Small-Molecule Solar Cells with Fill Factors up to 0.75 via a Layer-by-Layer Solution Process
Article first published online: 6 AUG 2013
© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Advanced Energy Materials
Volume 4, Issue 1, January 7, 2014
How to Cite
2014). Small-Molecule Solar Cells with Fill Factors up to 0.75 via a Layer-by-Layer Solution Process. Adv. Energy Mater., 4: 1300626. doi: 10.1002/aenm.201300626, , , , , (
- Issue published online: 6 JAN 2014
- Article first published online: 6 AUG 2013
- Manuscript Received: 5 JUN 2013
- 973 Project. Grant Number: 2011CB808401
- NSFC. Grant Numbers: 21025418, 51261130582, 21021091
- Chinese Academy of Sciences
- organic solar cells;
- small-molecule devices;
- solution processing;
- layer-by-layer processing
Organic solar cells (OSCs) are a promising cost-effective alternative for using solar energy; they are advantageous because of their low cost, light weight, and flexibility.[1-3] Both bilayer and bulk donor/acceptor (D/A) heterojunction OSCs can be fabricated by vacuum deposition or solution processing; solution processing is a better choice for large-area and low-cost devices. Over the last decade, much of the focus has been on the development of solution-processed bulk-heterojunction (BHJ) OSCs, which have seen a dramatic rise in efficiency. The encouraging power conversion efficiency (PCE) of over 9% has been achieved from devices based on blends of polymer donors and fullerene acceptors. Compared to their polymer counterparts, small molecules offer potential advantages such as defined molecular structure, definite molecular weight, easy purification, and good batch-to-batch reproducibility.[7-9] Recently, great efforts have been dedicated for developing small molecules for applications in solar cells, and so far the highest PCE values of small-molecule-based BHJ OSCs are up at 6–8%.[10-17] A signficant challenge for BHJ OSCs may be finding strategies for controlling morphology in order to maximize the D/A interfacial area while minimizing structurally induced recombination losses, allowing for a large short-circuit current density (JSC) and high fill-factor (FF).
In bilayer devices, each component of the bilayer can be separately optimized avoiding the problem of controlling the blend morphology. However, the reduced D/A interfacial area tends to give smaller JSC values. Vacuum-deposited bilayer-heterojunction OSCs have been investigated since the 1980s, and their efficiencies have reached 5–6%.[19, 20] Recently, solution-processed “bilayer” (layer-by-layer) OSCs were investigated in the typical system of poly(3-hexylthiophene) (P3HT)/fullerene derivatives.[21-25] According to these research works, after spin-coating overlayers on P3HT, significant interdiffusion of the fullerene acceptor into the P3HT layer naturally occurs. Thus, this layer-by-layer solution processing method is useful for creating p-i-n (D/D:A/A)-like structure, which can greatly improve exciton dissociation and charge transport. However, to our knowledge, there have been no reports on the design and synthesis of novel photovoltaic materials for solution-processed “bilayer” OSCs so far.
Oligothiophenes are one of the largest families of organic semiconductors, and oligothiophene-based push–pull molecules, including linear, X-shaped, star-shaped, and dendritic structures, have shown promising performance in OSCs. Chen and co-workers[10, 13, 27-29] and Bäuerle and co-workers[16, 30-32] have reported several linear oligothiophenes end-capped with electron-withdrawing groups for high-performance OSCs. Wong and co-workers have developed a novel 2D push–pull oligothiophene and investigated its photovoltaic properties. Roncali and co-workers have done systematic research on structure–property relationships in star-shaped triphenylamine-hybrid push–pull oligothiophenes.[34-38] We have also reported multidimensional oligothiophene-related donors for BHJ OSCs with PCE values of up to 4.3%.[39-42]
Here, we present the first example of the molecular design of novel materials for solution-processed “bilayer” (layer-by-layer) OSCs. To reduce the solubility and improve the planarity of the molecule, we removed six n-octyl side-chains on the oligothiophene and introduced two 5-alkylthiophene groups onto the benzodithiophene (BDT) core of the previously reported DCAO3T(BDT)3T to form a linear push–pull oligothiophene (BDT-3T-CA, Figure 1). Compound BDT-3T-CA shows good thermal stability and high crystallinity, selective solubility in common solvents, a relatively broad absorption spectra with a high extinction coefficient, a low HOMO (highest occupied molecular orbital) energy level, and a high hole mobility. Solution-processed layer-by-layer solar cells based on BDT-3T-CA/PC61BM (where PC61BM is the fullerene derivative, [6,6]-phenyl C61 butyric acid methyl ester) exhibit PCE values as high as 4.16%, with excellent FF values of up to 0.75. To our knowledge, the FF of 0.75 is one of the highest values reported for solution-processed OSCs so far, and it is a record FF for solution-processed small-molecule solar cells.
Compound BDT-3T-CA was synthesized by simple reactions (Scheme 1). The intermediate compound BDT-3T-CHO was synthesized in 77% yield through a Stille coupling reaction between 2,6-bis(trimethyltin)-4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)benzo[1,2-b:4,5-b′]dithiophene and 5′′-bromo-2,2′:5′,2′′-terthiophene-5-carboxaldehyde using Pd(PPh3)4 as the catalyst (Ph represents phenyl groups). A straightforward reaction of BDT-3T-CHO with n-octyl cyanoacetate afforded BDT-3T-CA in 91% yield. Compound BDT-3T-CA was characterized by matrix-assisted laser-desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS), 1H NMR, and elemental analysis (see Supporting Information (SI)). Due to the reduced number of alkyl chains, the solubility of BDT-3T-CA decreased relative to DCAO3T(BDT)3T. At room temperature, dichloromethane (DCM) can slightly dissolve BDT-3T-CA (<0.06 mg/mL), while the solubility of BDT-3T-CA in chloroform is significantly improved (ca. 0.8 mg/mL at room temperature versus 8 mg/mL at 50 °C). The thermal properties of BDT-3T-CA were investigated by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC). This compound exhibits good thermal stability with a decomposition temperature (5% weight loss) of 302 °C in nitrogen (Figure S1, SI). BDT-3T-CA is crystalline with a high melting point of 255 °C, as observed from the DSC heating trace (Figure S2, SI); during the cooling process, an exothermic peak due to crystallization was observed at 241 °C. Additionally, there are another couple of peaks at 88 °C (heating) and 81 °C (cooling), indicating that a solid–solid phase transition may occur.
Figure 2a shows the normalized UV–vis absorption spectra of BDT-3T-CA in dilute chloroform solution (10−6 m) and in a thin-film solid. Compared to DCAO3T(BDT)3T (absorption maximum, λmax = 478 nm and extinction coefficient, ε = 3.1 × 104m−1 cm−1), BDT-3T-CA in solution exhibits a stronger and red-shifted absorption with ε = 1.15 × 105m−1 cm−1 at λmax = 518 nm due to the reduced number of alkyl chains and improved planarity of BDT-3T-CA. The thin film of BDT-3T-CA shows intense absorption throughout the visible region (300–700 nm). The absorption peak (560 nm) of the BDT-3T-CA thin film is red-shifted by 42 nm relative to that in chloroform solution. In addition, the absorption spectra of the BDT-3T-CA film exhibits another sharp peak at 606 nm, which means that there is good alignment of the linear molecular chains. Compared to that in solution, the structured absorption spectrum of the film sample as well as its red-shift suggests that strong intermolecular interaction and aggregation exists in the film. The optical bandgap estimated from the absorption edge (662 nm) of the thin film is 1.87 eV. The photoluminescence (PL) spectra of BDT-3T-CA in solution and in film are shown in Figure 2b. The solution exhibits a maximum emission at 638 nm, while the film shows a red-shifted emission with peaks at 683 and 716 nm; this resembles the trend in absorption spectra and also indicates strong intermolecular interactions and an ordered alignment of the BDT-3T-CA chains in the film. The PL band located at ca. 800 nm is attributed to emission from the excimer caused by molecular aggregation.
The electrochemical properties of BDT-3T-CA was investigated by cyclic voltammetry (CV) as a film on a glassy carbon working electrode in 0.1 m [nBu4N]+[PF6]− CH3CN solution (nBu = n-butyl) at a potential scan rate of 100 mV s−1. As shown in Figure 2c, BDT-3T-CA exhibits two couples of reversible reduction waves and quasi-reversible oxidation waves. The onset oxidation and reduction potentials versus FeCp2+/0 are 0.40 and −1.90 V, respectively (Cp = cyclopentadienyl). The HOMO and LUMO (lowest unoccupied molecular orbital) energies were estimated to be −5.20 and −2.90 eV from the onset oxidation and reduction potentials, respectively, assuming the absolute energy level of FeCp2+/0 to be 4.8 eV below vacuum. The LUMO level of BDT-3T-CA is higher than that (−3.91 eV) of PC61BM; the LUMO gap between the donor and acceptor is large enough to guarantee photoinduced electron transfer between them. The difference between the HOMO of BDT-3T-CA and the LUMO of PC61BM is as large as 1.29 eV, which could lead to a high open-circuit voltage (VOC) in solar cells, and the theoretical value of VOC is ca. 0.99 V.[45, 46]
To investigate the structural ordering of BDT-3T-CA in solid state, we performed X-ray diffraction (XRD) analysis on the thin film produced by spin-coating the BDT-3T-CA chloroform solution (at 50 °C) onto octadecyltrichlorosilane(OTS)-treated SiO2/Si substrate. The XRD pattern (Figure 2d) exhibits a strong reflection peak (100) at 2θ = 6.0°, corresponding to a d100-spacing distance of 14.72 Å. The second-order diffraction peak (200) at 2θ = 11.9° and the third-order diffraction peak (300) at 2θ = 18.1°, corresponding to a d200-spacing value of 7.45 Å and a d300-spacing value of 4.90 Å, were also clearly observed, implying a highly ordered assembly of the linear π-conjugated molecule in the solid state. Due to the solid–solid phase transition in BDT-3T-CA speculated from the DSC traces, the XRD patterns of the BDT-3T-CA film were investigated at different temperatures under a nitrogen atmosphere (Figure S3, SI). Compared to the XRD patterns at room temperature, the XRD pattern at 120 °C showed a new reflection peak at 2θ = 3.6°, corresponding to a relatively large d-spacing distance of 25.20 Å. At 140 °C, the reflection peak at 2θ = 3.6° increased, while the reflection peak at 2θ = 6.0° decreased. This phenomenon shows that the structural order in the BDT-3T-CA film has changed with increasing temperature. However, the relatively large d-spacing distance of 25.20 Å at high temperature may exert negative impact on charge transport and OSC device performance.
To measure the hole mobility of BDT-3T-CA, organic field-effect transistors (OFETs) were fabricated by spin-coating the BDT-3T-CA chloroform solution onto OTS-treated SiO2/Si substrates. In a top-contact geometry using Au as the source and drain electrodes, compound BDT-3T-CA exhibits typical p-type semiconductor behavior (Figure 3a). The field-effect mobility was calculated to be 0.01 cm2 V−1 s−1 with a current on/off ratio of 103 and a threshold voltage of around −5 V from the transfer characteristics (Figure 3b). The hole mobility of 0.01 cm2 V−1 s−1 is among the highest values for small-molecule donors in OSCs, and it is comparable to that of PC61BM, which is beneficial for balancing hole and electron transport in “bilayer” OSCs. In addition, the hole mobility of BDT-3T-CA on PEDOT:PSS (poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate)) substrate was measured using the space-charge-limited current (SCLC) method in the system: indium tin oxide (ITO)/PEDOT:PSS/BDT-3T-CA/Au (Figure S4, SI). BDT-3T-CA as a cast film on PEDOT:PSS showed a hole mobility of up to 2.5 × 10−3 cm2 V−1 s−1, which is 1 order of magnitude higher than that (4.5 × 10−4 cm2 V−1 s−1) of DCAO3T(BDT)3T, probably due to the improved planarity, strong intermolecular interactions, and ordered alignment of the BDT-3T-CA chains caused by the fewer alkyl chains.
Due to the different solubilities of BDT-3T-CA (<0.06 mg/mL) and PC61BM (ca. 10 mg/mL) in DCM at room temperature, it is possible for solution-processed layer-by-layer solar cells to be fabricated by depositing a PC61BM overlayer on a BDT-3T-CA film from a DCM solution of PC61BM. The morphological effect of DCM dropping on the BDT-3T-CA film was investigated by atomic force microscopy (AFM) in tapping mode. The as-cast BDT-3T-CA film with a thickness of 40 nm on a PEDOT:PSS-coated glass substrate shows typical cluster structure with many crystalline domains and a relatively rough surface with a root-mean-square (RMS) roughness of 2.51 nm (Figure 4a). After a drop (30 μL) of DCM solvent was spin-coated onto the BDT-3T-CA film, it exhibited similar crystalline domain size and film thickness to that beforehand, but it obviously had a rougher surface with higher RMS roughness (5.56 nm; Figure 4b). This phenomenon implies that after spin-coating PC61BM from the DCM solution on the BDT-3T-CA underlayer, significant interdiffusion of PC61BM into the BDT-3T-CA layer could occur and form an intermixed blend layer of BDT-3T-CA:PC61BM between the BDT-3T-CA and PC61BM layers, leading to a quasi p-i-n device architecture.
The photophysics of the solution-processed BDT-3T-CA/PC61BM “bilayer” was investigated by UV–vis absorption and PL spectra. The absorption spectra of the “bilayer” film (Figure S5, SI) show two strong peaks at ca. 560 and 610 nm, which are similar to those of BDT-3T-CA neat film. However, in the control absorption spectra of BDT-3T-CA:PC61BM (1:1, w/w) blend film, the peak (ca. 610 nm) due to molecular ordering and aggregation is suppressed to some extent because PC61BM molecules in the blend film reduce intermolecular interactions among the BDT-3T-CA molecules, which is verified by XRD investigation. The XRD patterns of neat BDT-3T-CA thin films and BDT-3T-CA/PC61BM “bilayer” films on PEDOT:PSS-coated ITO/glass substrates show a reflection peak at 2θ = 6.0° (d-spacing distance of 14.72 Å), similar to the BDT-3T-CA films on OTS-treated SiO2/Si substrates; meanwhile the blend films of BDT-3T-CA:PC61BM (1:1, w/w) exhibit a reflection peak at 2θ = 5.6°, corresponding to a larger d-spacing distance of 15.77 Å (Figure S6, SI). These results indicate that molecular ordering of BDT-3T-CA is almost insusceptible to the PC61BM molecules in the “bilayer” film. Deposition of the PC61BM overlayer results in 90% quenching of BDT-3T-CA fluorescence; this is a little less than that of BDT-3T-CA:PC61BM (1:1, w/w) blend films with the same thickness, which indicates that effective photoinduced charge transfer occurred between BDT-3T-CA and PC61BM in the “bilayer” film (Figure S7, SI).
To demonstrate potential application of BDT-3T-CA in OSCs, we used BDT-3T-CA as an electron donor and PC61BM as an electron acceptor, and fabricated solution-processed layer-by-layer OSCs with a structure of ITO/PEDOT:PSS/BDT-3T-CA/PC61BM/Ca/Al. The donor underlayer was spin-coated onto PEDOT:PSS-coated substrates from hot chloroform solution (at 50 °C, 7.5 mg mL−1) of neat BDT-3T-CA, and then the acceptor overlayer was spin-coated (at 3500 rpm) onto the donor layer from a DCM solution (30 μL) of PC61BM at room temperature. For device optimization, we varied donor layer thickness by adjusting rotation speed and acceptor layer thickness by adjusting solution concentration. Table 1 summarizes the VOC, JSC, FF, and PCE values of the devices at different thicknesses of the donor and acceptor layers. These devices yield relatively high VOC as a result of the significant difference between the HOMO of BDT-3T-CA and the LUMO of PC61BM, and the VOC values (0.87–0.88 V) are almost insensitive to the thickness of the donor and acceptor layers. The thickness of donor and acceptor layers affects JSC, FF, and PCE of the devices. The layer-by-layer device at donor/acceptor layer thicknesses of 45 nm/40 nm gives the best performance: VOC, JSC, FF, and PCE reach 0.88 V, 6.30 mA cm−2, 0.75, and 4.16%, respectively, without any post-treatment (Figure 5a). For the donor/acceptor layer thicknesses of 45 nm/40 nm, more than 10 devices were fabricated, and their average PCE was 4.02% (3% device variation) with VOC values of 0.87–0.88 V, JSC values of 6.06–6.35 mA cm−2, and FFs of 0.738–0.756.
|Thickness of BDT-3T-CA layer [nm]||Thicknes of PC61BM layer [nm]||VOC [V]||JSC [mA cm−2]||FF||PCE [%]|
After thermal annealing at 120 °C for 10 min, the device with donor/acceptor layer thicknesses of 45 nm/40 nm exhibited a PCE of 3.35%, with a VOC of 0.91 V, a JSC of 5.85 mA cm−2, and a FF of 0.63 (Figure S8, SI). Compared to the as-cast devices, thermal annealing led to a decrease in JSC and FF values. We investigated the effect of thermal annealing on hole mobility and morphology of the BDT-3T-CA layer. Thermal annealing the BDT-3T-CA film at 120 °C for 10 min showed a lower SCLC hole mobility of 1.2 × 10−3 cm2 V−1 s−1 (Figure S4, SI) and a rougher surface with an RMS roughness of 3.31 nm (Figure S9, SI) than the as-cast film (hole mobility: 2.5 × 10−3 cm2 V−1 s−1; RMS roughness: 2.51 nm). According to early research by Yan et al., the decrease in device performance as a result of thermal annealing can be partly attributed to the decrease in hole mobility of BDT-3T-CA and/or the increase in interface roughness.
To the best of our knowledge, the FF of 0.75 is one of the highest values reported for solution-processed OSCs so far, and it is a record FF value for solution-processed small-molecule solar cells. The “bilayer” architecture provides continuous pathways for hole and electron transport, and it exhibits a photocurrent that saturates at relatively small electric fields, leading to high FF. High hole mobility and balanced hole and electron transport in the “bilayer” are another origin for this high FF. In the equivalent circuit model, the estimated FF value is up to 0.872 in ideal solar cell devices with no series resistance and shunt resistance based on BDT-3T-CA/PC61BM. The difference between the experimental and calculated values from an ideal solar cell model can be accounted for by the energy losses from unavoidable parasite resistances (such as the bulk resistances of materials and electrodes, or current leakage) in the real solar cells.
The incident-photon-to-converted-current efficiency (IPCE) spectra of the layer-by-layer solar cell based on BDT-3T-CA/PC61BM (45 nm/40 nm) is shown in Figure 5b. The “bilayer” film as cast has a broad IPCE spectra from 300 to 700 nm and a peak value of 36.9% at 564 nm along with a shoulder of 35.6% at 605 nm.
In summary, we have presented the first example of a molecular design of novel materials for solution-processed layer-by-layer OSCs; a linear push–pull oligothiophene (BDT-3T-CA) with 5-alkylthiophene-2-yl-substituted BDT as the core and terthiophene end-capped with n-octyl cyanoacetate as the arms was synthesized. BDT-3T-CA has good thermal stability, selective solubility in common solvents, highly crystallinity, broad and strong absorption with a narrow bandgap, suitable energy levels matched to that of PC61BM and a high hole mobility of up to 0.01 cm2 V−1 s−1. Due to the significant difference in solubility of BDT-3T-CA and PC61BM in DCM at room temperature, solution-processed layer-by-layer solar cells are fabricated by depositing PC61BM overlayer on BDT-3T-CA film from DCM solution. Effective photoinduced charge transfer occurred between BDT-3T-CA and PC61BM, while molecular ordering of BDT-3T-CA is almost insusceptible to PC61BM molecules in the “bilayer” film. The solution-processed layer-by-layer OSCs based on BDT-3T-CA/PC61BM afford a PCE of up to 4.16% with a FF of up to 0.75; the FF of 0.75 is one of the highest values reported for solution-processed OSCs so far and a record FF for solution-processed small-molecule solar cells. This preliminary work demonstrates that BDT-3T-CA is a suitable donor material for solution-processed layer-by-layer OSCs.
The authors thank the 973 Project (2011CB808401), the NSFC (21025418, 51261130582, 21021091), and the Chinese Academy of Sciences for financial support.
As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.