High Yield Synthesis of Water‐Processable Donor:Acceptor Janus Nanoparticles with Tuned Internal Morphology and Highly Efficient Charge Separation/Transfer

Here, for the first time, the high‐yield (87.6%) robust and facile synthesis of water‐processable donor–acceptor Janus nanoparticles (JNP) that are of high potential for optoelectronic applications is reported. The water‐processable JNPs have easily controlled Janus ratios and are of excellent quality, which is shown by energy‐filtered transmission electron microscopy. The JNPs exhibit improved charge separation and transfer properties compared to the conventional donor–acceptor nanoparticles which is characterized via both steady‐state and transient photoluminescence spectroscopy. The Janus character of particles allows the combination of two materials into one composite and programs morphology of structures, which can be formed on the basis of the particles. Finally, outstanding performance of JNP‐based photovoltaic cells with 53% improvement of efficiency is shown.


Introduction
In recent years, organic semiconductor nanoparticles dispersed in aqueous solvents, have attracted significant interest given that they are environmentally friendly and amenable for largescale manufacturing. [1] The organic donor:acceptor nanoparticles (D:A NPs) prepared recently by Kosco et al. show significantly enhanced photocatalytic activity in the hydrogen evolution reac-scale. However, there is little contrast for typical materials of photoactive layers through normal bright-field mode. [11] The energy-filtered TEM (EFTEM) technique, utilizing the elemental difference in a multi-component system, seems to be a good choice to provide material contrast in many situations. It has been successfully applied to create spectroscopic contrast between D:A in BHJ structure as well as D:A NPs. [11,12] This is the first study on the rational design of high-quality D:A Janus NPs (JNPs) with controlled Janus ratios. In this work, we address the challenges in the controlled design of D:A NPs with tunable internal structures and distribution between donor and acceptor parts. It ensures excitons reach the D:A interface and achieve efficient dissociation. Furthermore, EFTEM and photoluminescence spectroscopy are carried out to verify the internal morphology and excellent charge separation properties of D:A NPs. Finally, we correlate the internal morphological studies with the corresponding NP-based polymer solar cells (PSCs) performance and compare their properties to reference devices processed from halogenated solvents.
To prepare PTB7:PC 71 BM NPs, the emulsion of chloroform droplets containing PTB7 and PC 71 BM were first dispersed in an aqueous solution of sodium dodecyl sulfate (SDS) The fullerene moieties tend to merge in a continuous phase during chloroform evaporation. As a result, the droplets transformed to PC 71 BM-core:PTB7-shell (PC 71 BM-c:PTB7-s) structure when chloroform evaporated completely ( Figure S3, Supporting Information). This fullerene-core and polymer-shell structure is common in the mini-emulsion method. Because the polymer phase (PTB7) tends to minimize interfacial tension between the continuous (water) and disperse phases (PC 71 BM) of the emulsion when the organic solvent (chloroform) is evaporated. [2] Therefore, the polymer PTB7 would separate from the surface of the chloroform droplets, eventually forming NPs consisting of a PC 71 BM core and PTB7 shell.
In fact, it is challenging to determine the structure of PTB7:PC 71 BM nanoparticles from bright-field TEM images ( Figure 1c) due to the lack of contrast in polymeric materials. To explore their internal morphologies, the obtained NPs are investigated by energy-filtered TEM (EFTEM). This technique is based on detecting low atomic number elemental differences in the chemical structures. [14] Sulfur is present in PTB7 while not in PC 71 BM (Figure 1a and Figures S1 and S2, Supporting Information), thus the contrast in the elemental sulfur map corresponds to donor material. The obvious contrast between donor and acceptor domains is shown through selectively recording energy losses of the desired elements. This powerful EFTEM technique reveals that the PC 71 BM-rich core is surrounded by PTB7 shell in PC 71 BM-c:PTB7-s NP (Figure 1d) The acceptor core and donor shell exhibit obvious phase separation in this core-shell structure.
To the best of our knowledge, there is no other report on the design of high-yield Janus structures of D:A NPs. Most of the previous publications focused on designing D:A core-shell structures. However, the core-shell D:A NPs must undergo the annealing process to break the shell to generate interconnecting pathways. [7b] It is expected that the Janus structures show a certain phase separation of donor and acceptor. This means that donor and acceptor parts can generate a continuous phase to exhibit higher efficiency in organic photovoltaics.
We aim to convert the obtained core-shell structure into Janus structure using the different solubility properties between donor and acceptor materials. PC 71 BM has a quite high solubility in N-methyl-2-pyrrolidone (NMP) (>21 mg mL −1 ), while PTB7 can hardly dissolve (<0.1 mg mL −1 ) (Table S1, Supporting Information). [15] The prepared core-shell structure converts into the hollow structure when a certain amount of NMP (20 vol%) is involved (Figure 1e). In general, a gradually increased pressure will be created inside the shell upon an explosion to the NMP/water mixture. The pressure becomes high enough during a continuous stirring, the spherical symmetry will break while pushing out the PC 71 BM through the opening hole. It is consistent with the removal of PC 71 BM core by annealing. [7d] In detail, the PC 71 BM core is dissolved and escapes from one side of the nanoparticle, while the PTB7 collapses thus forming the hollow or half-moon structure. A unique large hole with a cavity inside can be clearly observed in the shell of each particle from scanning electron microscope (SEM) images ( Figure 1e). For SEM measurement, samples were prepared by applying 10 µL aqueous dispersion of nanoparticles (1 mg mL −1 ) on silicon wafers. NMP has a quite high boiling point (204.27 °C) and very low vapor pressure (0.29 Torr, 20 °C), [16] that means this solvent is difficult to evaporate (Table S1, Supporting Information). Therefore, a dialysis procedure to remove involved NMP is applied due to its miscibility with water ( Figure 1b). Interestingly, the hollow NPs convert back into the spherical morphology after dialysis, to reach the lowest energy configuration (Figure 1f,h). [17] It is concluded that the PC 71 BM core is not completely dissolved in 20% NMP/water mixture and part of PC 71 BM is still in the cavity hole.
Eventually, the PC 71 BM (previously dissolved in NMP) turns into spheres to reduce their surface area. [18] The PC 71 BM sphere fills into the vacancy part and thus a Janus structure is generated (Figure 1b,g). The domain differences of PTB7 and PC 71 BM is most pronounced in PTB7:PC 71 BM JNPs ( Figure 1g). Corresponding elemental mapping confirm that the donor and acceptor are well phase-separated in the nanoparticles after NMP removal (Figure 1g and Figure S4, Supporting Information). The EFTEM results show the core-shell and Janus structures of nanoparticles (Figure 1d,g).
Thus, the internal morphology of PTB7:PC 71 BM nanoparticles can be converted from core-shell to Janus structures by selective dissolution of PCBM by NMP. The obtained core-shell NPs show a high yield (82.8%) resulting from the mini-emulsion process, according to the statistical analysis of EFTEM images (Figure 2a). Moreover, these EFTEM images indicate that the PTB7 shell thickness is ≈12.1 ± 3.6 nm, whereas the diameters of individual cores are around 36.2 ± 8.3 nm ( Figure 2a).
Remarkably, the JNPs show a significantly high yield (87.6%) (Figure 2b). More EFTEM images ( Figure S4, Supporting Information) and large-area SEM images ( Figure S5, Supporting Information) give further supports for their high quality and yield. In addition, these JNPs remain highly stable even after four months of storage at 4 °C.  (Table S2, Supporting Information).
Next, the internal structures of donor and acceptor domains of the NPs were studied by UV-vis spectrophotometry. [8b] For UV-vis absorption, samples were measured in quartz cuvettes, containing 1 mL aqueous dispersion of nanoparticles (0.2 mg mL −1 ). The single-component NPs show characteristic peaks of PC 71 BM at the wavelength of 270 nm, whereas a characteristic peak of PTB7 is at 607 nm ( Figure 2d). The absorption spectrum of PTB7 is in good agreement with values from the literature. [19] Importantly, the characteristic peaks of PTB7:PC 71 BM NPs exhibit a redshift from 607 to around 620 nm. This redshift of polymer donor could result from interchain interactions as well as planarization of the polymer chains. [8a] It could be attributed to the formation of aggregates, that is further confirmed by corresponding SEM and bright-field TEM results ( Figure S9, Supporting Information). Furthermore, the absorption spectra between PC 71 BM-c:PTB7-s NPs and PTB7:PC 71 BM JNPs show that there is almost no change in absorption features. The strongest characteristic absorbance peak is observed at 270 nm due to fullerene. In addition, the characteristic peak at 620 nm with a shoulder at 670 nm is ascribed to the semicrystalline PTB7 domains.
Then the photoluminescence (PL) quenching studies are performed on our nanoparticles, which aims to investigate their charge separation behavior (Figure 2e). [7c] The samples (1 mL, 1 mg mL −1 of PTB7) were measured in quartz cuvettes in PL characterization. The emission peak of pure PTB7 NPs is at 805 nm. Excellent PL quenching of PTB7 is observed for all structures of D:A NPs. It is ascribed to the formation of an interface between donor and acceptor. Next, the quenching effect between PC 71 BM:PTB7 NPs in different structures is compared to evaluate charge separation properties. The PTB7:PC 71 BM JNPs show a more obvious PL quenching compared to PC 71 BMc:PTB7-s NPs, indicating a better charge separation behavior (inset figure in Figure 2e).
To further investigate the charge transfer properties within our D:A NPs, time-correlated single-photon counting (TCSPC) measurements are performed among different NPs (Figure 2f). For TCSPC measurement, the samples (0.2 mL, 1 mg mL −1 of PTB7) were characterized in eight-well chambered cover glass. All PL decay curves are well fitted to a mono-exponential decay function (Table S3, Supporting Information). It implies the existence of a simple radiative channel in a two-level system, characterized by a single decay rate. [20] For pure PTB7 NPs, the calculated PL lifetime (τ) is 0.86 ns. As expected, the decay time decreases upon involving PC 71 BM in NPs (Figure 2f). This indicates the diminishing of the photoluminescent quantum efficiency (PLQE) by 46.5% through an effective and fast charge transfer due to the presence of PC 71 BM in the NPs. [21] In detail, the PTB7:PC 71 Table S3, Supporting Information). This result suggests the efficient charge transfer is related to the internal structure of nanoparticles. [22] According to previous publications, the electron and hole mobilities strongly depend on the acceptor fraction in D:A blends. [23] Therefore, tuning the D:A ratio is very important to optimize the performance of organic solar cells. [24] In this work, we study various D:A mass ratios (1:1, 2:1, and 3:1) of Janus nanoparticles, in different lobe sizes, and characterize them through EFTEM. (Figure 3a-c).
The PC 71 BM domain is located at the edge of every nanoparticle. In addition, the size of the fullerene part reduces when less amount of fullerene is used (Figure 3a-c). It means that our strategy is demonstrated to fabricate high-yield (87.6%) Janus nanoparticles. As a unique advantage of this procedure, JNPs can be prepared with widely different mass fractions of donor and acceptor and thus control the Janus balance. The PTB7:PC 71 BM NP morphological differences are attributed to solubility differences as well as to the mass ratios of donor and acceptor. We can control these properties to prepare various internal morphologies on the nanoscale, to exhibit different properties.
Here, we also obtain the zeta-potential curves versus pH of PTB7:PC 71 BM JNPs among different Janus ratios ( Figure S10, Supporting Information). These curves of JNPs trend close to that of PTB7 NPs with an increase of PTB7 ratios. The same tendency is confirmed from obtained IEP results (Table S2, Supporting Information). It indicates that the increase of PTB7 ratios leads to a larger IEP.
To ensure a scientific comparison of PL quenching between PTB7:PC 71 BM JNPs in different ratios (1:1, 2:1, and 3:1), we control the same concentration of PTB7, which is confirmed by corresponding UV-vis absorption results ( Figure S11, Supporting Information). The characteristic peak (at 620 nm) as well as the shoulder (670 nm) of PTB7 exhibit the same intensity among different nanoparticles. On the other hand, the characteristic band of PC 71 BM shows higher intensity when its mass ratio increases. Afterwards, we investigate the PL behavior between the JNPs and a series of mixtures of pure NPs with various D/A mass ratios (1:1, 2:1, and 3:1) (Figure 3d-f). The emission intensity of the PTB7:PC 71 BM JNPs is dramatically reduced compared to that of mixed pure donor and acceptor NPs. This result suggests incomplete charge generation and separation in a simple mixed system.
The Janus structure exhibits efficient exciton dissociation, hence quenching the PL emission, even when the PTB7:PC 71 BM Janus ratio is 3:1 (Figure 3f). The intensity of emission in JNPs decreases slightly as the PC 71 BM mass ratio increases ( Figure S12  Supporting Information). Moreover, for these two kinds of D:A NPs (PC 71 BM-c:PTB7-s NPs and PTB7:PC 71 BM JNPs), the emission peak exhibits a redshift compared to pure PTB7 NPs, from 805 to 824 nm, even though they have the similar particle size ( Figure S8, Supporting Information). It can be attributed to the radiative recombination of charge transfer states. [25] The TCSPC delay of JNPs and mixed NPs is exhibited in Figure 3g-i. We notice that JNPs exhibit a much faster decay rate than blended donor NPs and acceptor NPs (Figure 3g-i and Table S3, Supporting Information). This change in PL lifetime is in accordance with the hypothesis of a more efficient charge transfer, [26] signifying a potent electron transfer from the polymer toward the fullerene acceptor in the Janus structure. The latter is due to a higher D/A interface in the Janus structure. Thus, the PTB7:PC 71 BM JNPs exhibit a superior charge separation behavior, which is beneficial for optoelectronic applications.
Finally, to study the effect of the internal structure of designed NPs in applications, we fabricate polymer solar cells (PSCs) using an inverted device architecture comprising indium tin oxide (ITO)/zinc oxide (ZnO)/ PTB7:PC 71 BM NPs/ molybdenum trioxide (MoO 3 )/silver (Ag) anode (Figure 4a). their performances are studied under illumination at AM 1.5G (100 mW cm −2 ). The Janus ratio 1:1 is applied in this system due to its best photoluminescence properties. The BHJ structured solar cells, formed by spin-coating from either chloroform or chlorobenzene solutions, were prepared as reference. The current density-voltage (J-V) curves of the NPs and BHJ structured PSCs are depicted (Figure 4b) with the key performance parameters (Table S4, Supporting Information). It can be seen that the PTB7:PC 71 BM NP based devices showed much higher PCE values than BHJ structured PSCs processed from chloroform (0.57%), but still lower than devices processed from more toxic solvent chlorobenzene (Figure 4b and Table S4, Supporting Information). Importantly, PTB7:PC 71 BM Janus NP devices yield much higher PCE values (2.03%) than the core-shell structured NP cell (1.43%). There is a 53% enhancement, through turning D:A NPs from core-shell into the Janus structures. Among device parameters, J SC of JNPs has been Adv. Optical Mater. 2022, 10, 2101922  PL lifetime imaging was performed to characterize the charge carrier properties of these different structured active layers on a micrometer scale (25 µm × 25 µm) (Figure 4c,d). The chlorobenzene processed film shows the fastest PL decays among these four active layers, followed by the Janus and core-shell NP films. The D:A film processed from chloroform has the longest PL lifetime, resulting from increased exciton recombination. [27] The green region with a short PL lifetime for the chlorobenzene processed BHJ film transits to a red region with a much longer PL lifetime with chloroform. The JNP film shows a larger area in green than that of core-shell NP film, indicating a faster decay rate on average. This indicates more efficient charge transfer in JNP films, leading to a higher current density and thus better device performance. The following TR-PL spectra further support this conclusion (Figure 4d and Table S5, Supporting Information).

Conclusions
To summarize this work, we presented a facile and robust route for the synthesis of D:A water-processable nanoparticles with tunable internal morphologies, efficient charge separation, and transfer properties. Internal patchy morphologies could be easily varied between core-shell as well as Janus structures and different Janus ratios. The Janus balance can be easily controlled by synthetically adjusting the D:A mass ratios. To our best knowledge, this is the first report on the fabrication of the D:A JNPs with a remarkably high yield (87.6%) and excellent quality.
The various internal structures of PTB7:PC 71 BM NPs dramatically improve the charge separation and transfer limitation in D:A dispersions system, verified by PL lifetime and TR-PL decay studies. These PTB7:PC 71 BM JNPs exhibit the dramatic PL quenching of donors and fast PL decay rates even when the D:A ratio is as high as 3:1. Credited to such excellent charge separation and transfer properties, the JNP-based devices yield PCE 53% that is higher than traditional core-shell NPs. Thus, creating a special patchy morphology with anisotropic Janus topology, that is different from the conventional core-shell design, might allow for a formation of a sharp interface between donor and acceptor domains. In this way, the charge separation and charge transfer properties in single D:A nanoparticles and their assembled structures can be well controlled. Our proposed approach is a promising new strategy to access water-processable JNPs and provides more profound knowledge about the design of high-performance NPs-based PSCs for future optoelectronic applications.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.