Alkyl-Side-Chain Engineering of Nonfused Nonfullerene Acceptors with Simultaneously Improved Material Solubility and Device Performance for Organic Solar Cells.

Two nonfullerene small molecules, TBTT-BORH and TBTT-ORH, which have the same thiophene-benzothiadiazole-thiophene (TBTT) core flanked with butyloctyl (BO)- and octyl (O)-substituted rhodanines (RHs) at both ends, respectively, are developed as electron acceptors for organic solar cells (OSCs). The difference between the alkyl groups introduced into TBTT-BORH and TBTT-ORH strongly influence the intermolecular aggregation in the film state. Differential scanning calorimetry and UV-vis absorption studies reveal that TBTT-ORH exhibited stronger molecular aggregation behavior than TBTT-BORH. On the contrary, the material solubility is greatly improved by the introduction of a BO group in TBTT-BORH, and the inevitably low molecular interaction and packing ability of the as-cast TBTT-BORH film can be effectively increased by a solvent-vapor annealing (SVA) treatment. OSCs based on the two acceptors and PTB7-Th as a polymer donor are fabricated owing to their complementary absorption and sufficient energy-level offsets. The best power conversion efficiency of 8.33% is obtained with the SVA-treated TBTT-BORH device, where, together with a high open-circuit voltage of 1.02 V, the charge-carrier mobility and the short-circuit current density were greatly improved by the SVA treatment to levels comparable to those of the TBTT-ORH device because of the suppressed charge recombination and improved film morphology. In this work, the simultaneous improvement of both material solubility and device performance is achieved through alkyl side-chain engineering to balance the trade-offs among material solubility/crystallinity/device performance.

Introduction

Organic solar cells (OSCs) have attracted intensive attention and have been rapidly developed because of their various advantages, which include being low cost, lightweight, and mechanically flexible. Recently, nonfullerene small molecules (NFSMs) have been widely used as a promising alternative to fullerene-derivative electron acceptors in OSCs. Most of the NFSMs consist of extended ladder-type fused-ring cores, and high-efficiency OSCs based on nonfused NFSMs, which can be relatively easily synthesized, have also been reported.[1−5] The use of NFSMs can not only overcome some of the critical drawbacks of fullerene derivatives, which include weak absorption of visible light, difficult synthesis, high cost, and strong self-aggregation, but also improve the power conversion efficiency (PCE) of the corresponding devices to greater than 15%.[6−8] The physical properties of the acceptor (A)–donor (D)–acceptor (A)-type NFSMs, including their UV–vis absorption properties, energy levels, and packing behavior, can be easily controlled through (i) the selection of specific combinations of various electron D and A units,[9−12] (ii) the introduction of electron-withdrawing (e.g., F and Cl) or electron-donating (e.g., OR) substituents,[13−17] and (iii) the side-chain engineering at both the core and end of the molecules.[18−21]

Alkyl chain engineering (e.g., changing of the length, position, and/or bulkiness of introduced side chains) is considered a prerequisite for optimizing the chemical structure of the molecules because it can impart them with sufficient solubility to enable the fabrication of the devices by solution processing while also providing control over the crystallinity and packing behavior of the molecules, which affect the charge-carrier mobility and the ultimate device performance.[22−37] In particular, alkyl variations in the same backbone enable control of intermolecular aggregation and film morphology while maintaining other physical properties such as optical and electrochemical properties. In general, the interdigitation of linear n-alkyl side chains endows polymers or small molecules with a highly crystalline morphology,[38,39] which is advantageous for charge transport in devices. However, strong intermolecular interactions induced by such highly ordered structures increase the phase-transition temperature and lower the solubility, making such materials unsuitable for solution processing (especially large-area or eco-friendly processes for future commercialization). Therefore, optimizing the chemical structure to exhibit both adequate solubility for solution processing and suitable crystallinity to enable effective charge transfer between molecules by balancing the trade-off between material solubility and crystallinity is important. For example, Qu et al. have reported a series of A–D–A-type indacenodithieno[3,2-b]thiophene-2-(1,1-dicyanomethylene)rhodanine-based NFSMs (IDBTR-C2, IDBTR-C4, IDBTR-C6, and IDBTR-C8) with n-alkyl chains ranging in length from ethyl to butyl, hexyl, and octyl chains on both ends.[40] Their four NFSMs exhibited similar optical and electrochemical properties, but the film morphologies differed depending on the introduced alkyl chain lengths. In their work, IDBTR-C6 (with hexyl chains) exhibited the best PCE of 9.29% because of its suitable crystallinity and domain size in the blend film. Instead of controlling the length of the n-alkyl side chains, researchers have introduced bulky branched side chains such as 2-ethylhexyl or 2-butyloctyl (BO) groups into the molecular backbone to achieve sufficient solubility by suppressing strong aggregation.[41] However, because the bulky side chains, which were introduced to increase material solubility and provide a uniform film morphology, may inevitably reduce molecular aggregation, crystallinity, and intermolecular charge transfer,[42−44] each side chain for each given molecular backbone should be optimized to balance solubility and crystallinity.

Benzo[1,2,5]thiadiazole (BT) has been widely used as an A unit of organic semiconductors because of its strong electron-withdrawing ability.[45] For both fullerene- and nonfullerene-based OSCs, D–A-type polymers containing BT or its derivatives are regarded as one of the most promising donor materials; their low bandgaps and high crystallinity have resulted in PCEs as high as 11%.[46−49] The combination of BT-containing small-molecule donors and fullerene acceptors has also led to high-performance OSCs. Bazan and coworkers reported a series of dithienosilole–thiophene–BT-based small-molecule donors in which a small change in the molecular structure (e.g., a change in regiochemistry) strongly influenced the film morphology and the OSC performance.[50,51] Consistent with recent trends, various nonfullerene acceptors based on BT units also have been developed, resulting in high-efficiency devices. In most cases, the BT groups were used as a π-bridge to connect the electron-rich core and the electron-withdrawing end groups in the A–D–A-type acceptors;[52−55] a few examples of acceptors with BT core groups have been reported.[56−61]

In the present work, a series of BT-based NFSMs, TBTT-BORH and TBTT-ORH, were synthesized as acceptors for solution-processed OSCs. Both molecules are composed of a nonfused thiophene–BT–thiophene (TBTT) core and alkyl-substituted rhodanine (RH) ends. To examine the effect of introducing alkyl groups onto the same backbone, we replaced the n-octyl chain of TBTT-ORH with the bulky side chain of the BO group, resulting in TBTT-BORH. The physical properties of the synthesized small molecules were systematically studied by thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), UV–vis spectrophotometry, cyclic voltammetry (CV), atomic force microscopy (AFM), and transmission electron microscopy (TEM), and OSC devices were fabricated using them as acceptors with a PTB7-Th polymer donor. As expected, the introduction of the bulky side chains of the BO group resulted in a relatively low phase-transition temperature and low crystallinity of the TBTT-BORH, consistent with a relatively smaller short-circuit current density (JSC) in the device than in the corresponding device with TBTT-ORH. More importantly, however, the inevitably lowered molecular aggregation and JSC in the TBTT-BORH devices could be effectively improved by a solvent-vapor annealing (SVA) treatment to levels comparable to the aggregation and JSC in TBTT-ORH, resulting in substantial improvements in the JSC and fill factor (FF) values. Therefore, the SVA-treated TBTT-BORH device with a high open-circuit voltage (VOC) of 1.02 V exhibited relatively a better PCE of 8.33% than the SVA-treated TBTT-ORH device.

Results and Discussion

Materials Synthesis and Physical Properties

Two NFSMs, TBTT-BORH and TBTT-ORH, were synthesized with a TBTT core and alkyl-substituted RH ends. TBTT-BORH and TBTT-ORH have different terminal alkyl chains of BO and octyl groups, respectively, on both RH ends. The alkyl side chains were introduced in a direction parallel to the TBTT backbone, resulting in rod-shaped small molecules. The two alkylrhodanines, BORH and ORH, were obtained through the ring-closing reaction of bis(carboxymethyl)trithiocarbonate and the corresponding alkylamines (RNH2) (i.e., 2-butyloctylamine (BONH2) and n-octylamine, respectively) via the procedure described in our previous report.[3] The n-octylamine was purchased for US$3 per 10 g, and the BONH2 was synthesized via a two-step Gabriel reaction from 2-butyloctanol, which was commercially available for $1 per g. The intermediate of B-T-BORH was synthesized by Knoevenagel condensation of B-T-CHO and BORH; the Pd-catalyzed Suzuki coupling reaction between the boronic ester of B-T-BORH (2 equiv) and the dibromide of BT (1 equiv) yielded TBTT-BORH. TBTT-ORH was prepared using the same methods. The chemical structures of the two NFSMs and their synthesis routes are described in Scheme . Both molecules are readily soluble in common organic solvents; TBTT-BORH exhibits better material solubility than TBTT-ORH because of its bulky BO-group side chains. Details of the synthesis and characterization, including the 1H and 13C NMR spectra and elemental analysis results, are provided in the Supporting Information (Figures S1–S6).

Scheme 1

Synthetic Route of TBTT-BORH and TBTT-ORH

The thermal characteristics of the two small molecules were studied by TGA and DSC, as shown in Figure . Both molecules exhibited excellent thermal stability, losing less than 5% of their weight when heated to 370 °C (Figure a). The DSC thermogram (Figure b) clearly shows that the phase-transition temperatures of TBTT-BORH were lower than those of TBTT-ORH. The melting temperatures (Tm) of TBTT-BORH and TBTT-ORH were 183 and 220 °C, respectively. The crystallization temperatures (TCr) of TBTT-BORH and TBTT-ORH were 111 and 174 °C, respectively. The introduction of the bulky side chain of the BO group lowered the crystallinity of the molecule, as indicated by an increase in the solubility of TBTT-BORH compared with that of TBTT-ORH. The introduction of branched alkyl side chains has been widely reported to reduce intermolecular interaction and thereby increase solubility.[62] The thermal properties of the two small molecules are summarized in Table .

Figure 1

(a) TGA and (b) DSC curves of TBTT-BORH and TBTT-ORH acceptors.

Table 1

Thermal and Optical Properties of TBTT-BORH and TBTT-ORH

	Tda [°C]	Tmb [°C]	Tcrystc [°C]	abs. peaksd [nm]		λonsete [nm]	Egoptf [eV]
				soln	film
TBTT-BORH	374	159,183	111	432, 540	575, 626	676	1.83
TBTT-ORH	373	220	174	432, 540	585, 635	693	1.79

Temperature resulting in 5% weight loss based on the initial weight.

Temperature at the melting endothermic peak.

Temperature at the recrystallization exothermic peak.

Absorption peaks measured for samples in chloroform solution and in the film state; λmax is underlined.

Absorption onset of the films.

Eg,opt = 1240/λonset.

Optical and Electrochemical Properties of the Small Molecules

Figure a shows the UV–vis absorption spectra of the two molecules in chloroform (CF) solution and in the film state. Irrespective of the alkyl side chains, TBTT-BORH and TBTT-ORH show the same UV–vis absorption characteristics in solution, with two main absorption bands. The high-energy band centered at 432 nm is attributed to π–π* transitions, and the low-energy band with an absorption maximum (λmax) at 540 nm originates from the intramolecular charge transfer between the D and A units.[63−65] The film samples were prepared by spin-coating a CF solution. The absorption spectra of both films were red-shifted compared with those of the solutions, indicating J-type aggregation of the molecules in the films. The UV–vis absorption of the TBTT-ORH film was more red-shifted than that of TBTT-BORH by approximately 10 nm, suggesting stronger aggregation behavior in the TBTT-ORH film than in the TBTT-BORH film. In addition, the low-energy absorption bands of the films displayed a more resolved double-peak vibronic structure, assigned as 0–1 and 0–0 transitions (e.g., 575 and 626 nm for TBTT-BORH and 585 and 635 nm for TBTT-ORH, respectively). The ratio between the intensity of the 0–0 peak and that of the 0–1 peak (A0–0/A0–1) was higher for TBTT-ORH (A0–0/A0–1 = 1.05) than for TBTT-BORH (A0–0/A0–1 = 0.98). The stronger absorption of the lower-energy (0–0) band for TBTT-ORH suggests more effective π–π stacking between molecular backbones.[66−69] This trend well matches the higher phase-transition temperatures observed in the DSC thermograms of TBTT-ORH, and the greater crystallinity and more efficient molecular packing could be beneficial for achieving a high JSC in OSC devices. A slightly red-shifted UV–vis absorption of TBTT-ORH yielded a relatively lower optical energy bandgap (Eg,opt) of 1.79 eV compared with 1.83 eV for TBTT-BORH. The optical properties of the two NFSMs are summarized in Table .

Figure 2

(a) UV–vis absorption spectra, (b) cyclic voltammograms, and (c) energy-level diagrams of the acceptors TBTT-BORH and TBTT-ORH. The voltammogram and energy diagram of the PTB7-Th donor are included for comparison.

The effect of annealing on the optical properties of the films was also studied, as shown in Figure S7a and b for TBTT-BORH and TBTT-ORH, respectively. For both molecules, two neat films were prepared by spin-coating under the same conditions: one was used without annealing (as-cast) and the other was solvent-vapor annealed using a procedure similar to that used for device fabrication. No substantial annealing-induced wavelength shifts were observed for the film of either material; however, increases in the absorption intensities were observed in the annealed films, and the magnitude of this increase was found to be larger in the TBTT-BORH film than in the TBTT-ORH film. In particular, the lowest-energy (0–0) band for the TBTT-BORH film (at 626 nm) was more enhanced; the A0–0/A0–1 ratio increased from 0.98 to 1.13 after annealing (Figure S7a). By contrast, the intensity ratio of the TBTT-ORH film remained similar before and after annealing (Figure S7b). That is, because of the strong aggregation behavior of the TBTT-ORH molecules, as confirmed by DSC analysis, a highly crystalline as-cast film could be obtained; such strong intermolecular aggregation and well-aligned crystalline morphology in the as-cast TBTT-ORH film were not further improved by an additional annealing process. By contrast, although the introduction of bulky side chains resulted in relatively weak intermolecular packing behavior of the TBTT-BORH molecules in the as-cast film, the annealing process greatly improved the intermolecular interaction and molecular packing among these molecules. Therefore, the material solubility of TBTT-BORH was effectively increased by appropriate side-chain engineering, rendering it more suitable for solution processing. At the same time, the crystallinity of the film, which usually decreases with increasing solubility, could be effectively improved through an appropriate annealing process. Thus, both good material solubility and high crystallinity could be achieved through proper modification of the molecular chemical structure. The increase in crystallinity induced by the annealing process is related to the domain size of the blend films and to the charge mobilities of the devices and plays an important role in the device performance, as described later.

The electrochemical properties of the small molecules were characterized by CV; the results are shown in Figure b. The molecular orbital energy levels of the two small molecules were found to be similar irrespective of the introduced alkyl side chains. The highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) energy levels of TBTT-BORH were calculated to be −5.76 and −3.58 eV, respectively. Similarly, the calculated HOMO and LUMO energy levels of TBTT-ORH were −5.74 and −3.60 eV, respectively. For comparison, the HOMO and LUMO energy levels of the previously reported bithiophene–RH-based rod-shaped acceptor, T2-ORH,[70] were determined at the same time and were found to be −5.67 and −3.60 eV, respectively. The introduction of the central BT group lowered the HOMO levels of the TBTT-based NFSMs relative to that of T2-ORH, whereas the LUMO levels of the three small molecules were approximately the same. The energy diagrams of TBTT-BORH and TBTT-ORH are shown in Figure c, together with that of the PTB7-Th donor. Notably, in our previous report, despite the existence of a relatively small LUMO energy offset between the PTB7-Th donor and the T2-ORH acceptor, D-to-A electron transfer effectively occurred in the PTB7-Th:T2-ORH device, resulting in a high JSC because of the contribution of both the donor and acceptor to the photocurrent of the devices.[3] The electrochemical properties of the acceptors are summarized in Table .

Table 2

Electrochemical Properties of TBTT-BORH and TBTT-ORH

	Eonset,oxa [V]	Eonset,reda [V]	EHOMO,CVb [eV]	ELUMO,CVb [eV]	Eg,CVc [eV]
TBTT-BORH	1.06	–1.12	–5.76	–3.58	2.17
TBTT-ORH	1.04	–1.10	–5.74	–3.60	2.14

Eonset,ox and Eonset,red are the onset potentials of oxidation and reduction, respectively, vs Ag/AgCl electrode.

Calculated using the empirical equations EHOMO,CV = −(Eonset,ox – E1/2,ferrocene + 4.8) eV and ELUMO,CV = −(Eonset,red – E1/2,ferrocene + 4.8) eV.

Eg,CV = ELUMO,CV – EHOMO,CV.

OSC Performance

The OSC devices were fabricated with an inverted configuration of ITO/ZnO NPs/PEIE/PTB7-Th:acceptor (1:2)/MoO/Ag, where ITO is indium tin oxide, ZnO NPs are zinc oxide nanoparticles, and PEIE is polyethylenimine (80% ethoxylated). CF was used as a processing solvent. The devices were subjected to SVA for further optimization of the device performance. A detailed explanation of the device fabrication procedure is provided in the Supporting Information. The current density–voltage (J–V) and external quantum efficiency (EQE) curves are shown in Figure , and the photovoltaic performances are summarized in Tables and S1.

Figure 3

(a) J–V curves and (b) EQE spectra of PTB7-Th:acceptor devices.

Table 3

Photovoltaic Properties of TBTT-BORH and TBTT-ORHa

acceptor	annealing	VOC [V]	JSC [mA/cm2]	FF [%]	PCE [%]	μhb [cm2/V s]	μec [cm2/V s]	μh/μe
TBTT-BORH	W/O	1.02	13.30 (12.77)d	44	5.97	2.08 × 10–5	5.34 × 10–6	3.9
	SVA	1.02	15.27 (14.66)d	54	8.33	7.41 × 10–5	3.22 × 10–5	2.3
TBTT-ORH	W/O	0.92	14.99 (14.39)d	45	6.21	5.00 × 10–5	1.73 × 10–5	2.9
	SVA	0.95	15.82 (15.19)d	51	7.60	1.40 × 10–4	5.12 × 10–5	2.7

Inverted device architecture is ITO/ZnO NPs/PEIE/PTBT-Th:acceptor (1.0:2.0, CF, d ≈ 100 nm)/MoO/Ag.

Hole-only device is ITO/PEDOT:PSS/PTBT-Th:acceptor (1.0:2.0, CF, d ≈ 100 nm)/Au.

Electron-only device is ITO/ZnO NPs/PEIE/PTBT-Th:acceptor (1.0:2.0, CF, d ≈ 100 nm)/LiF/Al.

JSC values calculated from the EQE spectra.

The as-cast devices based on TBTT-BORH and TBTT-ORH exhibited PCEs of 5.97 and 6.21%, respectively. The SVA treatment effectively increased the PCE to 8.33 and 7.60% for the TBTT-BORH and TBTT-ORH devices, respectively, where both the JSC and FF values of the devices were increased while maintaining the VOC values. For instance, the TBTT-BORH-based devices exhibited a VOC of 1.02 V irrespective of annealing and their JSC increased from 13.30 mA/cm2 before annealing to 15.27 mA/cm2 after annealing. The as-cast TBTT-ORH device exhibited a higher PCE than the as-cast TBTT-BORH device. However, for SVA-treated films, the TBTT-BORH device outperformed the TBTT-ORH device.

Both as-cast and SVA-treated TBTT-BORH devices exhibited a higher VOC value (1.02 V) than the corresponding TBTT-ORH devices (0.92–0.95 V). The higher VOC values for the TBTT-BORH devices are consistent with the somewhat higher-lying LUMO energy level of TBTT-BORH (−3.58 eV) compared with that of TBTT-ORH (−3.60 eV). Note that the difference in the VOC values of the devices (0.10–0.07 V) was found to be greater than the difference in the LUMO energy levels (0.02 eV), which can be explained by the strong aggregation behavior of TBTT-ORH relative to that of TBTT-BORH. The strong aggregation of molecules and corresponding enhancement of molecular–orbital interaction are known to destabilize the HOMO energy levels and stabilize the LUMO energy levels of the aggregates.[71] The LUMO level of the TBTT-ORH aggregate appears to be stabilized more than expected because of the relatively strong molecular interactions, which lowered the VOC of the devices based on TBTT-ORH compared with that of the devices based on TBTT-BORH for a given HOMO level of the donor, despite the insignificant difference in LUMO energy levels (0.02 eV) measured by CV.

As shown in Figure b, all of the devices based on TBTT-BORH and TBTT-ORH exhibited broad spectral responses from 350 to 800 nm. The complementary absorption of the PTB7-Th donor and two acceptors is desirable for harvesting more sunlight and increasing the JSC values in the devices.[72,73]Figure S8 compares the EQE curves of the PTB7-Th:acceptor devices and the UV–vis absorption spectra for the neat films of the components. Strong EQE responses were observed in the wavelength regions corresponding to the absorption of the donor and the acceptor. The EQE peaks in the regions 500–650 and 650–800 nm matched the absorption regions of the acceptor and donor, respectively. Additional strong responses were observed in the shorter-wavelength region from 400 to 500 nm in the EQE profiles, which corresponds to the absorption region attributed to the π–π* transitions in the neat acceptor films. Although both the neat acceptors and the blend films exhibited relatively weak absorption at ∼450 nm (Figure S7), the photocurrent contribution due to the absorption in these short-wavelength regions was found to be strong. The EQE responses at a wavelength of 450 nm were greater than 60% for all of the devices, leading to broad EQE profiles covering the wavelength range from 350 to 750 nm.

The similar absorption profiles of TBTT-BORH and TBTT-ORH resulted in similar corresponding EQE profiles; however, the EQE intensities and the JSC values calculated from the EQE curves (values in parenthesis in Table ) were higher in the TBTT-ORH devices than in the TBTT-BORH devices (Figures b and S8) for both the as-cast and SVA-treated devices. For example, the as-cast TBTT-ORH device exhibited a higher maximum EQE value (66% at 580 nm) and a higher calculated JSC value (14.39 mA/cm2) than the corresponding as-cast TBTT-BORH device (61% at 580 nm, 12.77 mA/cm2). This result can be explained by the higher crystallinity and better molecular packing behaviors of TBTT-ORH compared with TBTT-BORH, as demonstrated by the higher phase-transition temperatures and greater red-shifts in the film absorption spectra of TBTT-ORH.

For both acceptors, the EQE responses over the entire investigated wavelength range were increased by annealing, resulting in increased JSC values after annealing. Notably, the increase of the JSC by annealing was greater in the TBTT-BORH devices (15% increase) than in the TBTT-ORH devices (5.5% increase); thus, the difference between the JSC values of the TBTT-BORH and TBTT-ORH devices was reduced after the SVA treatment, although the EQE intensities and the calculated JSC values were still higher in the TBTT-ORH devices than in the TBTT-BORH devices. After the SVA treatment, the maximum EQE values reached 72% for both acceptors, and the difference in the calculated JSC values between the two acceptors (14.66 and 15.19 mA/cm2 for TBTT-BORH and TBTT-ORH, respectively) was not as great as in the case before SVA treatment (12.77 and 14.39 mA/cm2 for TBTT-BORH and TBTT-ORH, respectively). The EQE curves for the two acceptors are more or less overlapped after annealing, as shown in Figure S8b. The trend observed in the JSC values calculated from the EQE curves was consistent with the JSC values obtained from J–V curves (Table ). This result is in good agreement with the previously observed phenomenon in the absorption spectra, where the more pronounced vibronic 0–0 band observed in the spectrum of the SVA-treated TBTT-BORH film indicated greatly improved aggregation of the TBTT-BORH molecules after annealing. On the contrary, the as-cast TBTT-ORH film exhibited a highly crystalline morphology, and a further increase of molecular aggregation by annealing was not as pronounced as in the case of TBTT-BORH.

TBTT-BORH was more soluble in organic solvents than TBTT-ORH because of the introduction of the bulky BO side chain and relatively weak molecular interaction; however, such a weak intermolecular packing ability inevitably diminished charge transport by the hopping mechanism between aligned neighboring molecules, resulting in a lower JSC of the as-cast TBTT-BORH device. Therefore, in the as-cast condition, despite its relatively higher VOC of 1.02 V (see previous discussion), the TBTT-BORH device exhibited a lower PCE of 5.97% compared with the TBTT-ORH device (6.21%). However, this low crystallinity of TBTT-BORH could be sufficiently improved by the SVA annealing process, resulting in an enhancement of the JSC in the SVA-treated TBTT-BORH device, which was still lower than but similar to the JSC of the TBTT-ORH device. Given that a high VOC of 1.02 V was maintained, such an increase in the JSC resulted in a PCE of 8.33% for the TBTT-BORH device, which is even higher than the PCE for the TBTT-ORH device (7.60%). The large increase in the intermolecular interaction and device efficiency observed in the SVA-treated TBTT-BORH device (especially the JSC) is in good agreement with the improved film morphology and greater electron mobility of TBTT-BORH, which will be demonstrated later using space-charge-limited current (SCLC) charge carrier mobility measurements and AFM and TEM surface analyses.

The JSC and VOC values were measured under different light intensities (Plight) to investigate charge recombination of the as-cast and SVA-treated devices. Figure a shows a plot of the JSC dependence on Plight. In the formula JSC ∝ Plightα, the α value is equal to 1 when all free charge carriers are swept up and collected from the electrode and is less than 1 when bimolecular charge recombination is predominant in the devices. For both acceptors, although the as-cast devices exhibit α values of 0.87 and 0.90 for TBTT-BORH and TBTT-ORH, respectively, after SVA annealing, the α values for both devices became closer to 1 (0.97 and 0.94 for TBTT-BORH and TBTT-ORH, respectively); thus, the bimolecular recombination of the devices could be reduced by SVA annealing for both acceptors. In particular, the bimolecular recombination in the TBTT-BORH device, which was predominant under the as-cast condition (α = 0.87), was greatly reduced by annealing (α = 0.97), contributing to the increase of the JSC and FF values of the SVA-treated TBTT-BORH device.

Figure 4

(a) JSC and (b) VOC dependency on Plight for the devices based on TBTT-BORH (BO) and TBTT-ORH (O) in the as-cast (W/O) and SVA conditions.

The power-law dependence of VOC on Plight is shown in Figure b. In the formula VOC ∝ (nkT)/(q ln(Plight)), k is the Boltzmann constant, T is the absolute temperature, and q is the elementary charge. In the plots of VOC as a function of Plight, a slope close to 2 kT/q means the trap-assisted (monomolecular) recombination is dominant, whereas a slope close to 1 kT/q indicates that bimolecular recombination is dominant. In Figure b, for the TBTT-BORH and TBTT-ORH acceptors, the slopes decreased from 1.78 kT/q and 1.72 kT/q before annealing to 1.32 kT/q and 1.50 kT/q after annealing, respectively; that is, the slopes diverged from 2 kT/q, indicating that annealing suppressed monomolecular recombination in the devices. This result, in combination with the suppressed bimolecular recombination previously discussed, indicates that both bimolecular and monomolecular recombination could be effectively reduced by annealing for both acceptors. Notably, the relatively large (bimolecular and monomolecular) recombination observed in the as-cast TBTT-BORH device was effectively suppressed after annealing, consistent with the large increase of the JSC and FF in the SVA-treated TBTT-BORH device.

Hole- and electron-only devices were fabricated to investigate the SCLC charge-carrier mobilities. The configurations of the hole- and electron-only devices were ITO/PEDOT:PSS/active layer/Au and ITO/ZnO NPs/PEIE/active layer/LiF/Al, respectively (Figure ), where PEDOT:PSS is poly(3,4-ethylenedioxythiophene) polystyrene sulfonate. As summarized in Table , in the as-cast devices, the PTB7-Th:TBTT-ORH device exhibited greater charge-carrier mobilities than the PTB7-Th:TBTT-BORH device. The electron mobility (μe) of the as-cast TBTT-ORH device was 1.73 × 10–5 cm[2]/(V·s), which is 3 times greater than that of the as-cast TBTT-BORH device [5.34 × 10–6 cm2/(V·s)], which can be explained by the strong molecular aggregation behavior of TBTT-ORH. The higher μh and μe and smaller (i.e., more balanced) μh/μe ratio of the as-cast TBTT-ORH device compared with the case of TBTT-BORH well matched the JSC of the TBTT-ORH device. Compared to the as-cast TBTT-BORH film, the TBTT-ORH-based film exhibited a relatively smaller μh/μe ratio of 2.9 and a relatively high μe even before annealing, explaining the relatively high JSC of the as-cast TBTT-ORH device.

Figure 5

Dark J–V characteristics of (a) hole-only and (b) electron-only devices.

Moreover, consistent with the SVA-induced increase in the JSC, both the μh and μe values increased and became more balanced after the SVA treatment for both acceptors. Similar to the as-cast devices, the SVA-treated TBTT-ORH device also exhibited higher μh and μe values than the corresponding TBTT-BORH device (consistent with its higher JSC); however, the increase in mobility after annealing was greater in the TBTT-BORH device than in the TBTT-ORH device. The μh and μe of the TBTT-ORH device were increased approximately twofold by annealing, whereas the μh and μe of the TBTT-BORH device were increased more dramatically. Specifically, the μe of the TBTT-BORH device increased 6-fold after SVA treatment, from 5.34 × 10–6 cm2/(V·s) to 3.22 × 10–5 cm2/(V·s), which matches the increased molecular aggregation of TBTT-BORH after annealing, as explained in the previous discussion of the absorption results. In the as-cast condition, the introduction of the bulky side chain of the BO group resulted in lower crystallinity of the TBTT-BORH, which in turn resulted in reduced intermolecular charge transport (i.e., electron transport) among these molecules; however, upon annealing of the TBTT-BORH film, the μh and μe of the TBTT-BORH device could be greatly improved to levels comparable to those of the TBTT-ORH device. Because of the remarkably increased μe, the most balanced μh/μe ratio of 2.3 was obtained in the SVA-treated TBTT-BORH device.

Because the performance characteristics of the devices—in particular, their JSC and FF—were strongly dependent on the crystallinity of the two components and the bulk heterojunction (BHJ) morphology of the blend films,[74] we recorded height- and phase-mode AFM images for both acceptor-based photoactive films in the as-cast and SVA conditions (Figure ). The film samples were prepared in the same manner as the films used in the devices. Compared with the as-cast TBTT-BORH-based film (Figure a), which showed a uniform and smooth film morphology and relatively low film root-mean-square (rms) roughness (0.85 nm), the as-cast PTB7-Th:TBTT-ORH blend film (Figure c, especially in the phase mode) exhibited relatively large domains and a rougher film morphology (rms roughness = 1.17 nm). Such a relatively rougher surface of the TBTT-ORH blend film is attributed to the stronger aggregation behavior of TBTT-ORH relative to that of TBTT-BORH, as demonstrated by the UV–vis absorption and DSC measurements, which led to higher charge mobilities and better JSC of the PTB7-Th:TBTT-ORH device than the PTB7-Th:TBTT-BORH device.[75,76] A similar trend was observed in the TEM images: the TBTT-ORH-based film exhibited a relatively large domain size and a rougher morphology than the TBTT-BORH-based film.

Figure 6

(left to right) Height- and phase-mode AFM (2 μm × 2 μm) and TEM images of the (a) as-cast and (b) SVA-treated PTB7-Th:TBTT-BORH (BO) films and the (c) as-cast and (d) SVA-treated PTB7-Th:TBTT-ORH (O) films.

Moreover, the annealing treatment yielded rougher surfaces (i.e., higher rms values) and larger domain sizes than the corresponding as-cast films. The electron and hole transport could effectively occur within the large-sized domains, increasing the JSC and the PCEs of the devices. Again, the increase in rms roughness after annealing was more pronounced in the TBTT-BORH blend films than in the TBTT-ORH films, as was the increase in absorption intensity, JSC, and EQE intensity; specifically, a larger (about 71%) increase in rms roughness from 0.85 nm (as-cast) to 1.45 nm (SVA) was observed for the TBTT-BORH blend films, whereas a 19% increase (from 1.17 to 1.39 nm) was observed for the TBTT-ORH blend films.

In the present work, given that the material solubility was greatly improved by the introduction of a BO group in TBTT-BORH, the inevitably low intermolecular aggregation and electron mobility of the as-cast TBTT-BORH device could also be greatly increased by the relatively simple SVA post-treatment. The effects of the improved BHJ morphology, high charge-carrier mobilities, and suppressed charge recombination were demonstrated in the SVA-treated TBTT-BORH device, which exhibited a higher PCE than the TBTT-ORH device. Thus, the modification of the molecular structure by the simple side-chain engineering to enhance solubility and enable easier processing ultimately led to an increase in device efficiency. Thus, we have demonstrated the importance of determining the optimum device manufacturing conditions along with the proper molecular chemical structure to balance the trade-off between material solubility and crystallinity.

Conclusions

We synthesized two NFSMs, TBTT-BORH and TBTT-ORH, composed of a nonfused TBTT core flanked with alkyl-substituted RHs. Higher phase-transition temperatures and UV–vis absorption spectra with a greater red shift were observed for TBTT-ORH because of its relatively stronger molecular packing ability compared with that of TBTT-BORH with bulky BO side chains. On the contrary, despite the weak molecular aggregation behavior in the as-cast TBTT-BORH films, the molecular aggregation and charge-carrier mobilities of the TBTT-BORH films were substantially increased after the SVA treatment. Because TBTT-BORH and TBTT-ORH had a sufficient energy-level offset and complementary absorption properties with PTB7-Th, both NFSMs could be used as acceptors for OSCs with a PTB7-Th polymer donor. The best PCE of 8.33% was obtained with the device based on TBTT-BORH after SVA treatment, where, in addition to a high open-circuit voltage of 1.02 V for this device, its JSC and FF were also greatly improved by the SVA treatment. This device exhibited increased charge carrier mobilities and suppressed bimolecular and monomolecular charge recombination because of the increased molecular aggregation and improved film morphology of the TBTT-BORH after the SVA treatment.

Experimental Section

Materials

Bis(carboxymethyl)trithiocarbonate, 5-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)thiophene-2-carboxaldehyde (B-T-CHO), 1,2-dimethoxyethane (DME), 2-butyloctanol, hydrazine monohydrate, triphenylphosphine (PPh3), diethylene glycol monomethyl ether, and tris(dibenzylideneacetone)dipalladium(0) [Pd2(dba)3] were purchased from Tokyo Chemical Industry (Tokyo, Japan). Phthalimide and azodicarboxylic acid diisopropyl ester (DIAD) were purchased by Alfa Aesar (Gyeonggi-do, Korea). Piperidine, n-octylamine, potassium phosphate tribasic (K3PO4), tri-tert-butylphosphonium tetrafluoroborate (P(t-Bu)3·HBF4), and 4,7-dibromobenzo[c]-1,2,5-thiadiazole were purchased from Sigma-Aldrich (Gyeonggi-do, Korea). Triethylamine (TEA) was purchased from Samchun Chemicals (Seoul, Korea). All chemicals were used without further purification, and all reactions were performed under a nitrogen atmosphere with anhydrous solvents. 3-Octylrhodanine (ORH) was synthesized according to our previous report.[3]

Synthesis

Synthesis of 2-(2-Butyloctyl)isoindoline-1,3-dione (BOphth)[77]

Phthalimide (3.67 g, 25.0 mmol), PPh3 (7.86 g, 30.3 mmol), and 2-butyloctanol (4.65 g, 25.0 mmol) were dissolved in a solution of distilled tetrahydrofuran (THF) (100 mL); DIAD (6.06 g, 30.0 mmol) was then slowly added under stirring. The resultant solution was refluxed and stirred overnight under a N2 atmosphere. After the reaction solvent was evaporated, the crude product was dissolved in ethyl acetate (20 mL) and petroleum ether (100 mL). After the product was allowed to stand at 4 °C for 2 h, the precipitated yellow solid was removed by filtration and the filtrate was concentrated by evaporation to obtain an orange oil. The product was further purified by column chromatography [dichloromethane (DCM)] to obtain BOphth as a colorless oil. Yield: 6.64 g (84%). 1H NMR (400 MHz, CDCl3): δ (ppm) 7.86–7.81 (m, 2H), 7.73–7.68 (m, 2H), 3.57 (t, J = 7.3 Hz, 2H), 1.89–1.86 (m, 1H), 1.37–1.25 (m, 16H), 0.89–0.84 (m, 6H).

Synthesis of 2-Butyloctylamine (BONH2)[78]

BOphth (6.64 g, 21.1 mmol) was dissolved in a solution of methanol (100 mL), and then hydrazine monohydrate (3.16 g, 63.2 mmol) was added slowly under stirring; the resultant mixture was refluxed overnight under a N2 atmosphere. After the mixture was evaporated, the crude product was dispersed in DCM and then washed with 10 wt % KOH solution and saturated NaCl aqueous solution. The organic layer was dried with anhydrous MgSO4 and concentrated by evaporation to obtain BONH2 as a pale-yellow oil. Yield: 3.64 g (93%). This compound was used without further purification 1H NMR (400 MHz, CDCl3): δ (ppm) 2.60 (d, J = 4.7 Hz, 2H), 1.32–1.26 (m, 17H), 1.12 (s, 2H), 0.91–0.86 (m, 6H).

Synthesis of 3-(2-Butyloctyl)rhodanine (BORH)[79]

Bis(carboxymethyl)trithiocarbonate (4.4 g, 20 mmol) was completely dissolved in a solution of DME (40 mL). After TEA (2.0 g, 20 mmol) was added, the solution was stirred for 5 min, and then 2-butyloctylamine (3.6 g, 20 mmol) was added. The resultant solution was refluxed and stirred for 3 h under a N2 atmosphere. The reaction mixture was then extracted with DCM, washed with water, and dried over MgSO4. After removal of the solvent, the residue was purified by column chromatography (DCM/hexane = 3:1) to afford BORH as a yellow oil (4.8 g, 82% yield). 1H NMR (400 MHz, CDCl3): δ (ppm) 3.97 (s, 2H), 3.89 (d, J = 7.5 Hz, 2H), 2.07–1.98 (m, 1H), 1.35–1.18 (m, 16H), 0.90–0.86 (m, 6H). 13C NMR (100 MHz, CDCl3): δ (ppm) 201.60, 174.27, 48.91, 35.31, 35.16, 31.76, 31.37, 31.06, 29.58, 28.38, 26.16, 22.95, 22.61, 14.07, 14.01.

Synthesis of B-T-BORH[3]

B-T-CHO (0.53 g, 2.2 mmol) was dissolved in a solution of anhydrous CF (40 mL), a few drops of piperidine and then BORH (1.67 g, 5.52 mmol) were added; the resultant solution was refluxed and stirred for 1 h under a N2 atmosphere. The reaction mixture was then extracted with DCM, washed with water, and dried over magnesium sulfate. After removal of the solvent, it was recrystallized by methanol to afford B-T-BORH as a yellow solid (433 mg, 38% yield). 1H NMR (400 MHz, CDCl3): δ (ppm) 7.88 (s, 1H), 7.64 (d, J = 3.7 Hz, 1H), 7.43 (d, J = 3.7 Hz, 1H), 4.01 (d, J = 7.4 Hz, 2H), 2.12–2.09 (m, 1H), 1.36–1.25 (m, 28H), 0.90–0.85 (m, 6H). 13C NMR (100 MHz, CDCl3): δ (ppm) 193.00, 168.11 143.90, 137.99, 134.31, 124.64, 122.71, 84.72, 49.04, 35.66, 31.79, 31.44, 31.14, 29.59, 28.46, 26.24, 24.75, 22.98, 22.63, 14.09, 14.05.

Synthesis of B-T-ORH

This compound was synthesized in the same manner as B-T-BORH using compound B-T-CHO (1.0 g, 4.2 mmol), CF (50 ml), and ORH (2.6 g, 11 mmol) to give a yellow solid (1.43 g, 73% yield). 1H NMR (400 MHz, CDCl3): δ (ppm) 7.90 (s, 1H), 7.64 (d, J = 3.7 Hz, 1H), 7.44 (dd, J = 3.7, 0.5 Hz, 1H), 4.13–4.05 (m, 2H), 1.73–1.66 (m, 2H), 1.37–1.27 (m, 22H), 0.88 (t, J = 6.9 Hz, 3H). 13C NMR (100 MHz, CDCl3): δ (ppm) 192.62, 167.63, 143.87, 138.02, 134.40, 124.70, 122.87, 84.76, 44.85, 31.79, 29.14, 27.00, 26.79, 24.78, 22.65, 14.12.

Synthesis of TBTT-BORH

B-T-BORH (266 mg, 0.480 mmol), 4,7-dibromobenzo[c]-1,2,5-thiadiazole (71.0 mg, 0.24 mmol), and P(t-Bu)3·HBF4 (4.20 mg) were dissolved in a solution of distilled THF (11 mL), and then 1 M K3PO4 aqueous solution (0.7 mL) was added. After Pd2(dba)3 (6.60 mg) was added and the solution was degassed, the resultant solution was refluxed and stirred overnight under a N2 atmosphere. The reaction mixture was then extracted with CF, washed with water, and dried over magnesium sulfate. After removal of the solvent, it was purified by column chromatography (DCM) to afford TBTT-BORH as a red solid (129 mg, 58% yield). 1H NMR (400 MHz, CDCl3): δ (ppm) 8.19 (d, J = 4.1 Hz, 1H), 7.95 (s, 1H), 7.87 (s, 1H), 7.47 (d, J = 4.1 Hz, 1H), 4.03 (d, J = 7.5 Hz, 2H), 2.13 (m, 1H), 1.37–1.27 (m, 16H), 0.91–0.86 (m, 6H). 13C NMR (100 MHz, CDCl3): δ (ppm) 192.51, 167.91, 152.03, 146.18, 139.61, 134.71, 128.96, 125.82, 125.71, 124.68, 121.70, 49.13, 35.71, 31.81, 31.50, 31.18, 29.62, 28.49, 26.28, 23.00, 22.65, 14.10, 14.06. HRMS–FAB (m/z): [M + H]+ calcd for C46H59N4O2S7, 923.2683; found, 923.2664. Anal. Calcd for C46H58N4O2S7: C, 59.83; H, 6.33; N, 6.07; S, 24.31. Found: C, 60.32; H, 6.16; N, 6.06; S, 24.07.

Synthesis of TBTT-ORH

This compound was synthesized in the same manner as TBTT-BORH using compound B-T-ORH (1.70 g, 0.590 mmol), 4,7-dibromobenzo[c]-1,2,5-thiadiazole (600 mg, 1.30 mmol), P(t-Bu)3·HBF4 (10.0 mg), THF (30 ml), 1M K3PO4 in water (1.7 mL), and Pd2(dba)3 (16.0 mg) to give a red solid (0.30 g, 63% yield). 1H NMR (400 MHz, CDCl3): δ (ppm) 8.20 (d, J = 4.1 Hz, 1H), 7.96 (s, 1H), 7.89 (s, 1H), 7.49 (d, J = 4.1 Hz, 1H), 4.14–4.10 (m, 2H), 1.76–1.69 (m, 2H), 1.39–1.28 (m, 10H), 0.89 (t, J = 6.9 Hz, 3H). 13C NMR (125 MHz, tetrachloroethane): δ (ppm) 195.33, 170.41, 155.20, 149.33, 142.72, 137.84, 132.20, 129.10, 128.92, 127.92, 124.94, 48.07, 34.80, 32.14, 30.06, 29.86, 25.67, 17.15. HRMS–FAB (m/z): [M + H]+ calcd for C38H43N4O2S7, 811.1431; found, 811.1427. Anal. Calcd for C38H42N4O2S7: C, 56.26; H, 5.22; N, 6.91; S, 27.67. Found: C, 56.58; H, 5.12; N, 6.80; S, 27.92.