Applications of functional fullerene materials in
polymer solar cells
Yu-Ying Lai, Yen-Ju Cheng* and Chain-Shu Hsu*
Bulk heterojunction (BHJ) polymer solar cells (PSCs) on the basis of polymer:fullerene blends have delivered numerous impressive results in the last decade and thus have drawn much attention of the scientific and industrial communities. However, BHJ PSCs often suffer from several intrinsic problems, such as the low open-circuit voltage (Voc), energetically-unfavorable and chemically-incompatible interfaces, and morphological phase separation of the polymer:fullerene blend driven by thermal treatment, hindering further advancement. In this review, we summarize the recent progress on applications of fullerene-based materials in the BHJ PSCs. Bis-adduct fullerenes possessing high-lying LUMO energy levels can substitute for mono-substituted fullerenes and higher Vocvalues can thus be achieved. Incorporation of an n-type fullerene interlayer between inorganic and organic active layers in either inverted or conventional PSCs can help in minimization of charge recombination losses at the interface and improvement of charge transfer from the active layer to the inorganic layer, thus resulting in superior PCEs. The optimal active-layer morphology can be maintained through several approaches, such as in situ polymerization of cross-linkable fullerenes, suppression of large-scale PC61BM aggregation by a polyethylene glycol-functionalized fullerene, supramolecular pentafluorophenyl-fullerene stabilization, and light-induced oligomerization of fullerenes. The morphological phase separation of the active-layer materials can thus be minimized. The research of employing fullerene materials will continue playing a critical role in BHJ PSCs. With further investigation and advancement in thisfield, it is expected that the ultimate goal of commercialization can be realized.
Broader context
Since bulk heterojunction solar cells were developed in 1995, fullerene derivatives have been dominating in n-type materials due to their low-lying lowest unoccupied molecular orbital (LUMO) energy levels, fast photo-induced electron-transfer and high intrinsic electron mobility. The intrinsic molecular prop-erties of fullerene materials can be custom-tailored by covalently incorporation of functional groups on the buckyballs. The electronic and steric effects exerted by the attached adducts could be used to modulate the optical properties, frontier orbitals, intermolecular interaction, and recrystallization of the fullerenes. Double functionalization of fullerenes has been an effective strategy to arise the LUMO energy levels with greater Vocvalues of the devices. With the aid of the
functional groups to ensure solution processability, the fullerene materials can also act as an electron-selective buffer layer in either regular or inverted devices. More importantly, the functional groups can form chemical interaction/bonding with the metal oxides or electrodes to strengthen interfacial characteristics and facilitate electron-transportation, leading to the improvement of the device performance. Through covalent cross-linking or noncovalent interactions to suppress the aggregation of fullerenes, the optimized morphology of active layers can be permanently preserved against long-term thermal heating. The incorporation of well-designed functional groups has provided fullerene derivatives with versatile functions to achieve highly efficient and stable solar cells.
1.
Introduction
Research on polymer solar cells (PSCs) consisting of organic p-type polymer (donor) and n-p-type fullerene (acceptor) semi-conductors has attracted tremendous scientic and industrial interest in recent years.1,2 The general working principle for PSCs rst involves photoexcitation of the donor material by absorption of light energy to generate excitons. This
Coulomb-correlated electron–hole pair diffuses to the donor–acceptor (D– A) interface, where exciton dissociation occurs. Since the limited lifetime scale only allows excitons to diffuse a short distance (5–14 nm),3–8donor excitons created remote from the heterojunction interface decay to the ground state before they reach the D–A interfaces, leading to the loss of absorbed photons and quantum efficiency. In order to increase the D–A interfacial area for efficient charge separation, bulk hetero-junction (BHJ) is thus adopted to establish an interpenetrating network of donor and acceptor materials.1,2
A conventional BHJ PSC with an active layer sandwiched by a low-work-function aluminum cathode and a hole-conducting Department of Applied Chemistry, National Chiao Tung University, 1001 Ta Hseuh
Road, Hsin-Chu, 30010 Taiwan. E-mail: yjcheng@mail.nctu.edu.tw; cshsu@mail. nctu.edu.tw
Cite this: Energy Environ. Sci., 2014, 7, 1866 Received 12th September 2013 Accepted 6th February 2014 DOI: 10.1039/c3ee43080d www.rsc.org/ees
Environmental
Science
PERSPECTIVE
Published on 06 February 2014. Downloaded by National Chiao Tung University on 25/12/2014 02:57:57.
View Article Online
poly(3,4-ethylenedioxythiophene):poly(4-styrenesulfonic acid) (PEDOT:PSS) layer on top of an indium tin oxide (ITO) substrate is the most widely used and investigated device conguration. The combination of poly(3-hexylthiophene) (P3HT) as electron donor and [6,6]-phenyl-C61 butyric acid methyl ester (PC61BM) as electron acceptor in the active layer represents one of the most efficient BHJ solar cells with power conversion efficiencies (PCEs) approaching 5%.9–13 However, in order to achieve the ultimate goal of commercialization, further optimization is utterly required. To begin with, we will describe the most critical difficulties encountered in the research of the PSCs and discuss the potential solutions towards these challenges.
1.1 Low open-circuit voltage (Voc)
One of the most noticeable restrictions for numerous poly-mer:fullerene-based devices is the low open-circuit voltage (Voc),
limiting further advancement of device performance. It has been demonstrated that the magnitude of Vocvalue shows a linear relationship with the energy difference between the highest occupied molecular orbital (HOMO) energy level of polymers and the lowest unoccupied molecular orbital (LUMO) level of fullerenes.14–16 From the perspective of adjusting molecular energy level, either lowering the HOMO energy level of the p-type polymer (donor) or raising the LUMO energy level of the n-type fullerene (acceptor) could in principle reach larger Vocvalues. Given that fullerene derivatives have been regarded as irreplaceable n-type materials for high-performance BHJ solar cells, developing new fullerene-based materials that possess intrinsically high-lying LUMO energy levels would be a more practicable alternative to optimize the Vocvalue.
1.2 Energetically unfavorable and chemically-incompatible interfaces
Along with high performance, long-term stability is another primary area of concern for PSCs. Rapid oxidation of the low-work-function Al cathode and etching of ITO by the acidic PEDOT:PSS layer are the most common causes of instability in conventional unencapsulated solar devices. An effective approach to solving these problems and improving the device lifetime is to fabricate inverted PSCs.17–20 By reversing the polarity of charge collection in a regular cell, air-stable Ag combining with an adjacent hole-transporting layer, such as PEDOT:PSS or MoO3, can substitute for air-sensitive Al as the anodic electrode for efficient hole collection. In such an inver-ted conguration, it is oen required to insert inorganic metal oxide (TiOxor ZnO) between ITO and the active layer to function as an electron-selective contact.21–25 However, the energetical unfavorability and chemical incompatibility existing in inter-faces may lead to the depreciation of device performance.26For instance, surface hydroxyl groups on metal oxide are known to
Yen-Ju Cheng received his PhD degree in chemistry from the National Taiwan University in 2004 under supervision of Prof. Tien-Yau Luh. Aer spending another year as a postdoc with Prof. Luh at NTU, he joined Prof. Alex K.-Y. Jen's group as a post-doctoral researcher at the University of Washington in 2005. In the summer of 2008, he joined the Department of Applied Chemistry, National Chiao Tung University in Taiwan as an assistant professor and was promoted to associate professor in 2011. His current research interest is focused on the design, synthesis and characterization of organic and poly-meric functional materials for optoelectronic and photovoltaic applications.
Chain-Shu Hsu received his Ph.D.
degree from Case Western
Reserve University in 1987 and conducted his postdoctoral work at National Tsing Hua University
in Taiwan. He joined the
Department of Applied Chemistry of the National Chiao Tung University, Taiwan in 1988 as an associate professor and was promoted to full professor in 1991. Currently he is serving as a vice president and chair professor of the National Chiao Tung University. His research interests include liquid crystalline polymers and conjugated polymers, polymer light-emitting diodes, and organic photovoltaics. He received the Excellent Research Award of National Science Council, Taiwan in 1994, the Franco-Taiwan Scientic Award on nano-materials in 2006, the Teco and Hou Chin Tui Awards in 2007, and the Academic Award of the Ministry of Education, Taiwan in 2008. Yu-Ying Lai received his diploma
and Master degree in Chemistry from National Taiwan Normal University and National Taiwan University, respectively. He obtained his PhD degree from the department of chemistry and applied biosciences at ETH Z¨urich, Switzerland under the supervision of Prof. Peter Chen. Aerwards, he joined Prof. Yen-Ju Cheng's group in National Chiao-Tung University as a Postdoctoral fellow. His research interest focuses on the use of experimental and computational approaches to improve the organic-solar-cell technology.
act as electron traps, being harmful to the charge transport at the interface between the metal oxide and the organic layer.27–29 Interfacial modication by incorporating an additional n-type fullerene interconnection layer between the inorganic metal oxide and the organic active layer can provide a practical solu-tion for improving the inverted device performance by modu-lating the electronic and orbital interactions at the upper and lower interfaces.30 Moreover, it can also be applied to the conventional conguration. The improvement of device char-acteristics can be achieved by installation of a fullerene-based interconnection layer between the active layer and the cathode, being expected to increase the charge-transporting efficiency.31–33
1.3 Morphology evolution
Prudently controlling the phase separation between two components in the bulk to reach ideal D–A morphology plays an signicant role in the success of various polymer:fullerene systems. For example, thermal annealing of a P3HT:PC61BM composite drives P3HT to reorganize and self assemble, resulting in the enhancement of crystallinity of P3HT. Through this morphological evolution, a featureless and homogeneous blend gradually reaches a bicontinuous interpenetrating D–A network with optimal D–A domain size.9Accordingly, maximal D–A interfacial area for efficient charge separation and optimal D–A arrangement for superior charge transportation can be achieved.34Unfortunately, this optimal morphology is a kineti-cally trapped intermediate which readily moves toward a more thermodynamically stable state if thermal annealing at elevated temperatures is applied constantly.35–40As illustrated in Fig. 1, the P3HT domain continues to undergo further crystallization, while the spherical PC61BM, with its high molecular mobility, tends to diffuse out of the polymer matrix and aggregate into larger clusters.38,39,41 Such progressive phase segregation between P3HT and PC61BM eventually leads to the reduction of the percolating electron transport pathways within the mixed regions and thus adversely augments charge recombination losses. Generally speaking, given that a photovoltaic device must be exposed to long-term sunlight irradiation, the accu-mulated heat generated by long-term sunlight irradiation may destroy the optimal D–A morphology and deteriorate the device
performance.42 To reduce the crystallinity of polymers43 or fullerene derivatives44,45can help in maintaining the active-layer morphology. However, from the perspective of general appli-cability, stabilizing the optimal D–A morphology by introducing a compatibilizer to impart secondary interactions between donor and acceptor constituents or in situ chemical cross-link-ing between components in active layers are a more straight-forward strategy. In comparison to the plentiful p-type materials, n-type materials are still dominated by fullerene-based derivatives. Therefore, incorporation of compatibilizer or cross linker moieties into fullerene-based materials is consid-ered to be a feasible approach to stabilize the active-layer morphology.
In view of the great potential and applicability of the fullerene derivatives in the further improvement of the BHJ device performance and long-term stability, we review the recent research progress on applications of fullerene-based materials in the BHJ PSCs. The discussion will be focused on three topics: bis-adduct fullerenes, fullerene-based interlayer modication, and morphological stabilization. It should be noted that the synthetic complexity and efficiency are important issues for commercialization. The synthetic approaches for most fullerene materials mentioned in the following sections are rather easy and straightforward, which is advantageous to the ultimate goal of commercialization.
2.
Bis-adduct fullerenes
Incorporation of electron-donating groups (EDGs) on the peripheral side of pristine C60has been utilized to elevate the LUMO energy level of fullerenes.44,46,47 However, owing to no direct conjugation between the EDG and the core C60, the through-space electronic effect is relatively weak. Therefore, the variations in the LUMO energy level of fullerene derivatives are not signicant enough to affect the Voc. Recently, bis-adduct C60 derivatives have emerged as new materials with higher-lying LUMO energy levels than their corresponding mono-adduct C60 analogues.48–52The second functionalization on the core struc-ture of the mono-substituted C60further reduces the p-conju-gation and electron delocalization in the C60. The formed bis-adduct C60 derivatives can in principle have larger
Fig. 1 Morphological evolution of the P3HT:PC61BM blend under thermal treatment. Reprinted with permission from ref. 129. Copyright 2013, Wiley-VCH Verlag GmbH & Co. KGaA.
electrochemical reduction potentials and thus higher-lying LUMO levels.
2.1 BisPCBM
BisPCBM is the bis-adduct analogue of PC61BM (Scheme 1). The LUMO level of bisPCBM is ca. 100 mV higher than that of PC61BM. Owing to its structural similarity to PC61BM, bisPCBM was tested in conventional PSC devices.48A device on the basis of ITO/PEDOT:PSS/P3HT:bisPCBM (1 : 1.2, w/w)/Sm/Al cong-uration gave a PCE of 4.5% (Entry 1, Table 1). This PCE value is about a factor of 1.2 larger than that of the reference P3HT:PC61BM cell (3.8%).48As expected, this improvement is mainly due to the increase of the Vocvalue.
2.2 ICBA, IC70BA, and IPCBM
Subsequent to the development of bisPCBM, Li and co-workers designed and synthesized another C60 bis-adduct, ICBA (Scheme 1).49The LUMO level of ICBA is raised by 0.17 eV and 0.07 eV, respectively, in comparison to that of PC61BM and bisPCBM. The PCE value for the P3HT:ICBA blend was initially determined to be 5.44% based on the ITO/PEDOT:PSS/ P3HT:ICBA (1 : 1, w/w)/Ca/Al conguration (Entry 2, Table 1). By solvent and thermal annealing of the P3HT:ICBA blend, the PCE value was further improved to 6.48% (Entry 3, Table 1).51Again, in comparison to the P3HT:PC61BM reference cell (Entry 4, Table 1),49the improvement in PCE mainly comes from the signicant increment in Voc.
It is acknowledged that C60 derivatives have weak visible-light absorption than their C70 analogues. An indene-C70 bis-adduct IC70BA (Scheme 1) was thus developed in order to enhance the absorption ability of the active layer.53As expected, IC70BA has enhanced visible-light absorption and its LUMO energy level is ca. 0.19 V higher than that of PC61BM. Initially, a
device based on P3HT:IC70BA furnished a high Vocof 0.84 V and a PCE of 5.64% (Entry 5, Table 1).53Further device optimization resulted in a PCE of 6.69% (Entry 6, Table 1) with 3-methyl-thiophene as a processing additive54and 6.68% (Entry 7, Table 1) with MoO3to replace PEDOT:PSS as an anode buffer layer.55 More importantly, when 1-chloronaphthalene was adopted to be the processing additive for the P3HT:IC70BA blend, an impressive PCE of 7.40% (Entry 8, Table 1) was obtained.56
A PC61BM derivative, indene-PCBM (IPCBM, Scheme 1) was also developed.57 It is highly soluble in common organic solvents and its LUMO energy level is 0.12 eV higher than that of PC61BM. In comparison to ICBA and IC70BA, the slightly lower LUMO energy level of IPCBM is benecial for charge separation from donor to acceptor, especially for small bandgap polymers. A device based on P3HT:IPCBM delivered a Vocof 0.72 V and a PCE of 4.39% (Entry 9, Table 1).57 By installing an electron transportation layer of PFCn6:K+(Scheme 1) between the active layer and the cathode, the PCE was further improved to 6.63% with a Vocof 0.86 V, a Jscof 10.93 mA cm2, and an FF of 70.6% (Entry 10, Table 1).58
2.3 Bis-TOQC
A thieno-o-quinodimethane functionalized C60 bisadduct (bis-TOQC, Scheme 1) also shows potential in enhancing the PSC performance.59 A conventional device based on the ITO/ PEDOT:PSS/P3HT:bis-TOQC (1 : 0.6, w/w)/Ca/Al conguration delivered a good PCE of 5.1% under simulated 86 mW cm2AM 1.5G illumination (Entry 11, Table 1). Compared to a P3HT:PC61BM-based regular device, the performance enhancement mainly comes from the increment of the Vocvalue that is directly linked to the high LUMO energy level of bis-TOQC. It should be noted that the attempt to install solubilizing groups at the 5-position of thiophene of bis-TOQC degraded the
Scheme 1 Structures of bisPCBM, ICBA, IC70BA, IPCBM, PFCn6:K +
, bis-TOQC, and DMPCBA.
device efficiency signicantly. The presence of aliphatic side chains may hamper the intermolecular stacking between fullerenes and thus result in inferior electron transport ability, bringing negative effect on the PSC performance.60
2.4 DMPCBA
The function of bis-adduct fullerenes is not merely limited in the increment of Voc. We recently developed a di(4-methyl-phenyl)methano-C60bis-adduct, DMPCBA (Scheme 1).61Similar to other bis-adduct fullerenes, DMPCBA has a higher LUMO energy level by ca. 100 mV over PC61BM. Noticeably, differential scanning calorimetry (DSC) measurements revealed that DMPCBA doesn't exhibit melting thermal transition, while a sharp melting point at 280C was detected for PC61BM. The absence of the melting transition for DMPCBA indicates that it is an amorphous glass that has no tendency toward thermal-driven crystallization. Wudl and co-workers have demonstrated that the planes of the two phenyl rings in diphenylmethano-fullerene prefer to lie parallel to the diphenylmethano-fullerene surface.62 There-fore, the geometry of the two germinal diphenyl groups in DMPCBA can sterically protect and shield the core C60structure from severe intermolecular aggregation, rendering it intrinsi-cally amorphous and highly soluble. It can thus be anticipated that progressive phase segregation to micrometer-sized D–A domains may thus be alleviated with DMPCBA as the donor material in the active layer.
The possibility of employing DMPCBA in BHJ solar devices was then examined.61 A conventional device based on ITO/ PEDOT:PSS/P3HT:DMPCBA (1 : 1.2, w/w)/Ca/Al exhibited a Voc of 0.87 V, a Jscof 9.05 mA cm2, and a high FF of 65.5%, yielding a high PCE of 5.2% (Entry 12, Table 1). Yet again, the improved PCE in the P3HT:DMPCBA cell to that of the P3HT:PC61BM reference device (Entry 13, Table 1) mainly results from the signicant enlargement of Voc. Furthermore, the morphological stability was also tested. The devices were isothermally heated
at 160C for 10 and 20 h prior to the deposition of the top electrode. The efficiency of the P3HT:PC61BM reference device decreased dramatically, from 3.9% to 0.7% aer 20 h of isothermal heating. In sharp contrast, the DMPCBA-based cell exhibited much more stable device characteristics and the PCE smoothly degraded to 4.7% aer 10 h of heating and to 4.2% over 20 h of heating. Optical microscopy revealed that thermal annealing of the P3HT:PC61BM blend induced a severe macro-graphic alteration, forming needle-shaped PC61BM crystals hundreds of micrometers in length.63–65 Nevertheless, the morphology of the P3HT:DMPCBA (1 : 1.2, w/w) blend remains almost unchanged before and aer the thermal annealing. The high thermal stability is ascribed to the amorphous nature of DMPCBA that preserves the device characteristics.
3.
Fullerene-based interlayer
modi
fication
Hashimoto and co-workers have demonstrated that self-orga-nization of a uorinated PC61BM buffer layer on top of a P3HT:PC61BM active layer is able to improve the photovoltaic performance of a regular PSC.30Moreover, Jen and co-workers have synthesized a series of C60 derivatives having acid or alcohol groups, which can form a self-assembled monolayer (SAM) on the TiOxor ZnO surface in an inverted PSC and the corresponding device efficiency was then improved.66–68 Subse-quently, much research effort has evolved in the area of inter-facial modication by incorporating an additional n-type fullerene interlayer between inorganic and organic active layers. In this section, we will discuss the recent development on the fullerene-based interfacial modication in regular as well as inverted devices. The discussions are classied according to the functional groups covalently attached to the fullerenes. Gener-ally speaking, the capabilities of these fullerenes can be sorted into two categories: one is to enhance the interfacial charge-Table 1 Device characteristics of the P3HT-based PSCs with different acceptors
Entry Active layer
Weight ratio
(in wt%) Voc(V) Jsc(mA cm2) FF (%) PCE (%)
Electron mobility (cm2V1s1) 1 P3HT/bisPCBMa,48 1 : 1.2 0.72 9.14 68.0 4.50 7.0 104 2 P3HT/ICBAb,49 1 : 1 0.84 9.67 67.0 5.44 N/Ah 3 P3HT/ICBAb,c,51 1 : 1 0.84 10.61 72.7 6.48 N/A 4 P3HT/PC61BMb,49 1 : 1 0.58 10.8 62.0 3.88 N/A 5 P3HT/IC70BAb,53 1 : 1 0.84 9.73 69.0 5.64 N/A
6 P3HT/IC70BAb,e,54 1 : 1 0.86 10.79 72.1 6.69 N/A
7 P3HT/IC70BAd,55 1 : 1 0.85 10.61 74.1 6.68 N/A 8 P3HT/IC70BAb,f,56 1 : 1 0.87 11.35 75.0 7.40 N/A 9 P3HT/IPCBMb,57 1 : 1.5 0.72 9.49 64.0 4.39 N/A 10 P3HT/IPCBMg,58 1 : 1 0.86 10.93 70.6 6.63 N/A 11 P3HT/bis-TOQCb,j,59 1 : 0.6 0.86 7.70 66.0 5.10 N/A 12 P3HT/DMPCBAb,61 1 : 1.2 0.87 9.05 65.5 5.20 9.0 105i 13 P3HT/PC61BMb,61 1 : 1 0.60 9.65 67.4 3.90 1.3 104i
aITO/PEDOT:PSS/P3HT:bisPCBM/Sm/Al.bITO/PEDOT:PSS/P3HT:acceptor/Ca/Al.cThermal annealing of the active layer at 150C for 10 min.
dITO/MoO
3/P3HT:acceptor/Ca/Al. eWith 3 vol% 3-methylthiophene as the processing additive. fWith 3 vol% 1-chloronaphthalene as the processing additive.gITO/PEDOT:PSS/P3HT:acceptor/PFCn6:K+/Ca/Al.hN/A¼ not available. iBased on ITO/Al/P3HT:acceptor/Ca/Al device by using SCLC theory.jDevice parameters were characterized under simulated 86 mW cm2AM 1.5G illumination.
transport efficiency and the other is to modify the work function of cathodes.
3.1 Acid and alcohol groups
It was found that the C60-SAMs with various anchoring groups, i.e., carboxylic acid (C60-acid, PCBM-acid, and C60-acid_2), phosphonic acid (C60-PA), and catechol (C60-catechol), can be formed onto the surface of ZnO or TiOx by simple solution-based process (Scheme 2).66–68 For instance, inverted P3HT:PC61BM-based BHJ solar cells with intercalation of the C60-acid or C60-catechol SAM between the active layer and ZnO were fabricated. Under optimized SAM assembly conditions, the device efficiency improved from 3.47 (without C60SAM) to ca. 4.4% (with C60-acid) and 4.19% (with C60-catechol) (Entry 1–3, Table 2).68The PCE improvement when modied with a C
60 -SAM is mainly attributed to the enhancement of Jsc and FF (Entry 1–3, Table 2). The C60-SAM interlayer acts as an electron selective and hole blocking layer, and its presence can help in minimization of charge recombination losses at the interface owing to the improved charge transfer from the active layer to the ZnO layer. The larger photocurrent can thus be achieved.
Not merely limited in inverted solar devices, acid or alcohol functionalized fullerenes can also be utilized in conventional devices. Nevertheless, potential difficulties derived from solu-tion processing of the fullerene-based interlayer need be taken into account. Active-layer organic materials, e.g., P3HT:PC61BM, are usually soluble in common organic solvents. Organic solvents that dissolve the active-layer materials must be avoided when incorporating the fullerene-based interlayer on top of the active layer in the regular devices. Otherwise, severe intermixing between the two layers may occur, leading to the structural deterioration of the active layer. Accordingly, we developed two hydrophilic fullerene derivatives, bis(triethylene glycol) malo-nate C60 (EGMC-OH) and bis(triethylene glycol phthalic acid)-malonate C60(EGMC-COOH), with hydroxyl and acid groups as the outmost moieties respectively (Scheme 2).69With the aid of triethylene glycol-based side chains, these C60 materials with enhanced hydrophilic nature can be dissolved in the polar 2-ethoxyethanol/H2O solvent and then spin-cast to form an interlayer without destroying the underneath active layer. BHJ solar devices on the basis of ITO/PEDOT:PSS/P3HT:PC61BM/ interlayer/Ca/Al were then fabricated. The power conversion efficiency (PCE) was increased from 3.61% to 3.71% and 3.80% by using EGMC-OH and EGMC-COOH as the electron-selective modication layer (EML) materials, respectively (Entry 4–6, Table 2). Moreover, the electron transporting properties of EGMC-OH and EGMC-COOH can be further enhanced by the incorporation of alkali carbonates as the n-dopants. Regardless of the counter ions, Cs2CO3, Li2CO3, as well as K2CO3 are all effective in obtaining high PCEs. The device using EGMC-COOH doped with 40 wt% Li2CO3 achieved a highest PCE of 4.29% (Entry 7, Table 2). The alcohol/water soluble EGMC-COOH is applicable to other conventional devices as well. With the incorporation of the EGMC-COOH EML doped with 40 wt% Li2CO3, the PCDCTBT-C8 (Scheme 2):PC71BM-based device exhibited a superior PCE of 4.51%, which outperformed the
corresponding non-modied device with a PCE of 3.63% (Entry 8–9, Table 2).
3.2 Phosphoric ester groups
An alcohol soluble fullerene material, B-PCPO (Scheme 2) was developed and used in PSCs owing to the fact that it has the properties of good alcohol processability, ITO modication function, and good electron transporting ability.70An inverted device in the absence of EML by using the ITO/PCDTBT (Scheme 2):PC71BM/MoO3/Al conguration gave a PCE value of 4.83% with a Vocof 0.71 V, a Jscof 10.20 mA cm2, and an FF of 56.1% (Entry 10, Table 2). By using B-PCPO as the EML between ITO and the active layer (PCDTBT:PC71BM), the corresponding device delivered a high PCE of 6.20% with a Vocof 0.89 V, a Jscof 9.50 mA cm2, and an FF of 61.7% (Entry 11, Table 2). As listed in Table 2, with the assistance of B-PCPO, Vocand FF were both enhanced, leading to a higher overall performance, even though Jsc was slightly decreased. X-ray photoelectron spectroscopy (XPS) revealed that the work function of ITO is decreased by 0.4 eV by introduction of a thin B-PCPO layer, which may result from the dipole formation in the ITO/active layer inter-face.71,72 The lower work function of ITO is benecial for increasing the built-in potential of PSCs, in turn leading to the enhanced Voc.73,74Moreover, the decrease of ITO work function can also offer better energy level alignment with the LUMO level of PC71BM and thereby facilitate the electron transporting and collection, transforming to the enhancement of FF.
3.3 Polyethylene glycol groups
High work function metals, such as Ag, have been used as cathodes to substitute for Al in conventional PSCs on account of their high stability to ambient conditions. However, the high work function oen results in low Vocand poor device perfor-mance.74–76 Two polyethylene glycol (PEG)-based fullerene surfactants, ETL-1 and ETL-2 (Scheme 2), were found to be effective in tuning the work function of metal cathode and also increasing the photocurrent generation when used as an interfacial layer between the active layer and the metal cathode.77,78 Besides, the presence of PEG renders these two compounds highly soluble in alcoholic solvents, which is good for utilizing orthogonal solvent-processed strategies.79,80ETL-1 and ETL-2 with the cationic tetrahydro pyrrole and PEG addend can up-shi the work function of Ag by 0.8 eV.78 The device
conguration of ITO/PEDOT:PSS/PIDT-PhanQ (Scheme
2):PC71BM/interlayer/Ag was tested. In comparison to the reference cell without interlayer (Entry 12, Table 2), signicant improvement of PCE was observed for devices with installation of ETL-1 or ETL-2 between the active layer and the Ag cathode (Entry 13–14, Table 2). The presence of ETL-1 or ETL-2 can simultaneously improve Voc, Jsc, and FF, thus resulting in high-performance PSCs using the high work-function Ag cathode.
Calcium is known to be highly reactive to moisture and thermal evaporation under vacuum is an inevitable step for processing Ca in solar devices. A solution processable PEG-modied C60 (PEGN-C60, Scheme 2) was thus proposed to replace Ca in conventional solar devices.81It was tested under the
Scheme 2 (Continued)
ITO/PEDOT:PSS/active layer/interlayer/Al device conguration, where the active layers are P3HT:PC61BM, PBDTTT-C82(Scheme 2):PC71BM, and PBDTTT-C-T83 (Scheme 2):PC71BM and the interlayer is either Ca or PEGN-C60. As listed in Table 2 (Entry 15– 20), the PEGN-C60 based devices all demonstrated comparable device efficiencies to the Ca-based devices. Hence, PEGN-C60can be considered a promising alternative to Ca in conventional PSCs. 3.4 Amine groups
Subsequent to the success of PEGN-C60, an amine-functionalized C60(DMAPA-C60, Scheme 2) was employed as well in regular PSCs to substitute for Ca.84 The tested active-layer materials are P3HT:PC61BM, PBDTTT-C:PC71BM, and PBDTTT-C-T:PC71BM and the corresponding devices using DMAPA-C60as the interlayer
gave comparable PCE values to those fabricated by use of Ca on the basis of ITO/PEDOT:PSS/active layer/interlayer/Al device structure.84(Entry 21–23, Table 2) Moreover, the combination of DMAPA-C60with high work-function metal cathodes, such as Ag, also yielded decent device efficiency (Entry 24, Table 2), offering moreexibility for cathode selection in device fabrication. 3.5 Styrene groups
To realize a multilayer inverted device by cost-effective solution processing, the deposited fullerene derivative interlayer must possess sufficient resistance against the organic solvent used in processing the sequential active layer to prevent interfacial erosion.85,86 Development of a fullerene material that can be solution-deposited on a substrate and readily undergo chemical Scheme 2 Structures of fullerene derivatives– C60-acid, PCBM-acid, C60-acid_2, C60-PA, C60-catechol, EGMC–OH, EGMC–COOH, B-PCPO, ETL-1, ETL-2, PEGN-C60, DMAPA-C60, PCBS, PCBO, TSMC, PCBSD, C-PCBSD, PCBOD, C-PCBOD, PCBE-OH, and ZnO–C60precursor and low band-gap polymers– PCDCTBT-C8, PCDTBT, PIDT-PhanQ, PBDTTT-C, PBDTTT-C-T, PCPDTBT, PDITTDTBT, PTB7, and PTB7-Th.
cross-linking to form a robust and immobilized layer could be a straightforward and ideal strategy. Styrene is known to be a superb thermally curable group because it can undergo rapid polymerization in the solid state to form polystyrene without the use of any initiators.87–89To this end, a fullerene derivative, [6,6]-phenyl-C61-butyric styryl dendron ester, PCBSD (Scheme 2), that contains a small dendron functionalized with two thermally cross-linkable styryl groups was synthesized by our group and the corresponding cross-linked PCBSD (C-PCBSD, Scheme 2) was then integrated into inverted solar devices.72A solution of PCBSD was spin-coated on top of the ZnO surface, followed by thermal curing at 160 C for 30 min. This
cross-linked network produced a robust and adhesive thinlm with sufficient solvent resistance. A multilayer inverted solar cell based on the ITO/ZnO/C-PCBSD/P3HT:PC61BM (1 : 1, w/w)/ PEDOT:PSS/Ag conguration was fabricated and gave a PCE of 4.4% (Entry 26, Table 2). A reference device without the C-PCBSD interlayer gave a PCE of 3.5% (Entry 25, Table 2). With the additional interlayer of C-PCBSD, the PCE was substantially improved by 26%, which results from the simultaneously enhanced Voc, Jsc, and FF relative to the device without inter-layer. The use of the C-PCBSD interlayer provides an extra P3HT/ C-PCBSD interface area for ultrafast exciton dissociation, which is more efficient than the P3HT/ZnO interface.90,91Moreover, the Table 2 Device characteristics of the polymer-based solar cells with interlayer modification
Entry Active layer Interlayer
Weight ratio
(in wt%) Voc(V) Jsc(mA cm2) FF (%) PCE (%)
1 P3HT/PC61BMa,68 None 1 : 0.7 0.60 10.07 57.7 3.47 2 C60-acid 1 : 0.7 0.62 11.17 64.1 4.40 3 C60-cateholb 1 : 0.7 0.61 11.21 61.0 4.19 4 P3HT/PC61BMc,69 None 1 : 1 0.60 9.02 67.0 3.61 5 EGMC-OH 1 : 1 0.60 9.43 66.0 3.71 6 EGMC-COOH 1 : 1 0.60 9.61 66.0 3.80
7 EGMC-COOH doped with 40% Li2CO3 1 : 1 0.60 10.90 66.0 4.29
8 PCDCTBT-C8:PC71BMc,69 None 1 : 3 0.72 8.72 57.0 3.63
9 EGMC-COOH doped with 40% Li2CO3 1 : 3 0.72 11.10 56.0 4.51
10 PCDTBT:PC71BMd,70 None 1 : 4 0.71 10.2 56.1 4.83
11 B-PCPO 1 : 4 0.89 9.50 61.7 6.20
12 PIDT-PhanQ:PC71BMe,78 None 1 : 3 0.74 10.92 57.0 4.61
13 ETL-1 1 : 3 0.87 11.28 64.0 6.28 14 ETL-2 1 : 3 0.88 11.41 66.0 6.63 15 P3HT:PC61BMf,81 Ca N/Ag 0.63 9.01 67.5 3.80 16 PEGN-C60 N/A 0.63 9.21 66.6 3.84 17 PBDTTT-C:PC71BMf,81 Ca N/A 0.73 13.80 60.6 6.13 18 PEGN-C60 N/A 0.74 13.99 60.5 6.22 19 PBDTTT-C-T:PC71BMf,81 Ca N/A 0.79 14.83 62.4 7.28 20 PEGN-C60 N/A 0.79 14.79 63.4 7.45 21 P3HT:PC61BMf,84 DMAPA-C60 1 : 0.8 0.63 9.14 67.8 3.88 22 PBDTTT-C:PC71BMf,84 DMAPA-C60 N/A 0.72 14.08 61.8 6.29 23 PBDTTT-C-T:PC71BMf,84 DMAPA-C60 1 : 1.5 0.79 14.89 62.9 7.42
24 PBDTTT-C:PC71BMe,84 DMAPA-C60 N/A 0.73 14.26 59.5 6.20
25 P3HT:PC61BMa,72 None 1 : 1 0.58 11.60 52.0 3.50
26 C-PCBSD 1 : 1 0.60 12.80 58.0 4.40
27 PCPDTBT:PC71BMa,72 None 1 : 2 0.48 12.10 32.0 1.90
28 C-PCBSD 1 : 2 0.66 14.50 35.0 3.40
29 P3HT:ICBAa,50 None 1 : 1 0.82 10.60 55.0 4.80
30 C-PCBSD 1 : 1 0.84 12.40 60.0 6.22
31 P3HT:ICBAa,105 C-PCBSD nanorods 1 : 1 0.84 12.07 72.3 7.30
32 Planar C-PCBSD 1 : 1 0.84 11.17 66.0 6.20 33 P3HT:PC61BMh,113 None 1 : 1 0.58 10.65 57.76 3.60 34 SA-PCBO 1 : 1 0.60 11.14 60.73 4.10 35 C-PCBOD 1 : 1 0.61 12.25 61.26 4.50 36 P3HT:PC61BMi,122 None 1 : 1 0.58 9.87 55.00 3.20 37 SA-C-TSMC 1 : 1 0.60 10.42 61.70 3.90 38 PDITTDTBT:PC71BMh,122 None 1 : 4 0.87 9.94 50.27 4.30 39 SA-C-TSMC 1 : 4 0.91 11.54 55.02 5.80 40 PTB7:PC71BMj,124 ZnO 1 : 1.5 0.70 13.75 69.00 6.65 41 ZnO-C60 1 : 1.5 0.73 15.41 73.00 8.21 42 PTB7-Th:PC71BMj,124 ZnO 1 : 1.5 0.79 14.02 69.10 7.64 43 ZnO-C60 1 : 1.5 0.80 15.73 74.30 9.35
aITO/ZnO/interlayer/active layer/PEDOT:PSS/Ag.bThe C
60-SAM was prepared by solution immersion.68cITO/PEDOT:PSS/active layer/Ca/Al.dITO/ interlayer/active layer/MoO3/Al. eITO/PEDOT:PSS/active layer/interlayer/Ag. fITO/PEDOT:PSS(40 mm)/active layer/interlayer/Al.gNot available. hITO/TiO
x/interlayer/active layer/MoO3/Ag.iITO/TiOx/interlayer/active layer/PEDOT:PSS/Ag.jITO/interlayer/active layer/MoO3/Ag.
LUMO energy level of C-PCBSD (3.8 eV) is located between the LUMO of P3HT (3.3 eV) and the conduction band of ZnO (4.4 eV). Therefore, the C-PCBSD can function as an energy gradient intermediate, so that electrons can be efficiently transported to the ZnO through an energetically downhill cascade pathway.92
The device stability was also examined.72 The PCEs of the unencapsulated inverted solar cell devices were periodically measured for 35 days to monitor their long-term stability. With the presence of the C-PCBSD interlayer, the device lifetime was enhanced by 1.5 time. Device degradation is known to be related to leakage current caused by the formation of hot spots in bulk or at the interface.93The coverage of a three-dimensional C-PCBSD network on ZnO might passivate the hot spots generated in the ZnO to suppress the leakage current, thus improving the opera-tional lifetime. Addiopera-tionally, without this interlayer as a dense and robust spacer, gradual mutual penetration between the ZnO and the active layer may alter their chemical or physical prop-erties as the operational time increases.
This strategy was also applied to other blending systems. A inverted device based on ITO/ZnO/C-PCBSD/PCPDTBT94,95 (Scheme 2):PC71BM (1 : 2, w/w)/PEDOT:PSS/Ag was fabricated and exhibited a Vocof 0.66 V, a Jscof 14.50 mA cm2, and an FF of 35%, leading to a PCE of 3.4% (Entry 28, Table 2), which has a considerable improvement of 78% compared to the reference device without the C-PCBSD interlayer (Entry 27, Table 2). Moreover, the utilization of n-type bis-adduct fullerenes in an inverted device with the C-PCBSD interlayer modication was also carried out.50ICBA was selected as the n-type materials on account of its high-lying LUMO level that has been demon-strated to be benecial for the enhancement of Voc.49A device on the basis of ITO/ZnO/C-PCBSD/P3HT:ICBA (1 : 1, w/w)/ PEDOT:PSS/Ag conguration exhibited a Vocof 0.84 V, Jscof 12.4 mA cm2, and FF of 60%, achieving a high PCE of 6.22% (Entry 30, Table 2), which is a 29% increase over the reference device without the C-PCBSD interlayer (Entry 29, Table 2).
To provide a direct path for charge transport while main-taining a large interfacial area in BHJ solar cells, the ideal architecture of the donor and acceptor is the periodic, vertically aligned, and interpenetrating ordered bulk heterojunction (OBHJ).1,96–101 C-PCBSD combined with the anodic aluminum oxide (AAO) template-assisted approach102–104was employed to construct a well-organized nanostructured interface. A device on the basis of ITO/ZnO/C-PCBSD nanorods/P3HT:ICBA/
PEDOT:PSS/Ag conguration was completed (Fig. 2).105Under otherwise identical conditions, a control device, by using the ITO/ZnO/planar C-PCBSD/P3HT:ICBA/PEDOT:PSS/Ag congu-ration (without the C-PCBSD nanostructure), was fabricated for comparison. The device with the C-PCBSD nanorods exhibited an exceptional PCE of 7.3%, with Voc¼ 0.84 V, Jsc¼ 12.07 mA cm2, and FF¼ 72.3% (Entry 31, Table 2). The control device without the C-PCBSD nanostructure exhibited an inferior PCE of 6.2%, with a Vocof 0.84 V, a Jscof 11.17 mA cm2, and an FF of 66% (Entry 32, Table 2). The efficiency enhancement could be ascribed to the fact that the array of C-PCBSD nanorods pene-trating into the BHJ layer offers substantial P3HT/C-PCBSD interfacial area for extra exciton dissociation, thereby gener-ating a higher photocurrent, which was supported by steady-state photoluminescence (PL) measurements.105 Furthermore, the introduction of C-PCBSD nanorods enhanced the electron mobility from 9.4 104 to 2.6 103 cm2 V1 s1 while maintaining the hole mobility value of approximately 3 103 cm2V1s1, thus leading to more balanced charge transport. Such an enhancement in electron mobility has also been demonstrated in previous studies, showing that nanostructured organic semiconductors can possess superior charge-transport properties than their planar counterparts.99,100,107,108
The stability of the devices without encapsulation was also evaluated. The device stability was enhanced signicantly with the help of the C-PCBSD nanorods. As suggested by optical microscopy, the macrographic alteration of the active-layer morphology is the primary reason causing the reduction of effi-ciency of the devices. Considering that a part of the P3HT:ICBA BHJ inltrated into the C-PCBSD nanorods and is thus sur-rounded and conned between the rods, nanostructure-assisted morphological stability could be rationalized by the spatial connement effect, which might suppress the thermal-driven crystallization of ICBA and P3HT upon thermal annealing.
3.6 Oxetane groups
Oxetane functional group has been extensively utilized as an efficient cross-linker for the applications of organic elec-tronics.88,106–111In comparison to the radical polymerization of styrene, cationic polymerization of oxetane is less sensitive to air. Moreover, it is reasonable to postulate that hydroxyl groups on metal oxide can also serve as nucleophiles to open the ring of protonated oxetane, thereby forming a monolayer modier by covalent bonding.112
We developed a PC61BM-based n-type material, [6,6]-phenyl-C61-butyric oxetane dendron ester (PCBOD) (Scheme 2), func-tionalized with a small dendron containing two oxetane groups as the cross linkers.113To induce cationic oxetane ring-opening polymerization, a catalytic amount of photoacid generator (PAG) is required to generate acidic protons upon UV irradia-tion. Infrared spectroscopy suggested that when the polymeri-zation temperature is above the glass transition temperature (Tg) of PCBOD, the two arms of the dendron can provide adequateexibility for the oxetane groups to react in the solid state to furnish cross-linked PCBOD (C-PCBOD) (Scheme 2). A plausible mechanism was proposed to account for the self-Fig. 2 Schematic representation of the nano-structured device
architecture. Reprinted with permission from ref. 105. Copyright 2011, Wiley-VCH Verlag GmbH & Co. KGaA.
assembled monolayer and cross-linking of PCBOD on the TiOx surface (Fig. 3).113,114The anchoring of oxetane groups may take place by nucleophilic attack of the hydroxyl groups on the TiOx surface to open the protonated oxetane rings. Further bidentate anchoring is possible through the etherication condensation. Once the self-assembled monolayer is formed, the intermolec-ular cross-linking can efficiently occur through ring-opening polymerization between the oxetane groups to vertically grow a multi-molecular interlayer.
Subsequently, inverted solar cell devices based on the conguration of ITO/TiOx/interlayer/P3HT:PC61BM (1 : 1, w/w)/ MoO3/Ag was fabricated.113 The device with the C-PCBOD interlayer exhibited a Vocof 0.61 V, a Jscof 12.25 mA cm2, and
an FF of 61.26%, achieving a PCE of 4.5% (Entry 35, Table 2). The device without the C-PCBOD layer delivered an inferior PCE of 3.6% (Entry 33, Table 2). Noticeably, when PCBOD was replaced with PCBO (Scheme 2) to form a self-assembly PCBO monolayer (SA-PCBO), the corresponding device gave a PCE of 4.10% (Entry 34, Table 2), which drops by 10% compared to the C-PCBOD-based device. This result suggests that a dense, robust, and pinhole-free multimolecular interlayer is capable of further strengthening the interface characteristics.
3.7 Trichlorosilane
Trichlorosilane moieties are known to readily undergo facile hydrolysis to produce polysiloxane in ambient air. Because of its high reactivity and wide availability, it is frequently used to implement self-assembly and/or cross-linking for the purpose of surface modication.115–121 We recently have incorporated trichlorosilane functionalities into fullerene derivatives for PSC applications. Bis(2-(trichlorosilyl)propyl)-malonate C60 (TSMC, Scheme 2) containing two trichlorosilane groups was synthe-sized and spin-cast on top of the ITO/TiOxsurface to form a self-assembled/cross-linked network of TSMC (SA-C-TSMC).122 Atomic force microscopy (AFM) was employed to investigate the surface morphology. In AFM images, pronounced nano-structures were found on the surface of the SA-C-TSMC layer with a RMS roughness of 5.26 nm.122 These vertically grown semispherical nanoaggregates comprising the cross-linked TSMC aggregates are ca. 10–30 nm in height and ca. 20–60 nm in diameter. In contrast, the surface of the pristine TiOx exhibited much smooth morphology with a RMS roughness ca. 3.49 nm. A plausible mechanism accounting for the formation of SA-C-TSMC layer on the top of TiOxsurface is proposed in Fig. 4. Upon spin-coating, the nucleophilic hydroxyl groups on Fig. 3 Proposed mechanism for self-assembly and multi-molecular
cross-linking of PCBOD on the TiOxsurface. Reprinted with permis-sion from ref. 114. Copyright 2011, American Chemical Society.
Fig. 4 Hydrolysis of the trichlorosilane groups of TSMC: (a) self-assembled reaction on TiOx, (b) intermolecular cross-linking to form multilayer network. Reprinted with permission from ref. 123. Copyright 2013, American Chemical Society.
the surface can attack the electrophilic silicon atoms of tri-chlorosilanes to spontaneously form a self-assembled mono-layer via the Ti–O–Si linkages (Fig. 4a). In the mean time, intermolecular cross-linking between the TSMC molecules to form a multilayer network through the siloxane Si–O–Si link-ages also proceeds efficiently (Fig. 4b). Considering that hydrolysis of trichlorosilane groups with moisture is a fast and energetically favorable process, it is reasonable to expect that some of the TSMC molecules have started cross-linking and aggregating into nanoparticles suspended in the solution before spin-casting. The precross-linked nanoparticles depos-ited on the surface were integrated into the SA-C-TSMC layer, thereby forming the vertically aligned nanostructures on the surface. The nanoaggregates on the SA-C-TSMC layer can provide extra charge-generating interfacial area and straight electron transport pathways.
A multilayer inverted device with the conguration of ITO/ TiOx/SA-C-TSMC/P3HT:PC61BM (1 : 1, w/w)/PEDOT:PSS/Ag was fabricated. It exhibited a Vocof 0.60 V, a Jscof 10.42 mA cm2, an FF of 61.70%, and a PCE of 3.9% (Entry 37, Table 2), which represents a 22% improvement over the device without the SA-C-TSMC layer (PCE¼ 3.2%, Entry 36, Table 2). This strategy was further applied to the devices incorporating various photoactive materials. A p-type copolymer PDITTDTBT123 (Scheme 2) was chosen to test the general applicability of SA-C-TSMC. The device with the SA-C-TSMC interlayer (ITO/TiOx/SA-C-TSMC/ PDITTDTBT:PC71BM (1 : 4, w/w)/MoOx/Ag) and its correspond-ing control device in the absence of the interlayer were thus fabricated. Compared to the device without SA-C-TSMC (Entry 38, Table 2), the device with SA-C-TSMC exhibited Voc¼ 0.91 V, Jsc¼ 11.54 mA cm2, FF¼ 55.02% and PCE ¼ 5.8% (Entry 39, Table 2), which is a considerable enhancement of 35% in effi-ciency. Furthermore, this experiment also demonstrates that the TSMC interlayer is effective not only for the PC61BM-based system but also for the PC71BM-based system.
3.8 ZnO doped with fullerene materials
The combination of fullerenes and metal oxides in PSCs is an intriguing idea. Lately it has been realized by employing ZnO doped with PCBE-OH (Scheme 2) as the cathode in inverted PSCs.124This fullerene-doped ZnOlm (abbreviated as ZnO-C
60) was prepared by spin coating a mixture of PCBE-OH (0.5 wt%) and Zn(OAc)2 (99.5 wt%) in 2-methoxyethanol and ethanol-amine on ITO, followed by sintering at 180C under ambient conditions. It is believed that a condensation reaction would rst occur between Zn(OAc)2and PCBE-OH to form a Zn-C60 precursor (Scheme 2). Contact angle and XPS measurements suggest that the formed ZnO-C60 lm is fullerene-rich on its surface, indicating that the fullerene molecules are mainly sit-uated at the uppermost part of the ZnO-C60 lm. The HOMO and LUMO energy levels of the ZnO-C60lm were determined to be7.01 and 4.53 eV by UV-Vis spectroscopy. For the pristine ZnOlm, the HOMO and LUMO energy levels are 7.24 and 4.14 eV, respectively. It is evident that the doping of PCBE-OH to ZnO results in smaller energy band gap, accompanied by the upshi of the HOMO energy and the downshi of the LUMO
energy. The observed energy variation also reects with the fact that the fullerenes are located on the surface of the ZnO-C60 lm, thus leading to the change of the frontier-orbital energies. Furthermore, the electron mobility and surface conductivity of the ZnO-C60 lm were evaluated as well and found to be appreciably higher than those of the pristine ZnOlm.
Subsequently, the ITO/interlayer/active layer/MoO3/Ag device conguration was employed to examine the effect brought by the ZnO-C60lm, where the cathodic interlayer is either pristine ZnO or ZnO-C60 lm and the active layer is PTB7125 (Scheme 2):PC71BM or PTB7-Th124(Scheme 2):PC71BM. The device char-acteristics are listed in Table 2. Evidently, the device efficiencies increased signicantly when the cathodic interlayer varied from pristine ZnO to ZnO-C60. The best PCE of 9.35% was achieved by use of the ITO/ZnO-C60/PTB7-Th:PC71BM/MoO3/Ag device structure. As concluded from the attained experimental data, the success of this fullerene-dopant strategy mainly lies in the enhancement of electron collection of the cathode and the diminution of the interfacial series resistance between the cathode and the active layer.
4.
Morphological stabilization
Morphological instability of PSCs is considered to be one of the major obstacles hindering their path toward commercializa-tion. However, the development of practicable strategies to preserve optimal morphology is rather challenging. Fr´echet and co-workers found that a slight decrease of the regioregularity of P3HT to weaken the crystallization-driven phase separation not only retains the device efficiency but also enhances the thermal stability of morphology in the solar cells.43Moreover, Fr´echet et al. used a diblock copolymer containing fullerene and oli-gothiophene pendant groups to serve as a compatibilizer to improve adhesion between the P3HT and PC61BM bound-aries.126 Wudl and co-workers also used similar strategy to adjust the interfacial morphology between P3HT and PC61BM and found a substantial improvement in device performance.127 Lately, several elegant approaches, including in situ polymeri-zation,63 polyethylene glycol (PEG) end-capped fullerene,128 supramolecular pentauorophenyl-fullerene interactions,129 and light-induced oligomerization of fullerenes130 have been utilized in stabilizing the active-layer morphology and simul-taneously retaining the device efficiency, providing insight into the general applicability of morphological stabilization.
4.1 In situ polymerization
One of the most straightforward methods for stabilization of the active-layer morphology is to lock it by covalent bonding.131–135In situ chemical cross-linking in BHJlms right aer the formation of the optimal morphology is an ideal strategy to achieve this goal.131–135 Crosslinkable PCBSD (Scheme 2) with two styryl groups can not only be used in interfacial modication,72but also be utilized in active layers for morphological stabilization.63 Conventional devices on the basis of ITO/PEDOT:PSS/active layer/Ca/Al conguration were manufactured. Various amounts of PCBSD were doped into the
P3HT:PC61BM blend. The blending weight ratio of P3HT to the total n-type materials (i.e., PC61BM plus PCBSD) was xed at 1 : 1 and the relative proportion between the PC61BM and PCBSD was adjusted. Two-stage thermal annealing as the standard procedure for all blending systems was imple-mented.63The spin-coated active-layer blend was initially ther-mally annealed at 110 C for 10 min to develop suitable morphology, followed by thermal annealing at 150 C for another 10 min to trigger in situ polymerization. It was found that by doping small amounts of PCBSD into the P3HT:PC61BM blend (P3HT:PC61BM:PCBSD¼ 6 : 5 : 1 in weight), the initial morphology of the blend can bexed and preserved effectively aer polymerization. The device based on this blend showed highly stable device characteristics, delivering a PCE of 3.70% during 25 h isothermal heating at 150C (Entry 4, Table 3). In sharp contrast, under otherwise identical conditions, the P3HT:PC61BM blend underwent severe phase separation due to the crystallization of P3HT and the aggregation of PC61BM. The PCE of the P3HT:PC61BM-based device dropped dramatically from 4.08% to 0.69% (Entry 1 and 2, Table 3). PCBS (Scheme 2) with only one styryl moiety, was also used as a dopant in the active layer. At a low doping concentration of PCBS in the blend (P3HT:PC61BM:PCBS¼ 6 : 5 : 1 in weight), linearly polymerized PCBS can also stabilize the morphology against heating. This device exhibited a PCE of 3.61% during 25 h heating at 150C (Entry 6, Table 3).63
4.2 Polyethylene glycol (PEG) end-capped fullerene
Fullerenes are usually insoluble in polar solvent, such as water.136,137However, when they are functionalized with hydro-philic PEG groups, the resultant fullerene derivatives oen have high solubility in polar solvents.128A PEG-end capped PC
61BM
derivative, PCBPEG (Scheme 3), was developed and used as a dopant to stabilize the morphology of the P3HT:PC61BM blend. By employing the ITO/PEDOT:PSS/active layer/Al conguration, when the P3HT/PC61BM active-layer blend was doped with 5 wt % of PCBPEG, the corresponding device gave a Vocof 0.60 V, a Jsc of 11.44 mA cm2, an FF of 60.0%, and a PCE of 4.15%. Aer being heated at 60C for 300 h, the device can still maintain 50% of the original PCE value. For comparison, the device without doping of PCBPEG lost more than 85% of its original PCE value aer 300 h heating at 60C. Transmission electron
microscopy reveals that PCBPEG is able to suppress the formation of large-scale PC61BM aggregation at elevated temperatures, thus restraining the P3HT/PC61BM from severe phase separation at elevated temperatures.
4.3 Supramolecular pentauorophenyl-fullerene interactions
It has been reported by Nakamura and co-workers that through the pentauorophenyl (C6F5)-fullerene interaction, 1,4-bis-(pentauorobenzenyl) fullerene assembles into an interlocked one-dimensional zigzag array in the solid state.138 It is intriguing to further utilize this C6F5–C60interaction in stabi-lizing the active-layer morphology in PSCs. To this end, we developed a new PC61BM-based fullerene, [6,6]-phenyl-C61 butyric acid pentauorophenyl ester (PC61BPF, Scheme 3) and tested it in conventional solar devices (ITO/PEDOT:PSS/active layer/Ca/Al).129 A series of active-layer blends with different doping ratio of PC61BPFwere initially thermally annealed at 140
C for 15 min, and then subjected to isothermal heating at 150 C for the longest 25 hours prior to the deposition of top
elec-trodes. It was found that by doping small amounts of PC61BPF into the P3HT:PC61BM blend (P3HT:PC61BM:PC61BPF¼ 6 : 5 : 1
Scheme 3 Chemical structures of PC61BP F
, PC61BP, PCBPEG, and PDTBCDTBT.
Table 3 Photovoltaic parameters of PSCs based on various blends isothermally heated at 150C for different time
Entry Blend
Weight ratio
(in wt%) Time (h) Voc(V) Jsc(mA cm2) FF (%) PCE (%)
1 P3HT/PC61BM 1 : 1 — 0.60 10.26 66 4.08 2 25 0.62 2.27 49 0.69 3 P3HT/PC61BM/PCBSD 6 : 5 : 1 — 0.60 8.44 66 3.32 4 25 0.60 9.35 66 3.70 5 P3HT/PC61BM/PCBS 6 : 5 : 1 — 0.60 9.82 65 3.84 6 25 0.58 9.49 66 3.61
in weight), the device performance showed highly stable device characteristics, delivering a PCE of 3.68% (Entry 4, Table 4) aer 25 h isothermal heating at 150C. In sharp contrast, the PCE of the P3HT:PC61BM blend dropped dramatically from 4.08% (Entry 1, Table 4) to 0.69% (Entry 2, Table 4) aer 25 h isothermal heating. The decreased efficiency is mainly a result of the decrease of Jscvalue. In comparison to PC61BPF, PC61BP (Scheme 3) with non-uorinated phenyl group was synthesized and also used as the doping component in the P3HT:PC61BM blend. By using the same doping ratio (P3HT:PC61BM:PC61BP¼ 6 : 5 : 1 in weight), the device yielded a PCE of 3.45% initially (Entry 5, Table 4). Subsequent to 25 h thermal heating at 150C, it degraded signicantly to 0.44% (Entry 6, Table 4). These results clearly demonstrate that the presence of penta-uorophenyl group in PC61BPFto exert C6F5–C60 interactions plays a decisive role in controlling and stabilizing the morphology against thermal heating.
A rational mechanism for the morphological stabilization of the P3HT:PC61BM:PC61BPFblends is proposed and illustrated in Fig. 5. For the as-cast thin lms with dopant of PC61BPF, diffusion of the fullerene molecules dispersed in the mixed regions is partially restricted by the intermolecular C6F5–C60 interactions, reducing their tendency toward thermal-induced fullerene aggregation (Fig. 5a). Pre-annealing at 140C for 15 min may not be sufficient to achieve optimized morphology, which depends on the loading amount of PC61BPF(Fig. 5b). By further heating at 150C during 25 h, the fullerene molecules in the mixed regions might gradually overcome the supramolec-ular attraction and slowly aggregate into fullerene clusters with concurrent enhancement of P3HT crystallinity (Fig. 5c). DFT
calculations at the wB97XD/6-311G(d,P) level of theory was employed to investigate the stabilization brought by the pen-tauorophenyl group of PC61BPF.129The computational results reveal that the C60core can interact with the pentauorophenyl group to form a stable adduct and this intermolecular interac-tion was estimated to be10.58 kcal mol1. The presence of C6F5–C60interactions is thus supported by DFT calculations.
This strategy was also applied to other blending systems. A device on the basis of the ITO/PEDOT:PSS/PDTBCDTBT139 (Scheme 3):PC71BM (2 : 3 in wt%)/Ca/Al conguration was fabricated. The active layer, thermally annealed at 100C for 10 min, achieved a high efficiency of 5.34%. Aer isothermal heating of the active layer at 150 C for 24 h, the efficiency dramatically degraded to an extremely low value of 0.01%. Under otherwise identical conditions, the device by adding 10 wt% PC61BPFinto the active layer delivered an initial efficiency of 5.24%, indicating that the incorporation of PC61BPFinto this system does not alter the device performance. In sharp contrast, aer thermal annealing at 150C for 25 h, this PC
61BPF -incor-porated device still achieved a comparable PCE of 4.24%, that is about 81% of its original value.
4.4 Light induced oligomerization of fullerenes
Photoinduced oligomerization of fullerenes via [2 + 2] cycload-dition has been demonstrated.140Lately, this concept has been incorporated into PSCs with the aim of ensuring thermally stable D–A morphology.130 The morphological behaviour of PCDTBT (Scheme 2):PC61BM blend was investigated. The as-cast PCDTBT:PC61BM blend lm was illuminated under
Fig. 5 Morphological stabilization of P3HT:PC61BM:PC61BPFblends. Reprinted with permission from ref. 129. Copyright 2013, Wiley-VCH Verlag GmbH & Co. KGaA.
Table 4 Photovoltaic parameters of PSCs (ITO/PEDOT:PSS/active layer/Ca/Al) before and after isothermal heating at 150C for 25 h
Entry Blend Weight ratio (in wt%) Time (h) Voc(V) Jsc(mA cm2) FF (%) PCE (%)
1 P3HT:PC61BM 1 : 1 0 0.60 10.26 66.00 4.08 2 25 0.62 2.27 49.00 0.69 3 P3HT:PC61BM:PC61BPF 6 : 5 : 1 0 0.60 9.87 65.60 3.88 4 25 0.62 9.06 65.47 3.68 5 P3HT:PC61BM:PC61BP 6 : 5 : 1 0 0.60 8.92 64.40 3.45 6 25 0.62 1.78 41.48 0.44
nitrogen for various time periods by a moderate uorescent lamp with intensity of 10 mW cm2. No apparent morphological change subsequent to thermal stress at 140 C for 1 h was observed for the blend pre-treated by the above-mentioned light exposure. In sharp contrast, the blend without light exposure underwent fast growth of PC61BM crystallites under otherwise identical thermal stress conditions. The enhanced morpholog-ical stability is attributed to the formation of small-domain PC61BM oligomers that minimize the diffusion of PC61BM out of polymer matrix. The device stability were then examined on the basis of the ITO/PEDOT:PSS/PCDTBT:PC61BM (1 : 2) (with or without >120 min light exposure)/Ca/Al. The initial efficien-cies for the devices with or without illumination are about 6%, suggesting that the illumination step does not have negative effect on the device performance. Overall, the device stability under thermal stress at 80 C was enhanced nearly 10 fold through the pre-treatment of light exposure. The application of this approach to other blends, such as DPP-TT-T141/PC
61BM, was demonstrated as well, suggesting its general applicability.
5.
Conclusions
Research effort devoted to the eld of bulk heterojunction (BHJ) polymer solar cells (PSCs) has delivered promising results. Nevertheless, in order to achieve the ultimate goal of commer-cialization, further optimization is certainly required. Herein we reviewed the research progress on applications of fullerene-based materials in the BHJ PSCs. It was categorized into three areas: bis-adduct fullerenes, fullerene-based interlayer modi-cation, and morphological stabilization. For bis-adduct fuller-enes, the second functionalization on the core structure of the mono-substituted C60 can reduce thep-conjugation and elec-tron delocalization in the C60, resulting in the formed bis-adduct C60 derivatives with larger electrochemical reduction potentials and thus higher-lying LUMO levels. These bis-adduct fullerenes, i.e. bisPCBM, ICBA, IC70BA, IPCBM, bis-TOQC, and DMPCBA, were used in BHJ PSCs and found to be effective in increasing the Voc of the corresponding devices, in turn enhancing the PCE values. Moreover, the function of bis-adduct fullerenes is not merely limited in the enlargement of Voc. For example, due to the amorphous nature of DMPCBA, the morphology of the P3HT:DMPCBA blend can be preserved aer long-term thermal heating.
As for interlayer modication, incorporation of an additional n-type fullerene interlayer between inorganic and organic active layers either in regular or inverted solar cells can help in minimization of charge recombination losses at the interface and improvement of charge transfer from the active layer to the inorganic layer, thus resulting in higher PCEs. To this end, various functional groups, such as acid, alcohol, phosphoric ester, polyethylene glycol, amine, styrene, oxetane, tri-chlorosilane, and ZnO were employed to modify fullerenes, respectively, and the resultant fullerenes exhibited promising interfacial electronic properties in the improvement of device characteristics. Certain fullerene, such as cross-linked PCBSD (C-PCBSD) can increase the device stability since the hot spots generated in the inorganic layer (ZnO) may be passivated by the
cross-linked PCBSD interlayer and the leakage current is therefore suppressed, giving rise to longer device operational lifetime. Noteworthily, the incorporation of the anodic aluminum oxide template with C-PCBSD can not only increase the device efficiency but also the device stability. The efficiency enhancement could be ascribed to the fact that the C-PCBSD nanostructures penetrating into the BHJ layer offer substantial P3HT/C-PCBSD interfacial area for extra exciton dissociation, thereby generating a higher photocurrent and PCE. The morphological stability may result from the spatial connement effect, which might suppress the thermal-driven crystallization of ICBA and P3HT upon thermal annealing.
Lastly, with regard to morphological stability, in situ poly-merization was applied to lock the optimal morphology by covalent bonding. Functionalized fullerenes in which the functional groups are polyethylene glycol and peruorophenyl moieties were utilized to suppress the thermal-induced fullerene aggregation. Light-induced oligomerization of PC61BM was demonstrated to be effective in minimizing the aggregation of PC61BM under thermal stress. These strategies can enhance the device stability signicantly since the optimal active-layer morphologies are more robust and therefore less sensitive towards thermal heating.
The research of employing fullerene derivatives in BHJ PSCs has established many valuable landmarks. It is envisioned that further investigation and advancement in thiseld can increase both the device efficiency and stability and ultimately reach the goal of commercialization.
Acknowledgements
We thank the Ministry of Science and Technology and the Ministry of Education, and Center for Interdiscplinary Science (CIS) of the National Chiao Tung University, Taiwan, for nancial support. Y. J. C thanks the support from the Golden-Jade fellowship of the Kenda foundation, Taiwan.
Notes and references
1 K. M. Coakley and M. D. McGehee, Chem. Mater., 2004, 16, 4533.
2 S. G¨unes, H. Neugebauer and N. S. Saricici, Chem. Rev., 2007, 107, 1324.
3 J. J. M. Halls, K. Pichler, R. H. Friend, S. C. Moratti and A. B. Holmes, Appl. Phys. Lett., 1996, 68, 3120.
4 M. Theander, A. Yartsev, D. Zigmantas, V. Sundstr¨om, W. Mammo, M. R. Andersson and O. Ingan¨as, Phys. Rev. B: Condens. Matter, 2000, 61, 12957.
5 A. Haugeneder, M. Neges, C. Kallinger, W. Spirkl, U. Lemmer, J. Feldmann, U. Scherf, E. Harth, A. G¨ugel and K. M¨ullen, Phys. Rev. B: Condens. Matter, 1999, 59, 15346.
6 T. St¨ubinger and W. Br¨utting, J. Appl. Phys., 2001, 90, 3632. 7 D. E. Markov, E. Amsterdam, P. W. M. Blom, A. B. Sieval and
J. C. Hummelen, J. Phys. Chem. A, 2005, 109, 5266. 8 Y.-J. Cheng, S.-H. Yang and C.-S. Hsu, Chem. Rev., 2009, 109,
5868.
9 W. Ma, C. Yang, X. Gong, K. Lee and A. J. Heeger, Adv. Funct. Mater., 2005, 15, 1617.
10 A. C. Mayer, S. R. Scully, B. E. Hardin, M. W. Rowell and M. D. McGehee, Mater. Today, 2007, 10, 28.
11 G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery and Y. Yang, Nat. Mater., 2005, 4, 864.
12 M. Reyes-Reyes, K. Kim and D. L. Carroll, Appl. Phys. Lett., 2005, 87, 083506.
13 C. H. Woo, B. C. Thompson, B. J. Kim, M. F. Toney and J. M. J. Fr´echet, J. Am. Chem. Soc., 2008, 130, 16324. 14 C. J. Brabec, A. Cravino, D. Meissner, N. S. Saricici,
T. Fromherz, M. T. Rispens, L. Sanchez and
J. C. Hummelen, Adv. Funct. Mater., 2001, 11, 374.
15 M. C. Scharber, D. M¨uhlbacher, M. Koppe, P. Denk, C. Waldauf, A. J. Heeger and C. J. Brabec, Adv. Mater., 2006, 18, 789.
16 L. J. A. Koster, V. D. Mihailetchi and P. W. M. Blom, Appl. Phys. Lett., 2006, 88, 093511.
17 G. Li, C. W. Chu, V. Shrotriya, J. Huang and Y. Yang, Appl. Phys. Lett., 2006, 88, 253503.
18 H. H. Liao, L. M. Chen, Z. Xu, G. Li and Y. Yang, Appl. Phys. Lett., 2008, 92, 173303.
19 Z. He, C. Zhong, S. Su, M. Xu, H. Wu and Y. Cao, Nat. Photonics, 2012, 6, 591.
20 X. Guo, N. Zhou, S. J. Lou, J. Smith, D. B. Tice, J. W. Hennek, R. P. Ortiz, J. T. L. Navarrete, S. Li, J. Strzalka, L. X. Chen, R. P. H. Chang, A. Facchetti and T. J. Marks, Nat. Photonics, 2013, 7, 825.
21 G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese and C. A. Grimes, Appl. Phys. Lett., 2007, 91, 152111.
22 C. Waldauf, M. Morana, P. Denk, P. Schilinsky, K. Coakley, S. A. Choulis and C. J. Brabec, Appl. Phys. Lett., 2006, 89, 233517.
23 M. S. White, D. C. Olson, S. E. Shaheen, N. Kopidakis and D. S. Ginley, Appl. Phys. Lett., 2006, 89, 143517.
24 R. Steim, S. A. Choulis, P. Schilinsky and C. J. Brabec, Appl. Phys. Lett., 2008, 92, 093303.
25 T. Yang, W. Cai, D. Qin, E. Wang, L. Lan, X. Gong, J. Peng and Y. Cao, J. Phys. Chem. C, 2010, 114, 6849.
26 H.-L. Yip and A. K.-Y. Jen, Energy Environ. Sci., 2012, 5, 5994. 27 S. Lee, B. Koo, J. Shin, E. Lee, H. Park and H. Kim, Appl.
Phys. Lett., 2006, 88, 162109.
28 M. McDowell, I. G. Hill, J. E. McDermott, S. L. Bernasek and J. Schwartz, Appl. Phys. Lett., 2006, 88, 073505.
29 L. L. Chua, J. Zaumseil, J. F. Chang, E. C. W. Ou, P. K. H. Ho, H. Sirringhaus and R. H. Friend, Nature, 2005, 434, 194. 30 Q. Wei, T. Nishizawa, K. Tajima and K. Hashimoto, Adv.
Mater., 2008, 20, 2211.
31 Q. Mei, C. Li, X. Gong, H. Lu, E. Jin, C. Du, Z. Lu, L. Jiang, X. Meng, C. Wang and Z. Bo, ACS Appl. Mater. Interfaces, 2013, 5, 8076.
32 C. Duan, W. Cai, B. B. Y. Hsu, C. Zhong, K. Zhang, C. Liu, Z. Hu, F. Huang, G. C. Bazan, A. J. Heeger and Y. Cao, Energy Environ. Sci., 2013, 6, 3022.
33 C. Duan, K. Zhang, C. Zhong, F. Huang and Y. Cao, Chem. Soc. Rev., 2013, 42, 9071.
34 H. Hoppe and N. S. Saricici, J. Mater. Chem., 2006, 16, 45 and references therein.
35 E. Klimov, W. Li, X. Yang, G. G. Hoffmann and J. Loos, Macromolecules, 2006, 39, 4493.
36 Y.-C. Huang, S.-Y. Chuang, M.-C. Wu, H.-L. Chen, C.-W. Chen and W.-F. Su, J. Appl. Phys., 2009, 106, 034506. 37 H. Zhong, X. Yang, B. deWith and J. Loos, Macromolecules,
2006, 39, 218.
38 X. Yang, J. K. J. van Duren, R. A. J. Janssen, M. A. J. Michels and J. Loos, Macromolecules, 2004, 37, 2151.
39 A. Swinnen, I. Haeldermans, M. vande Ven, J. D'Haen, G. Vanhoyland, S. Aresu, M. D'Olieslaeger and J. Manca, Adv. Funct. Mater., 2006, 16, 760.
40 C. Muller, T. A. M. Ferenczi, M. Campoy-Quiles, J. M. Frost, D. D. C. Bradley, P. Smith, N. Stingelin-Stutzmann and J. Nelson, Adv. Mater., 2008, 20, 3510.
41 X. Yang, J. K. J. van Duren, M. T. Rispens, J. C. Hummelen, R. A. J. Janssen, M. A. J. Michels and J. Loos, Adv. Mater., 2004, 16, 802.
42 M. Jørgensen, K. Norrman and F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2008, 92, 686.
43 K. Sivula, C. K. Luscombe, B. C. Thompson and J. M. J. Fr´echet, J. Am. Chem. Soc., 2006, 128, 13988. 44 Y. Zhang, H.-L. Yip, O. Acton, S. K. Hau, F. Huang and
A. K. Y. Jen, Chem. Mater., 2009, 21, 2598.
45 C.-Z. Li, S.-C. Chien, H.-L. Yip, C.-C. Chueh, F.-C. Chen, Y. Matsuo, E. Nakamura and A. K. Y. Jen, Chem. Commun., 2011, 47, 10082.
46 F. B. Kooistra, J. Knol, F. Kastenberg, L. M. Popescu, W. J. H. Verhees, J. M. Kroon and J. C. Hummelen, Org. Lett., 2007, 9, 551.
47 C. Yang, J. Y. Kim, S. Cho, J. K. Lee, A. J. Heeger and F. Wudl, J. Am. Chem. Soc., 2008, 130, 6444.
48 M. Lenes, G.-J. A. H. Wetzelaer, F. B. Kooistra, S. C. Veenstra, J. C. Hummelen and P. W. M. Blom, Adv. Mater., 2008, 20, 2116.
49 Y. He, H.-Y. Chen, J. Hou and Y. Li, J. Am. Chem. Soc., 2010, 132, 1377.
50 Y.-J. Cheng, C.-H. Hsieh, Y. He, C.-S. Hsu and Y. Li, J. Am. Chem. Soc., 2010, 132, 17381.
51 G. Zhao, Y. He and Y. Li, Adv. Mater., 2010, 22, 4355. 52 Y. Li, Chem. - Asian J., 2013, 8, 2316.
53 Y. He, G. Zhao, B. Peng and Y. Li, Adv. Funct. Mater., 2010, 20, 3383.
54 Y. Sun, C. Cui, H. Wang and Y. Li, Adv. Energy Mater., 2011, 1, 1058.
55 X. Fan, C. Cui, C. Fang, J. Wang, S. Li, F. Cheng, H. Long and Y. Li, Adv. Funct. Mater., 2012, 22, 585.
56 X. Guo, C. Cui, M. Zhang, L. Huo, Y. Huang, J. Hou and Y. Li, Energy Environ. Sci., 2012, 5, 7943.
57 Y. He, B. Peng, G. Zhao, Y. Zou and Y. Li, J. Phys. Chem. C, 2011, 115, 4340.
58 S.-H. Liao, Y.-L. Li, T.-H. Jen, Y.-S. Cheng and S.-A. Chen, J. Am. Chem. Soc., 2012, 134, 14271.
59 C. Zhang, S. Chen, Z. Xiao, Q. Zuo and L. Ding, Org. Lett., 2012, 14, 1508.
