovoltaic performances and the safety of PV systems. It induces a decrease in efficiency over the years, owing to stress sources as thermal cycling, humidity, irradiation and mechanical shocks. Therefore, the investigation of PV system degradation has gained more interest, encouraging researchers of the National Renewable Energy Laboratory (NREL) [26] to analyze more than 11,000 degradation rates. They found a median degradation rate of 0.5–0.6%/year and a mean of 0.8–0.9%/year for crystalline silicon high-quality technologies. However, this rate could get higher in a hot and humid environment. The classical PV systems lifetime (T₈₀) warrantied by the manufacturers is about 20–25 years. That means in 25 years the efficiency decreases down to 80% of its initial value.

The standard qualification tests (ex: IEC 61215-crystalline silicon), developed initially by the Jet Propulsion Laboratory (JPL) in the early 1980s [27] are used under stress sources for a long duration to predict the PV system lifetime. However, the data obtained using these tests are not sufficient to deduce the PV system lifetime. Nevertheless, an accelerated test could predict the lifetime by artificially speeding up stress factors until the PV system failure. For instance, the empirical temperature-humidity acceleration model of Desombre [28] was used later in the JPL corrosion model [29]. The accelerated high temperature could damage the materials of the PV system. Further, the UV irradiation containing high energy photons could lead to degradation which does not realistically represent the solar irradiation.

Although the organic semiconductors are more sensitive to environmental conditions than their inorganic counterparts, there is no well-defined standard test for OSC lifetime. During the first three international summits on organic photovoltaic stability (ISOS), test protocols under stress have been established to increase the amount of stability data [30]. The OSC lifetime was investigated, for most cases, under continuous standard illumination of 100 mW/cm² and long operational lifetimes (T₈₀) have been demonstrated from hundreds of hours [31] to thousands of hours [32]. Du et al. found a predicted lifetime approaching ten years [33]. Very recently, a predicted lifetime as high as 30 years has even been published for a Non-Fullerene Acceptor (NFA) based OSC [34]. The OSC degradation phenomena, and the factors causing it, were highlighted in the dark and continuous simulated AM 1.5 irradiation, under either air exposure or controlled environment. Degradation can be due to the device structure. For instance, Kawano et al. concluded that the degradation of the OSC is related to the water diffusion into PEDOT:PSS layer resulting in the series resistance increase [35]. The OSC performance deterioration could also be related to chemical modification of the active layer. The PC₆₁BM-based OSC exposed to a sun-equivalent white light source for 120 h and under inert conditions showed the induced fullerene dimerization [36]. The exposure of PC₆₁BM-based OSC to natural outdoor conditions in the Negev-desert also induced the fullerene dimerization [37]. Accelerated tests have also been used for studying the OSC stability/degradation. Exposing OSCs to high light-intensity, so-called « photo-degradation » could accelerate their degradation process until the failure and enable an estimation of the lifetime [38].

In this work, we used a low bandgap fluorinated copolymer (PF2) [12,39] as an electron donor and ([6,6]-phenyl-C71-butyric acid methyl ester) (PC₇₁BM) as an electron acceptor. PF2 is known to self-organize in crystalline lamellas with face-on preferential orientation. That results in high hole mobility and the formation of pure domains in [PF2:PC₇₁BM] BHJ, leading to the efficient transport of charge carriers with limited trap-assisted recombination. In addition, PF2 showed good light-absorption complementarity with PC₇₁BM. On the one hand, due to the promising characteristics of the [PF2:PC₇₁BM] blend, an efficiency of above 10% has already been reported by our team [40]. On the other hand, encapsulated devices showed a relatively low stability with devices stored in the dark losing almost 40% of their initial PCE in 200 h. In the present work, we also used EH-IDTBR [41] as an alternative electron-acceptor with PF2. EH-IDTBR is one of the most reported small-molecule NFA, owing to its high electron affinity contributing to the high open-circuit voltage (V_oc) [42]. Moreover, EH-IDTBR has improved the domain crystallinity in ternary blends [PTB7-TH:F8IC:EH-IDTBR] reducing the energetic disorder and leading to a high PCE of 12.3% [43]. The stability of blends using EH-IDTBR as electron-acceptor is its most interesting property. It was shown that OSCs based on EH-IDTBR in blend with a low-bandgap donor polymer (PffBT4T-2DT) under continuous white light illumination had more than 90% of their initial PCE after 4000h [44]. Preliminary stability tests were also conducted on [PF2:EH-IDTBR] devices under continuous AM 1.5 illumination. These devices had a lower PCE around 6%. Encapsulated [PF2:EH-IDTBR] devices showed a relative loss in efficiency lower than 10% after 300 h continuous illumination [45].

In the present work, we decided to explore the properties of PF2/PCBM/EH-IDTBR ternary blends with the goal to identify the best compromise in terms of the high PCE obtained on [PF2:PC₇₁BM] devices and the promising stability observed on [PF2:EH-IDTBR] devices. The stability was estimated using accelerated photo-degradation on non-encapsulated devices in an inert atmosphere. Several binary ([PF2:PC₇₁BM] and [PF2:EH-IDTBR]) and ternary [PF2:PC₇₁BM:EH-IDTBR] blends were tested. The photovoltaic performances as a function of degradation time were measured. Fresh and aged samples were studied as a function of light-intensity to highlight the main charge-carrier recombination mechanism in each case. Finally, the fresh and aged samples’ morphology were investigated by atomic force microscopy (AFM) to correlate the morphology evolution with the performance degradation.