The interplay between nanomorphology and efficiency of polymer-fullerene bulk-heterojunction (BHJ) solar cells has been the subject of intense research, but the generality of these concepts for small-molecule (SM) BHJs remains unclear. Here, the relation between performance, charge generation, recombination, and extraction dynamics and nanomorphology, achievable with two analogous benzo[1,2-b:4,5-b]dithiophene-pyrido[3,4-b]-pyrazine SM donors (BDT(PPTh2)2, namely SM1 and SM2, differing by their side-chain substitution pattern, are examined as a function of solution additive composition. The results show that 1,8-diiodooctane (DIO), used as a processing additive acts as a plasticizer in the blends, increases domain size, and promotes ordering/crystallinity. Surprisingly, the system with high domain purity (SM1) exhibits both poor exciton harvesting and severe charge trapping, alleviated only slightly with increased crystallinity. In contrast, the system consisting of mixed domains and lower crystallinity (SM2) shows both excellent exciton harvesting and low charge recombination losses. Importantly, the onset of large, highly-ordered pure crystallites in the latter (SM2) system reduces solar cell efficiency, pointing to possible differences in the ideal morphologies for SM-based BHJ solar cells compared with polymer-fullerene devices. In polymer-based systems, tie chains between pure polymer crystals establish a continuous charge transport network, whereas SM-based active layers may in some cases require mixed domains at concentrations that enable both small molecule aggregation and charge percolation to the electrodes.

As organic photovoltaic performance approaches 20% efficiencies, causal structure–performance relationships must be established for devices to realize theoretical limits and become commercially competitive. Here, we reveal evidence of a causal relationship between mixed donor–acceptor interfaces and charge generation in polymer–fullerene solar cells. To do this, we combine a holistic loss analysis of device performance with quantitative synchrotron X-ray nanocharacterization to identify a >98% anticorrelation between field-dependent geminate recombination and nanodomain purity. Importantly, our analysis eliminates other possible explanations of the performance trends, a requirement to establish causality. The unprecedented granular level of our analysis also separates field-dependent and field-independent recombination at the interface, where we find for the first time that this system is free of field-independent recombination, a loss channel that plagues high-performance systems, including those with non-fullerene acceptors. This result broadens the case that minimizing mixed phases to promote sharp interfaces between pure aggregated domains is the ideal nanostructure for realizing theoretical efficiency limits of organic photovoltaics.

Although solvent additives are used to optimize device performance in many novel non‐fullerene acceptor (NFA) organic solar cells (OSCs), the effect of processing additives on OSC structures and functionalities can be difficult to predict. Here, two polymer‐NFA OSCs with highly sensitive device performance and morphology to the most prevalent solvent additive chloronaphthalene (CN) are presented. Devices with 1% CN additive are found to nearly double device efficiencies to 10%. However, additive concentrations even slightly above optimum significantly hinder device performance due to formation of undesirable morphologies. A comprehensive analysis of device nanostructure shows that CN is critical to increasing crystallinity and optimizing phase separation up to the optimal concentration for suppressing charge recombination and maximizing performance. Here, domain purity and crystallinity are highly correlated with photocurrent and fill factors. However, this effect is in competition with uncontrolled crystallization of NFAs that occur at CN concentrations slightly above optimal. This study highlights how slight variations of solvent additives can impart detrimental effects to morphology and device performance of NFA OSCs. Therefore, successful scale‐up processing of NFA‐based OSCs will require extreme formulation control, a tuned NFA structure that resists runaway crystallization, or alternative methods such as additive‐free fabrication.