Yikun 
            Guo
          
        
      a, 
      
        
          
            Yunke 
            Li
          
        
      b, 
      
        
          
            Omar 
            Awartani
          
        
      c, 
      
        
          
            Han 
            Han
          
        
      a, 
      
        
          
            Guangye 
            Zhang
          
        
      b, 
      
        
          
            Harald 
            Ade
          
        
      *c, 
      
        
          
            He 
            Yan
          
        
      *b and 
      
        
          
            Dahui 
            Zhao
          
        
      
*a
      
aBeijing National Laboratory for Molecular Sciences, Chemistry, Centre for the Soft Matter Science and Engineering and the Key Laboratory of Polymer Chemistry and Physics of the Ministry of Education, College of Chemistry, Peking University, Beijing 100871, China. E-mail: dhzhao@pku.edu.cn
      
bDepartment of Chemistry and Hong Kong Branch of Chinese National Engineering Research Center for Tissue Restoration & Reconstruction, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China. E-mail: hyan@ust.hk
      
cDepartment of Physics, North Carolina State University, Raleigh, NC 27695, USA. E-mail: hwade@ncsu.edu
    
First published on 16th February 2017
The side-chain structures of conjugated molecules are well recognized to sensitively influence the crystallinity, morphology and thus carrier transport properties of organic semiconductors. Here, by varying the alkyl side-chain length in the polymer acceptors, the effect of side-chain engineering on the photovoltaic performance is systematically studied in all-polymer solar cells. Clear trends of first an increase and then a decrease in the Jsc and FF values are observed as the branched alkyl groups are extended from 4 to 8 carbons. Correspondingly, the maximum average PCE (ca. 7.40%) is attained with an acceptor bearing a branched side-chain length of seven carbon atoms.
Numerous studies with organic semiconductors have demonstrated that side-chain engineering can bring about prominent differences in a material’s properties and performance in devices, such as thin film transistors (TFTs)18,19 and PSCs.20–22 In particular, the length of the side chains is one of the key factors that can be tweaked to optimize the photovoltaic performance of fullerene-based PSCs.23–25 Similarly, fine tuning the side-chain structures of polymer donors and acceptors plays an important role in optimizing the PCE of all-PSCs.4,26–28 Since the side groups sensitively influence the inter-chain interactions among polymer backbones, minor structure modifications to the side groups may create dramatic changes to the overall nanoscale polymer-blend morphology in the active layer of all-PSCs, hence sensitively influencing the device performance. Therefore, side-chain engineering represents an effective protocol and indispensable step in material structure optimizations aimed at enhancing the PCE of all-PSCs.
Very recently, by employing PDI-V comprising a backbone of alternating perylenediimide (PDI) and vinylene units as the polymer acceptor, along with a classical polymer donor PTB7-Th,29 we obtained an all-PSC exhibiting an impressive PCE of 7.57% (with an average PCE of 7.3%). Most remarkably, these PDI-V-based solar cells exhibited extraordinary stability toward oxygen and water in the air. Under ambient conditions with 90% humidity, a PCE of up to 7.49% was realized with cells without any encapsulation. This is a unique property particularly valuable for industrial processing.15
In light of such distinguished cell performance, we were encouraged to conduct a systematic side-chain engineering study on the PDI-V acceptor. Previously, the side-chain optimizations of polymer acceptors have prevailingly been conducted with naphthalene diimide (NDI)-based polymers.4,26–28 As an emerging family of potent acceptor materials, the side-chain effects with PDI-based polymers are basically unexplored, and relevant investigations are highly warranted. A series of PDI-vinylene polymers with varied alkyl chain lengths (named as PDI-V-Cx, with x representing the number of carbons in the branched alkyl side chains) were synthesized and investigated (Fig. 1a). Again, using PTB7-Th as the donor material, the performance of these polymer acceptors was examined, and a set of side-chain length dependent cell parameters were observed and recorded. Basically, while open-circuit voltage (Voc) exhibits minimal variations, both short-circuit current (Jsc) and fill factor (FF) manifest consistent trends of variation, namely first increasing and then decreasing, as the number of carbon atoms in the branched alkyl groups increases from 4 to 8. Atomic force microscopy (AFM) and grazing-incidence wide angle X-ray scattering (GIWAXS) were employed to characterize the effect of varying the side-chain length on the surface morphology, molecular packing and polymer aggregation of the neat and blend films. Certain correlations can be drawn between the device performance and morphological data. Optimal morphology appears to emerge with medium side chain lengths, offering higher and more balanced hole and electron mobilities, as well as favourable FF. The highest average PCE of 7.40% is thus achieved with the acceptor with an alkyl chain length of 7 carbon atoms in the branched side groups. Importantly, the current work shows that all examined PDI-polymer acceptors afford impressive PCEs of 5.5 to 7.4%, regardless of the side chain length. Such robust device performance and structural flexibility related to side-chain diversity implies a great advantage for large-scale solar cell production, because a great variety of device fabrication conditions can be applied and tested.
:
4, according to 1H NMR spectroscopy)30 were used in the polymerization reactions to afford the corresponding regio-irregular polymers. More synthesis details can be found in the ESI.† Due to the “dove-tailed” alkyl side chains on PDI, all PDI-V-Cx exhibited adequate solubility in common organic solvents, including o-dichlorobenzene (DCB), chlorobenzene (CB) and chloroform. The number-average molecular weights (Mn) of the polymer series were measured with gel permeation chromatography (GPC) in 1,2,4-trichlorobenzene at an elevated temperature of 150 °C, and the data are summarized in Table 1. The results show that all polymers of PDI-V-Cx with different side chains showed similar Mn and polydispersity (Mw/Mn), which excluded significant device performance variations caused by molecular weight differences.
        
| Polymer | M n (kDa) | M w/Mn | 
                    E
                    optg a (eV) | 
                  LUMOb (eV) | HOMOc (eV) | E-mobilityd (cm2 V−1 s−1) | 
|---|---|---|---|---|---|---|
| a Optical bandgap obtained in the film state. b Measured with cyclic voltammetry. c Estimated from (LUMO − Eoptg). d Calculated using the SCLC model. | ||||||
| PDI-V-C4 | 13.4 | 2.03 | 1.73 | −4.04 | −5.77 | 1.2 × 10−3 | 
| PDI-V-C5 | 13.4 | 1.96 | 1.72 | −4.04 | −5.76 | 1.1 × 10−3 | 
| PDI-V-C6 | 15.1 | 2.03 | 1.73 | −4.03 | −5.76 | 1.9 × 10−3 | 
| PDI-V-C7 | 13.7 | 2.23 | 1.73 | −4.02 | −5.75 | 2.7 × 10−3 | 
| PDI-V-C8 | 15.2 | 2.08 | 1.73 | −4.04 | −5.77 | 1.1 × 10−3 | 
:
1. Measurements of the device performance were carried out under the illumination of AM 1.5 simulated solar light at 100 mW cm−2. The J–V curves are shown in Fig. 2a, and the corresponding parameters are summarized in Table 2. In general, as both branches in the alkyl side chains are extended from 4 to 8 carbons, both Jsc and FF values are found to increase first before decreasing with chain length. As the Voc remains nearly constant over the entire range of examined side chain length, such variations in Jsc and FF naturally lead to a similar trend in the variation of average PCE, as the polymer side chain differs. In particular, among all examined devices, those comprising the active layers of PTB7-Th:PDI-V-C6 and PTB7-Th:PDI-V-C7 display Jsc values of 15.55 and 15.30 mA cm−2, respectively, which are evidently higher than those observed with the other three acceptors. Meanwhile, higher FF values ranging from 0.63 to 0.65 are obtained for PDI-V-C6/C7/C8, which are considerably higher than those for PDI-V-C4/C5. Overall, the polymers PDI-V-C7 and PDI-V-C6 provide better device performance than the other three acceptors bearing either longer or shorter side chains. The best average PCE of 7.40% is achieved with PTB7-Th:PDI-V-C7, which is slightly higher than the PCE of PTB7-Th:PDI-V-C6 devices (Table 2). Similar to the results shown in our previous work,15 the presence of various solvent additives, such as 1,8-diiodooctane and 1-chloronaphthalene, resulted in lowered PCEs for PTB7-Th:PDI-V-C7 (Table S1, ESI†). The fact that optimal device performance can be achieved without additives or post-treatments presents a great advantage for large-scale production.
        ![]()  | ||
| Fig. 2 (a) J–V curves and (b) EQE spectra of all-PSCs comprising the active layer of PTB7-Th:PDI-V-Cx. | ||
| Blends | V oc (V) | J sc (mA cm−2) | FF | PCEa (%) | μ e (cm2 V−1 s−1) | μ h (cm2 V−1 s−1) | μ h/μe | 
|---|---|---|---|---|---|---|---|
| a Data averaged over 10 cells, with the highest value shown in brackets. | |||||||
| PTB7-Th:PDI-V-C4 | 0.73 ± 0.01 | 14.44 ± 0.23 | 0.53 ± 0.03 | 5.58 ± 0.17 (5.75) | 0.8 × 10−4 | 6.9 × 10−4 | 8.6 | 
| PTB7-Th:PDI-V-C5 | 0.73 ± 0.01 | 14.88 ± 0.17 | 0.55 ± 0.02 | 5.97 ± 0.13 (6.06) | 1.5 × 10−4 | 8.3 × 10−4 | 5.5 | 
| PTB7-Th:PDI-V-C6 | 0.74 ± 0.01 | 15.55 ± 0.23 | 0.63 ± 0.02 | 7.25 ± 0.20 (7.38) | 1.8 × 10−4 | 7.6 × 10−4 | 4.2 | 
| PTB7-Th:PDI-V-C7 | 0.74 ± 0.01 | 15.30 ± 0.07 | 0.65 ± 0.01 | 7.40 ± 0.06 (7.44) | 2.4 × 10−4 | 1.0 × 10−3 | 4.2 | 
| PTB7-Th:PDI-V-C8 | 0.74 ± 0.01 | 14.83 ± 0.12 | 0.63 ± 0.03 | 6.91 ± 0.10 (6.99) | 2.1 × 10−4 | 1.1 × 10−3 | 5.2 | 
The external quantum efficiency (EQE) was measured to compare the spectral response of different acceptors. As shown in Fig. 2b, devices containing different acceptors all exhibit a wide photo-responsive range of 300 to 800 nm, illustrating that both the donor and acceptor contribute to light harvesting in these all-PSCs. The maximum EQE values of the different devices all reach over 69%, and the variation of the maximum EQE is well correlated with the changes in Jsc values observed with different acceptors. The photocurrent values calculated from the EQE spectra are also found to be consistent with the Jsc values obtained from the J–V measurements of the corresponding acceptors, showing differences of less than 4% (see Table S2, ESI†).
Further insight into the relationship between the polymer structure and device performance was gained by measuring the hole (μh) and electron mobility (μe) of the polymer blends using the SCLC method. As shown in Table 2, the bulk charge carrier transport properties of the active layer blends are generally consistent with the observed photovoltaic properties, especially with the FF values. Among all fabricated devices, the PTB7-Th:PDI-V-C7 blend shows the highest μe and the most balanced μh/μe, suggesting rapid charge transport and efficient charge extraction with minimized charge recombination within this polymer blend.31–33 The combination of these effects results in the highest FF and highest PCE values at this side chain length.
Photoluminescence (PL) quenching experiments of the polymer blend films were also performed to assess the percentage of excitons that can reach and dissociate at the donor/acceptor interface (Fig. S3, ESI†). Blending PDI-V-C7 with PTB7-Th results in fluorescence quenching of about 95% compared with the neat PDI-V-C7 films, indicating highly efficient photo-induced charge transfer between PTB7-Th and PDI-V-C7 in the blend film. In general, the quenching efficiencies of the different blends are found to be similar in the range of 93–95% (Fig. S3a, ESI†).
To understand the effect of side-chain length manipulation on molecular packing and device performance, GIWAXS34 measurements were carried out. The 2-D GIWAXS patterns and 1-D scattering profiles for neat PDI-V-C4-8 films are provided in Fig. 4 and 5, respectively. It is clear that all films have relatively low crystallinity, which is evident from the absence of higher order lamellar peaks in both the in-plane (IP) and out-of-plane (OOP) directions. The (100) lamellar peak is observed at a different q in the IP and OOP directions especially in PDI-V-C4 and PDI-V-C5, with an oval shape (100) signal in the 2D pattern. The difference between the IP and OOP (100) peak is reduced with an increased side-chain length. This is readily noticeable as the oval feature becomes more circular with increasing side-chain length. The IP lamellar spacing is reduced (from 23.4 Å to 22.6 Å) and the OOP lamellar spacing is increased (from 13.7 Å to 15.5 Å) with increasing side-chain length, especially from PDI-V-C5 to PDI-V-C6. Moreover, a (001) peak is present in all neat PDI-V-Cx films at q = 0.76–0.80 Å−1. This indicates that the polymer backbone direction is predominantly in the plane of the film. A relatively sharp feature is observed at q = 0.26 Å−1 and q = 0.25 Å−1 in the PDI-V-C4 and PDI-V-C5, respectively, and not in any of the other films. It is not entirely clear what these features are due to, but it is worth noting that the devices based on these two films have significantly lower FFs than those with the other films. It is possible that these features correspond to a polymorph with less interdigitated lamellar-packing. Further investigation is necessary to fully elucidate the nature of these features. The (010) peak in the OOP direction without a clear corresponding IP (100) signature indicates a preferential “rolling-log”, 2D orientation distribution in all five neat films. Increasing the side-chain length appears to also increase the (010) OOP spacing from 3.9 Å to 4.2 Å.
![]()  | ||
| Fig. 4 2-D GIWAXS patterns of neat PDI-V-Cx films: (a) PDI-V-C4, (b) PDI-V-C5, (c) PDI-V-C6, (d) PDI-V-C7 and (e) PDI-V-C8. | ||
When blending PTB7-Th with PDI-V-Cx, we observe mostly similar morphological features to the neat PDI-V-Cx films as shown in Fig. 6 and 7. However, an important difference in the neat films is the appearance of two peaks at ∼0.8 Å−1. The lower q peak corresponds to the (001) peak of PDI-V-Cx and the second peak is attributed to a (001) peak of PTB7-Th (see Fig. S4, ESI†). It is noticeable that as the side chain length increases, the PDI-V-Cx (001) peak weakens and the intensity of the second peak increases. This suggests that the in-plane stacking order of the PDI-V-Cx backbone is distorted as the side-chain length is extended. Meanwhile, the backbone ordering of PTB7-Th is improved. The coherence length and the d-spacing in the π–π (010) direction are similar for all blend films, which are not substantially influenced by varying the side-chain length. However, correlations between the (100) peak of the neat PDI-V-Cx films and FF and Jsc can help illuminate some structure-performance dependence. As the side-chain length is increased the lamellar spacing in the IP direction is decreased and the OOP lamellar spacing is increased. The OOP lamellar spacing positively correlates with an increase in Jsc and FF. This can be attributed to better OOP charge transport along the thickness of the film, which is beneficial in a solar cell architecture. This is in agreement with the positive correlations between both electron and hole mobilities and OOP lamellar spacing. While it is more relevant to look at these features in the blend films, the scattering profiles of the lamellar and π–π stacking for PTB7-Th largely overlap with those for PDI-V-Cx, making it challenging to isolate each of the components’ morphological evolution separately in the blends.
![]()  | ||
| Fig. 6 2-D GIWAXS patterns of PTB7-Th:PDI-V-Cx blend films: (a) PTB7-Th:PDI-V-C4, (b) PTB7-Th:PDI-V-C5, (c) PTB7-Th:PDI-V-C6, (d) PTB7-Th:PDI-V-C7 and (e) PTB7-Th:PDI-V-C8. | ||
:
1). To completely dissolve the polymers, the solutions were stirred on a hot plate at 110 °C for at least 3 hours. The active layers were spin-coated from the cooled polymer solution on the ZnO coated substrates in a glove-box under an N2 atmosphere at 1700 rpm to obtain a thickness of ∼120 nm. The blend films were then transferred to a vacuum chamber of a thermal evaporator in the same glove-box. At a vacuum level of ≤3 × 10−6 Torr, a thin layer (∼7 nm) of V2O5 was deposited as the anode interlayer, followed by the deposition of 100 nm of Al as the top electrode. All cells were encapsulated using epoxy inside the glove-box. The device J–V characteristics were measured under AM 1.5G (100 mW cm−2) using a Newport solar simulator. The light intensity was calibrated using a standard Si diode (with a KG5 filter purchased from PV Measurement) to bring spectral mismatch to unity. The J–V characteristics were recorded using a Keithley 2400 source meter unit. Typical cells have a device area of 5.9 mm2, which is defined by a metal mask with an aperture aligned with the device area. EQEs were measured using a Newport EQE system equipped with a standard Si diode. Monochromatic light was generated from a Newport 300 W lamp source.
      
    
  
    Footnote | 
| † Electronic supplementary information (ESI) available: Material syntheses and additional characterization data. See DOI: 10.1039/c6qm00355a | 
| This journal is © the Partner Organisations 2017 |