Recent developments in the optimization of the bulk heterojunction morphology of polymer: Fullerene solar cells

materials Review Recent Developments in the Optimization of the Bulk Heterojunction Morphology of Polymer: Fullerene Solar Cells Hugo Gaspar 1, Flávio Figueira 2, Luiz Pereira 3, Adélio Mendes 4, Júlio C. Viana 1and Gabriel Bernardo 4,* 1IPC/i3N—Institute for Polymers and Composites, University of Minho, Campus de Azurém, 4800-058 Guimarães, Portugal; [email protected] (H.G.); [email protected] (J.C.V.) 2QOPNA, Departament of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal; f[email protected] 3Department of Physics and i3N—Institute for Nanostructures, Nanomodelling and Nanofabrication, University of Aveiro, 3810-193 Aveiro, Portugal; [email protected] 4LEPABE, Department of Chemical Engineering, University of Porto, 4200-465 Porto, Portugal; [email protected] *Correspondence: gbernar[email protected] Received: 26 November 2018; Accepted: 13 December 2018; Published: 16 December 2018   Abstract: Organic photovoltaic (OPV) devices, made with semiconducting polymers, have recently attained a power conversion efficiency (PCE) over 14% in single junction cells and over 17% in tandem cells. These high performances, together with the suitability of the technology to inexpensive large-scale manufacture, over lightweight and flexible plastic substrates using roll-to-roll (R2R) processing, place the technology amongst the most promising for future harvesting of solar energy. Although OPVs using non-fullerene acceptors have recently outperformed their fullerene-based counterparts, the research in the development of new fullerenes and in the improvement of the bulk-heterojunction (BHJ) morphology and device efficiency of polymer:fullerene solar cells remains very active. In this review article, the most relevant research works performed over the last 3 years, that is, since the year 2016 onwards, in the field of fullerene-based polymer solar cells based on the copolymers PTB7, PTB7-Th (also known as PBDTTT-EFT) and PffBT4T-2OD, are presented and discussed. This review is primarily focused on studies that involve the improvement of the BHJ morphology, efficiency and stability of small active area devices (typically < 15 mm 2 ), through the use of different processing strategies such as the use of different fullerene acceptors, different processing solvents and additives and different thermal treatments. Keywords: organic photovoltaics; fullerenes 1. Introduction Organic photovoltaics (OPVs) represent a promising approach to generate renewable energy. Compared with currently available technologies, OPVs can be easily manufactured over much larger areas, on lightweight plastic substrates with high flexibility, using high-throughput roll-to-roll (R2R) compatible processing technologies [ 1 – 6 ]. These capabilities strongly suggest that OPVs will enable large reductions in module fabrication cost and a consequent reduction on the energy payback time. The low production costs associated with OPVs might be the key for opening solar energy to new markets such as, for example, in rural communities and developing countries with poor transmission infrastructures by allowing people to invest and generate their own electricity. On the other hand, gadget market, in a low power, low cost energy conversion and wearable structure incorporation, are perhaps the most suitable application. Materials 2018,11, 2560; doi:10.3390/ma11122560 www.mdpi.com/journal/materials Materials 2018,11, 2560 2 of 36 Although the progress in OPVs has been slower than for example in perovskite solar cells, the efficiency of single junction polymer solar cells has been increasing steadily in the last 10–15 years, evolving from 5% in 2005 to above 14% in 2017 [ 7 , 8 ] and tandem cells have reached an even higher efficiency, above 17% [ 9 ]. Criteria such as efficiency, lifetime and cost, need to be satisfied to successfully commercialize large scale organic photovoltaics but considering the steady progress that has been happening in the field, the future of the technology looks bright. Polymer solar cells were dominated, for over two decades, by donor: acceptor blends based on fullerene acceptors. However, due to rapid developments in non-fullerene (NF) small molecule acceptors, in recent years NF acceptors have outperformed the fullerene acceptors and a large number of reviews have recently addressed the application of this newer type of acceptors in OPVs [ 10 – 15 ]. Despite this, the research activity in the field of fullerene-based polymer solar cells remains very intense and important developments have occurred in the last few years. This review focus mainly on the efforts performed over the last three years to improve the bulk-heterojunction (BHJ) morphology and the efficiency of small active area (typically < 15 mm 2 ) polymer:fullerene solar cells based on three of the most recent and highest efficiency low bandgap polymers, namely PTB7, PTB7-Th (also known as PBDTTT-EFT) and PffBT4T-2OD. These studies include either changing the morphology through the use of different processing strategies (different solvents and additives; different thermal treatments; different donor: acceptor ratios; etc.) or changing the morphology using other fullerenes different from the standards PC 71 BM ([6,6]-phenyl-C 71 -butyric acid methyl ester) and PC 61 BM ([6,6]-phenyl-C 61 -butyric acid methyl ester). Strategies to improve the morphological stability of the BHJ are also reported. As a benchmark for high performance, when considering studies involving the standard fullerenes PC 71 BM and PC 61 BM, mostly research works reporting power conversion efficiencies (PCE) of 8% or higher are reviewed. This requirement is relaxed when considering studies involving other fullerenes. This review provides first a general background to the field (Chapter 1) while Chapter 2 goes deeper on the polymers and fullerenes considered along the review; Chapter 3 discusses the general use of additives in OPVs and Chapters 4–6 discuss the main works based on respectively polymers PTB7, PTB7-Th and PffBT4T-2OD; finally, Chapter 7 draws the main conclusions and future perspectives. 1.1. Device Architectures OPV devices can be typically manufactured with three different architectures, as sketched in Figure 1: standard (normal); inverted and tandem. The device architecture of a normal or standard device is shown in Figure 1a. In this configuration, ITO is usually coated with a hole transport layer/electron blocking layer (HTL/EBL) of poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS). The BHJ layer is coated on top of the HTL and finally a low work function metal electrode (usually, calcium and aluminium) is deposited on top for collecting the electrons generated in the BHJ (cathode). Suitable improvement in the cathode can also be done by using dielectric materials like LiF that promote a remarkable decrease in the electrode work function by tunnelling process. The exposure of the acidic PEDOT:PSS to ITO has been reported as a potential stability problem, especially at high temperatures and high relative humidities. Although the standard geometry is still the most widely used, its limited operational lifetime is a disadvantage. Compared to the standard device, the inverted OPV configuration shown in Figure 1b has some associated advantages. In this configuration, the ITO acts as an electron collector (cathode) and is usually coated with a transparent metal oxide layer like zinc oxide or titanium dioxide. The active BHJ layer is deposited on top of the metal oxide layer, followed by a HTL (usually PEDOT:PSS) and the device is finalized with the deposition of a high work-function metal anode. Compared to the normal structure, the inverted structure has the advantage of allowing the use, as top anode, of an ambient stable high work-function metal such as gold (Au), silver (Ag) and copper (Cu), which can be deposited under normal ambient atmosphere conditions using scalable deposition Materials 2018,11, 2560 3 of 36 technologies such as inkjet printing [ 16 – 18 ]. A recent and extensive review on this particular OPV device architecture can be found elsewhere [19]. The device architecture of a tandem device is represented in Figure 1c. The tandem architecture integrates two or more sub-cells, integrating photoactive layers with complementary absorption spectra, stacked in series. The main advantage of the tandem architecture, compared to the single-junction architecture, is that it can maximize the proportion of the solar irradiation that is absorbed and used to generate electricity. A recent study suggests that OPVs with PCE > 17% can be obtained using highly optimized tandem architectures [9]. Materials 2018, 11, x FOR PEER REVIEW 3 of 35 (Cu), which can be deposited under normal ambient atmosphere conditions using scalable deposition technologies such as inkjet printing [16–18]. A recent and extensive review on this particular OPV device architecture can be found elsewhere [19]. The device architecture of a tandem device is represented in Figure 1c. The tandem architecture integrates two or more sub-cells, integrating photoactive layers with complementary absorption spectra, stacked in series. The main advantage of the tandem architecture, compared to the singlejunction architecture, is that it can maximize the proportion of the solar irradiation that is absorbed and used to generate electricity. A recent study suggests that OPVs with PCE > 17% can be obtained using highly optimized tandem architectures [9]. Figure 1. Architectures used in the construction of OPV devices: (a) Standard (Normal); (b) Inverted and (c) Tandem. TCO and ITO stand respectively for “Transparent Conductive Oxide” and “Indium Tin Oxide.” 1.2. Macroscopic Device Physics Figure 2a shows the typical current density-potential (J-V) curve for an OPV device under dark and illumination conditions. The power conversion efficiency (PCE) of the device is the product of three figures of merit, namely short-circuit current (Jsc), open-circuit voltage (Voc) and fill factor (FF) divided by the incident power radiation Pi, that is: PCE = (Jsc × Voc × FF)/Pi. Improving the figures of merit of an OPV requires a critical analysis of its macroscopic behaviour. Figure 2b shows the equivalent circuit for a solar cell. In this, Jph is the photoinduced current density (all generated photocurrent from photon to electron energy transfer); Rs is the series resistance; Rsh is the shunt resistance; J is the current flow in the external load and V is the applied voltage. The series resistance Rs originates from the bulk resistance of the active layer, the bulk resistance of the electrodes and the contact resistances of all interfaces in the device. The value of Rs determines where the current mainly flows (considering the potential range in the 4th quadrant): to the diode if Rs is high or to the external load if Rs is low. Therefore, in the fabrication of solar cells, ideally Rs should be as low as possible (Rs → 0), as an effective way to improve the performance of the device. The shunt resistance Rsh may originate from manufacturing or natural intrinsic defects and includes current leakage from pinholes in the film, current leakage from the edge of the cell, current leakage by energy levels acting as traps or electron-hole recombination. It has the effect of dividing the current in the equivalent circuit and therefore ideally the value of Rsh should approach infinity so that there is no current loss in the device, that is, there is no current flow through Rsh. Figure 1. Architectures used in the construction of OPV devices: ( a ) Standard (Normal); ( b ) Inverted and ( c ) Tandem. TCO and ITO stand respectively for “Transparent Conductive Oxide” and “Indium Tin Oxide.” 1.2. Macroscopic Device Physics Figure 2a shows the typical current density-potential (J-V) curve for an OPV device under dark and illumination conditions. The power conversion efficiency (PCE) of the device is the product of three figures of merit, namely short-circuit current (J sc ), open-circuit voltage (V oc ) and fill factor (FF) divided by the incident power radiation Pi, that is: PCE = (Jsc ×Voc ×FF)/Pi. Improving the figures of merit of an OPV requires a critical analysis of its macroscopic behaviour. Figure 2b shows the equivalent circuit for a solar cell. In this, J ph is the photoinduced current density (all generated photocurrent from photon to electron energy transfer); Rsis the series resistance; Rsh is the shunt resistance; Jis the current flow in the external load and Vis the applied voltage. The series resistance R s originates from the bulk resistance of the active layer, the bulk resistance of the electrodes and the contact resistances of all interfaces in the device. The value of R s determines where the current mainly flows (considering the potential range in the 4th quadrant): to the diode if R s is high or to the external load if R s is low. Therefore, in the fabrication of solar cells, ideally R s should be as low as possible (Rs→0), as an effective way to improve the performance of the device. The shunt resistance R sh may originate from manufacturing or natural intrinsic defects and includes current leakage from pinholes in the film, current leakage from the edge of the cell, current leakage by energy levels acting as traps or electron-hole recombination. It has the effect of dividing the Materials 2018,11, 2560 4 of 36 current in the equivalent circuit and therefore ideally the value of R sh should approach infinity so that there is no current loss in the device, that is, there is no current flow through Rsh. Materials 2018, 11, x FOR PEER REVIEW 4 of 35 Figure 2. (a) Fundamental points in a solar cell I-V curve required for a full understanding of its figures of merit; (b) The usual OPV equivalent electrical circuit; (c) Impact of the variation of the series resistance (Rs) on FF; (d) Impact of the variation of the shunt resistance (Rsh) on the FF; (e) real I-V curve (as usually found) where both Rs and Rsh are not ideal. Figure 2. ( a ) Fundamental points in a solar cell I-V curve required for a full understanding of its figures of merit; ( b ) The usual OPV equivalent electrical circuit; ( c ) Impact of the variation of the series resistance (R s ) on FF; ( d ) Impact of the variation of the shunt resistance (R sh ) on the FF; ( e ) real I-V curve (as usually found) where both Rsand Rsh are not ideal. Materials 2018,11, 2560 5 of 36 Figure 2c,d represents the impact of the series resistance and shunt resistance on the fill factor, considering all the remaining parameters as constant. The FF represents how “rectangular” the J-V curve is and it characterizes how “difficult” or how “easy” the photogenerated carriers can be extracted from the solar cell. The ideal value for FF is 100% when the J-V curve is a perfect rectangle—however in practice this never happens and in OPVs the FF values are typically in the range 40–80%. From Figure 2c it can be seen that at negative and low positive voltages the J-V curve is a straight line with slope 1/Rsh; at intermediate positive voltages the J-V curve is an exponential line that is determined by the diode; and at higher positive voltages the J-V curve is another straight line with slope 1/R s . Therefore, to improve the FF value the shunt resistance should be as high as possible and the series resistance should be as low as possible. It must be noted that in a real device R s is higher than the ideal and R sh is lower than the ideal. Therefore, an I-V curve as shown in Figure 2e is the usual. 1.3. Relationship between Nanoscopic and Macroscopic Device Physics The typical active layer in a OPV device is composed of a blend of an electron donating polymer and an electron acceptor which can be either fullerene-based (PC 71 BM, PC 61 BM or others) or a non-fullerene small organic molecule. This donor: acceptor blend is known as bulk-hetero-junction (BHJ). Figure 3a represents the typical band diagram for a donor: acceptor bulk-hetero-junction in an OPV device. Incident photons with energy higher than the bandgap (E g ), that is, E > E g , are absorbed by the BHJ causing the generation of excitons, that is, electron-hole pairs (step 1 in Figure 3a). Typical binding energy values for excitons are on the order of 0.5 eV or larger [ 20 , 21 ]. Then, the excitons created must diffuse to a donor: acceptor interface where they dissociate efficiently into holes and electrons (steps 2 and 3 in Figure 3a). At this point, it is worth emphasizing that until a few years ago the LUMO-LUMO offset ( ∆ LUMO) between the donor and the acceptor was considered to be the driving force needed to split the tightly bound excitons and a ∆ LUMO ≥ 0.3 V was regarded as the minimum necessary to ensure efficient charge transfer [ 22 ]. However, recent studies have shown that other factors also play their role and highly efficient OPVs can have ∆ LUMO < 0.1 eV [ 23 ]. Finally, the separated electrons and holes must travel along the interpenetrating network towards the metal cathode and anode, respectively, where they are collected giving rise to a photocurrent and photo-voltage (step 4 in Figure 3a). Due to the low dielectric constant of conjugated polymers, the typical exciton diffusion length (L D ) in a BHJ, that is, the average distance an exciton can diffuse through the material before recombination happens, is as low as ca. 10 nm [ 24 – 28 ], a length-scale that necessarily defines the optimal size of phase-separation for maximizing device efficiency. Materials 2018, 11, x FOR PEER REVIEW 5 of 35 Figure 2c,d represents the impact of the series resistance and shunt resistance on the fill factor, considering all the remaining parameters as constant. The FF represents how “rectangular” the J-V curve is and it characterizes how “difficult” or how “easy” the photogenerated carriers can be extracted from the solar cell. The ideal value for FF is 100% when the J-V curve is a perfect rectangle— however in practice this never happens and in OPVs the FF values are typically in the range 40–80%. From Figure 2c it can be seen that at negative and low positive voltages the J-V curve is a straight line with slope 1/Rsh; at intermediate positive voltages the J-V curve is an exponential line that is determined by the diode; and at higher positive voltages the J-V curve is another straight line with slope 1/Rs. Therefore, to improve the FF value the shunt resistance should be as high as possible and the series resistance should be as low as possible. It must be noted that in a real device Rs is higher than the ideal and Rsh is lower than the ideal. Therefore, an I-V curve as shown in Figure 2e is the usual. 1.3. Relationship between Nanoscopic and Macroscopic Device Physics The typical active layer in a OPV device is composed of a blend of an electron donating polymer and an electron acceptor which can be either fullerene-based (PC71BM, PC61BM or others) or a nonfullerene small organic molecule. This donor: acceptor blend is known as bulk-hetero-junction (BHJ). Figure 3a represents the typical band diagram for a donor: acceptor bulk-hetero-junction in an OPV device. Incident photons with energy higher than the bandgap (Eg), that is, E > Eg, are absorbed by the BHJ causing the generation of excitons, that is, electron-hole pairs (step 1 in Figure 3a). Typical binding energy values for excitons are on the order of 0.5 eV or larger [20,21]. Then, the excitons created must diffuse to a donor: acceptor interface where they dissociate efficiently into holes and electrons (steps 2 and 3 in Figure 3a). At this point, it is worth emphasizing that until a few years ago the LUMO-LUMO offset (ΔLUMO) between the donor and the acceptor was considered to be the driving force needed to split the tightly bound excitons and a ΔLUMO ≥ 0.3 V was regarded as the minimum necessary to ensure efficient charge transfer [22]. However, recent studies have shown that other factors also play their role and highly efficient OPVs can have ΔLUMO < 0.1 eV [23]. Finally, the separated electrons and holes must travel along the interpenetrating network towards the metal cathode and anode, respectively, where they are collected giving rise to a photocurrent and photovoltage (step 4 in Figure 3a). Due to the low dielectric constant of conjugated polymers, the typical exciton diffusion length (LD) in a BHJ, that is, the average distance an exciton can diffuse through the material before recombination happens, is as low as ca. 10 nm [24–28], a length-scale that necessarily defines the optimal size of phase-separation for maximizing device efficiency. Figure 3. Cont. Materials 2018,11, 2560 6 of 36 Materials 2018, 11, x FOR PEER REVIEW 6 of 35 Figure 3. (a) Typical band diagram for a donor:acceptor BHJ in organic solar cells; (b) Photocurrent generation processes in two BHJs with very different morphologies: a bad morphology (A) and a good morphology (B). This is illustrated in Figure 3b where the processes occurring in two very different BHJ morphologies are represented. In the ideal morphology, most of the excitons created can find nearby (distance < LD) a donor:acceptor interface where they dissociate generating free charge carriers and therefore the corresponding device has a high PCE. By contrast, in the morphology with very large phase segregation most of the excitons when created are too far away (distance > LD) from the nearest donor:acceptor interface and they recombine. Therefore, the ideal morphology for a BHJ is typically considered to be a bi-continuous and interpenetrating comb-like network morphology composed of donor-rich and acceptor-rich phases, with the donor phase smallest dimension of ~20 nm. Electrical carrier mobility depends on those domains. As widely accepted, the better figures of merit result from a compromise between D:A interface total area and D:A domains bulk volume. It is worth mentioning, at this point, that different experimental techniques can be used to probe the nanomorphology of BHJs. These include: Atomic Force Microscopy [29], High Resolution Electron Microscopy [30], Near-Edge X-ray Absorption Fine Structure (NEXAFS) spectroscopy [31], Grazing Incidence Wide and Small Angle X-ray Scattering (GI-WAXS/GI-SAXS) [32], Resonant Soft X-Ray Scattering (RSoXS) [33], Spectroscopic Ellipsometry [34,35], Neutron Reflectivity [36,37] and Small Angle Neutron Scattering (SANS) [38–40]; the discussion of these techniques is, however, beyond the scope of this review. The open circuit voltage (Voc) [41–44] is primarily dependent on the energy difference between the LUMO level of the acceptor and the HOMO level of the donor: the higher the difference, the higher the Voc. Therefore, the main strategy usually followed to increase Voc consists in utilizing either fullerene acceptors with higher LUMO levels or donor polymers with lower HOMO levels. As long as the LUMO of the fullerene acceptor remains lower than the LUMO of the polymer donor by an amount sufficient to promote charge separation, raising the fullerene LUMO level or lowering the polymer HOMO level should increase the Voc and thus the PCE. The Voc is prone to various loss mechanisms that affect the device efficiency and these have recently received special attention [21,45– 47]. The short-circuit current (Jsc) [48–51] depends on several factors such as the amount of light absorbed, the efficiency of exciton and free charge generation, the charge-carrier mobility and the efficiency of charge extraction at the electrodes. The higher these variables are, the higher is the corresponding Jsc value. Figure 3. ( a ) Typical band diagram for a donor:acceptor BHJ in organic solar cells; ( b ) Photocurrent generation processes in two BHJs with very different morphologies: a bad morphology (A) and a good morphology (B). This is illustrated in Figure 3b where the processes occurring in two very different BHJ morphologies are represented. In the ideal morphology, most of the excitons created can find nearby (distance < L D ) a donor:acceptor interface where they dissociate generating free charge carriers and therefore the corresponding device has a high PCE. By contrast, in the morphology with very large phase segregation most of the excitons when created are too far away (distance > L D ) from the nearest donor:acceptor interface and they recombine. Therefore, the ideal morphology for a BHJ is typically considered to be a bi-continuous and interpenetrating comb-like network morphology composed of donor-rich and acceptor-rich phases, with the donor phase smallest dimension of ~20 nm. Electrical carrier mobility depends on those domains. As widely accepted, the better figures of merit result from a compromise between D:A interface total area and D:A domains bulk volume. It is worth mentioning, at this point, that different experimental techniques can be used to probe the nano-morphology of BHJs. These include: Atomic Force Microscopy [ 29 ], High Resolution Electron Microscopy [ 30 ], Near-Edge X-ray Absorption Fine Structure (NEXAFS) spectroscopy [ 31 ], Grazing Incidence Wide and Small Angle X-ray Scattering (GI-WAXS/GI-SAXS) [ 32 ], Resonant Soft X-Ray Scattering (RSoXS) [ 33 ], Spectroscopic Ellipsometry [ 34 , 35 ], Neutron Reflectivity [ 36 , 37 ] and Small Angle Neutron Scattering (SANS) [ 38 – 40 ]; the discussion of these techniques is, however, beyond the scope of this review. The open circuit voltage (V oc ) [ 41 – 44 ] is primarily dependent on the energy difference between the LUMO level of the acceptor and the HOMO level of the donor: the higher the difference, the higher the V oc . Therefore, the main strategy usually followed to increase V oc consists in utilizing either fullerene acceptors with higher LUMO levels or donor polymers with lower HOMO levels. As long as the LUMO of the fullerene acceptor remains lower than the LUMO of the polymer donor by an amount sufficient to promote charge separation, raising the fullerene LUMO level or lowering the polymer HOMO level should increase the V oc and thus the PCE. The V oc is prone to various loss mechanisms that affect the device efficiency and these have recently received special attention [21,45–47]. The short-circuit current (J sc ) [ 48 – 51 ] depends on several factors such as the amount of light absorbed, the efficiency of exciton and free charge generation, the charge-carrier mobility and the efficiency of charge extraction at the electrodes. The higher these variables are, the higher is the corresponding Jsc value. Materials 2018,11, 2560 7 of 36 The fill factor (FF) [ 52 – 55 ], as mentioned above, depends on the series resistance, shunt resistance and diode. However, these also depend on a large number of factors that interact with each other intricately and for this reason the FF is more complex and less understood than the V oc and J sc [ 47 , 56 – 58 ]. Previous studies have shown that variables such as the blend composition, the blend morphology, the regioregularity of the conjugated polymer and the thickness of the BHJ influence significantly the FF affecting R s and R sh [ 52 ]. When the crystallinity of the blend layer increases, either by thermal annealing or through the use of a regioregular polymer, R s decreases [ 52 , 55 ]. When the thickness of the blend layer increases, R s increases and R sh decreases [ 52 , 59 ]. The quality of the two interfaces between the BHJ and the electrodes was also shown to play an important role in determining FF [ 52 , 60 ]. In practice, FF depends on the I-V shape in the 4th quadrant. As the D:A junction degrades (by molecular conformation and orientation and/or energetically inappropriate levels of D:A materials) the I-V behaviour becomes more space charge limited (SCLC) dependent and the I-V curve displays a typical “s-shape.” Common strategies to reduce R s and increase R sh include using buffer layers to reduce recombination [ 55 ], increasing the crystallinity of the BHJ materials [ 55 , 61 ], changing the donor: acceptor mass ratios [62] and optimizing the size of phase domains in the BHJ [55,61]. Among the several physical processes involved in the electrical power generation, the radiative (photons) absorption and charge collection at electrodes (involving the charge separation at D:A interfaces and electrical carriers transport in OPV volume) are often the most interesting for device improvement. In a simple model, the J sc value should be close to the J Ph, but from the equivalent electrical circuit it is easy to understand that J sc decreases, in a primary way, by the effect of series resistance R s . But, the I-V shape depends primarily on the R sh . In fact, this resistance comprehends all physical effects responsible for the electrical carrier recombination from the exciton separation (charge state at interface domain) until the collection at electrodes (from the electrical charge transport in volume). The value of R sh should, therefore, be as high as possible (as experimentally observed > 10 6Ω ). On other hand, increasing the photogenerated carriers should increase the J sc value. We must therefore primarily search for donors (polymers) with reduced (as possible) bandgap (HOMO—LUMO difference) to increase the solar light absorption in the visible spectrum. Donor-Acceptor (D:A) interface is also a key factor in the OPV energy performance. As soon as the bonded excitons reach the D:A interface, an extremely competitive process between separation (charge transfer—CT—states) and recombination will take place, determining how much efficient will be the free (net) charge creation. Once again, this process is almost (besides energy levels) dominated by the molecular structure/conformation of the donor/acceptor materials in the active film. Regarding the bulk transport, a competitive framework between electrical carriers drift and recombination must occur. To allow the highest charge collection possible at electrodes, the drift time should be smaller as possible to a positive competition with the high probable electron-hole recombination due to the complex distribution of a high density of electrical active energy levels in the BHJ [ 63 , 64 ]. As demonstrated from transient photoconductivity (see for example, [ 65 ]), the characteristic drift time is proportional to d3 2µE , where dis the organic layer thickness, µ the electrical carrier density and Ei=(VOC −Vext)×d−1 the internal electrical field, dependent on the V oc and the external applied voltage (V ext ). An important aspect, is that the drift time is inversely proportional to the electrical carrier mobility. Increasing the degree of structural ordering of the D–A matrix, an increase of electrical charge mobility is expected (hopping process facilitated) with a consequent decrease of the drift time. This must allow the electrical carriers to drift towards electrodes before they can recombine, increasing the FF. Morphological ordering of the active layer is therefore desired. All the referred parameters are intrinsically related to the donor and acceptor compounds used and their interaction in the solid state. Optimization and incrementation of these parameters require a deep comprehension of the device operation and photocurrent generation (J PH ) as well as its limitations [ 66 ]. Materials 2018,11, 2560 8 of 36 2. Polymers and Fullerenes In polymer:fullerene solar cells the role of light absorption has traditionally been assigned to the polymer donor, for the simple reason that fullerenes do not absorb strongly in the visible and near-infrared region of the spectrum. For this reason, a large amount of research in the OPV field has focused on synthesizing new polymers with a small optical bandgap E g [ 67 – 69 ]. Besides the low optical bandgaps to broaden the absorption range into the infrared spectrum, polymers used as electron donors in polymer:fullerene solar cells should exhibit suitable LUMO energy levels for efficient electron transfer to the fullerene moieties and crystalline characteristics to ensure good charge mobility [ 68 ]. Three of the most recent and successful small band gap copolymers that, when compared with the “old reference” poly(3-hexylthiophene-2,5-diyl) (P3HT), E g ca. 2.1 eV [ 70 ], have a smaller energy gap (allowing more of the sun’s spectral emission to be harvested) and an higher ionization potential (leading to an increased V oc and thus greater PCE) are: (i) poly[[4,8-bis[(2-ethylhexyl)oxy] benzo[1,2-b:4,5-b’]dithiophene-2,6-diyl][3-fluoro-2-[(2-ethylhexyl)carbonyl]thieno [3–b]thiophenedi-yl]], commonly known as PTB7 (E g ca. 1.6 eV); (ii) poly[4,8-bis(5-(2-ethyl hexyl)thiophen-2-yl)benzo [1–b;4,5-b’]dithiophene-2,6-diyl-alt-(4-(2-ethylhexyl)-3-fluorothieno[3–b]thiophene-)-2-carboxylate-2– 6-diyl)], commonly known as PTB7-Th or PBDTTT-EFT (E g ca. 1.6 eV) and iii) ((poly[(5,6-difluoro-2, 1,3-benzothiadiazol-4,7-diyl)-alt-(3,3 000 di(2-octyldodecyl) 2,2 0 ;5 0 ,2”;5”,2 000 -quaterthiophen-5,5 000 -diyl), commonly known as PffBT4T-2OD (Eg = 1.65 eV). The E g values indicated are only approximate as they depend on the polymer molecular weight [71]. Figure 4depicts the chemical structure of these three very popular “post-P3HT” conjugated polymers used in OPVs. In this review only devices based on these three polymers (PTB7 [ 72 ], PTB7-Th and PffBT4T-2OD) will be discussed. Materials 2018, 11, x FOR PEER REVIEW 8 of 35 2. Polymers and Fullerenes In polymer:fullerene solar cells the role of light absorption has traditionally been assigned to the polymer donor, for the simple reason that fullerenes do not absorb strongly in the visible and nearinfrared region of the spectrum. For this reason, a large amount of research in the OPV field has focused on synthesizing new polymers with a small optical bandgap Eg [67–69]. Besides the low optical bandgaps to broaden the absorption range into the infrared spectrum, polymers used as electron donors in polymer:fullerene solar cells should exhibit suitable LUMO energy levels for efficient electron transfer to the fullerene moieties and crystalline characteristics to ensure good charge mobility [68]. Three of the most recent and successful small band gap copolymers that, when compared with the “old reference” poly(3-hexylthiophene-2,5-diyl) (P3HT), Eg ca. 2.1 eV [70], have a smaller energy gap (allowing more of the sun’s spectral emission to be harvested) and an higher ionization potential (leading to an increased Voc and thus greater PCE) are: (i) poly[[4,8-bis[(2ethylhexyl)oxy] benzo[1,2-b:4,5-b’]dithiophene-2,6-diyl][3-fluoro-2-[(2-ethylhexyl)carbonyl]thieno [3–b]thiophenedi-yl]], commonly known as PTB7 (Eg ca. 1.6 eV); (ii) poly[4,8-bis(5-(2-ethyl hexyl)thiophen-2-yl)benzo[1–b;4,5-b’]dithiophene-2,6-diyl-alt-(4-(2-ethylhexyl)-3-fluorothieno[3– b]thiophene-)-2-carboxylate-2–6-diyl)], commonly known as PTB7-Th or PBDTTT-EFT (Eg ca. 1.6 eV) and iii) ((poly[(5,6-difluoro-2,1,3-benzothiadiazol-4,7-diyl)-alt-(3,3′′′di(2-octyldodecyl) 2,2′;5′,2′′;5′′,2′′′ -quaterthiophen-5,5′′′-diyl), commonly known as PffBT4T-2OD (Eg = 1.65 eV). The Eg values indicated are only approximate as they depend on the polymer molecular weight [71]. Figure 4 depicts the chemical structure of these three very popular “post-P3HT” conjugated polymers used in OPVs. In this review only devices based on these three polymers (PTB7 [72], PTB7Th and PffBT4T-2OD) will be discussed. Figure 4. Chemical structures of the polymer donors used in OPVs in this review. Figure 4. Chemical structures of the polymer donors used in OPVs in this review. Materials 2018,11, 2560 9 of 36 Buckminsterfullerene C 60 was the first fullerene used in an OPV device, in the seminal work by Sariciftci et al. [ 73 ], where these authors report the photo-induced electron-transfer from a conducting polymer to C 60 . However, the very low solubility of C 60 in common organic solvents makes it very difficult to process and therefore, soon after its introduction in the OPV field, the strategy of functionalizing C60 with solubilizing moieties was adopted. For this reason the fullerenes PC 61 BM [74,75] and its analogue PC 71 BM [ 76 ] soon emerged as the two most widely used electron accepting materials in organic photovoltaics (OPVs). These two fullerenes, PC 61 BM and PC 71 BM, are now utilized as reference acceptors for all kinds of other fullerene acceptors, because of their good solubility, high electron mobility and high chemical stability. A key difference between PC 61 BM and PC 71 BM is the ellipsoidal shape of the latter, as compared to the more spherical C 60 molecule [ 77 ]. The lower symmetry and more extended conjugation of C 70 enables energetic transitions that are forbidden in C 60 , leading to a broader photo-absorption profile in the visible region of the solar spectrum [ 78 ]. This allows increased photon harvesting and a potentially higher photocurrent for devices using PC 71 BM rather than PC 61 BM, an important attribute that has brought the C 70 analogue to the forefront of OPV research (despite its higher cost). Although PC 61 BM and PC 71 BM are the most commonly used fullerenes in organic solar cells, several other modified fullerenes such as those depicted in Figure 5are found in literature with specific characteristics and have been assessed as acceptors in BHJs. This will be discussed below in combination with the corresponding polymers depicted in Figure 4. As mentioned above, the V oc of a polymer:fullerene solar cell increases when the energy difference between the LUMO level of the fullerene and the HOMO level of the polymer increases (Figure 3a). Therefore, using fullerenes with a higher LUMO level results in devices with higher V oc values. A common synthetic approach used to increase the LUMO level of fullerenes consists in adding more addends to the fullerenecageinordertoobtain fullerene multiadducts. Such strategyreducesthe numberofdoublebonds and the level of conjugation in the fullerene molecule and therefore increases the corresponding LUMO level. The fullerene bisadduct 1 0 ,1”,4 0 ,4”-tetrahydro-di[1,4]methanonaphthaleno[5,6]fullerene-C 60 , commonly known as ICBA (Figure 5), has been the most popular bisadduct used in solar cells. However, as will be seen throughout this review, most often, although fullerene bisadducts such as ICBA produce devices with higher V oc values, the corresponding PCE values are lower than those of reference devices with PC 71 BM and PC 61 BM. The reasons for this are that fullerene multiadducts usually have much lower electron mobilities than monoadducts and this impacts the device’s J sc and FF adversely. Furthermore, these multiadducts are typically mixtures of several isomers and the small differences in the LUMO levels of these isomers contribute to increase the probability of exciton recombination and thus further decrease Jsc and FF. Materials 2018,11, 2560 16 of 36 Two new fullerene acceptors, namely N3 and N6 in Figure 5, were synthetized by Nagarjuna et al. and tested in solar cells based on PTB7. The cyclopropane ring in N3 is attached to two aryl rings: one aryl ring consists in fluorene with two long decyl chains, which increase the solubility of the molecule in organic solvents and the other aryl ring is a methyl benzoate substituent, that has an electron withdrawing CO 2 CH 3 group (EWG), which increases the electron accepting nature of the fullerene [ 102 ]. The fullerene N6 differs from N3 in that the methyl benzoate substituent is replaced by a long substituent containing NO 2 and CN groups, which are EWG. The highest PCE (4.1%) of devices with an inverted architecture was achieved with the system PTB7:N3 due to a slightly higher J sc and V oc . The poor shunt resistance in PTB7:N6 solar cells indicates more recombination of charge carriers in these devices. Lower solubility of N6 in 1,2-dichlorobenzene (o-DCB) may cause reduced donor-acceptor interface for excitons to be broken up into charges and insufficient percolation paths for charges to get collected at the electrode in PTB7:N6 films, resulting in higher recombination than in PTB7:N3 devices. On the other hand, series resistances were found to be 16.9 and 13.7 Ω· cm 2 for PTB7:N3 and PTB7:N6, respectively. A higher value of series resistance in the PTB7:N3 device is possibly due to a higher thickness of the film (85 nm) compared to that of PTB7:N6 blend films (70 nm) [ 102 ]. In a later work, the same authors synthetized trifluoromethyl derived fulleropyrrolidine (DIF-ful-C 60 in Figure 5) [ 103 ] fullerenes in which the attached − F and − CF 3 groups enhance the electron accepting character of the resultant fullerene due to the high electron deficient property of the ligands. The inverted devices with PTB7:DIF-ful-C 60 displayed a slightly higher efficiency than the control device of PTB7:PC 61 BM (6.8 vs. 6.2%) due to a higher V oc (0.82 vs. 0.70 V) and a higher J sc (−15.97 vs. −15. 40 mA·cm−2). In a highly original work, the impact of the endohedral fullerene Lu 3 N@C 80 PCBEH (Figure 5) on the morphology and efficiency of PTB7-based devices was tested by Roehling et al. [ 104 ]. This endohedral fullerene has a lower electron affinity than standard fullerenes, which can raise the V oc of photovoltaic devices. However, the PCE values obtained with standard devices processed with a o-DCB solution without additives were very poor (PCE of 0.4%) and, surprisingly, deteriorate further with the addition of additives. A morphological study showed that the poor performance results from a poor miscibility between PTB7 and Lu 3 N@C 80 PCBEH and in films with DIO added, the fullerene was observed to strongly aggregate into micrometre sized crystals. Ternary polymer solar cells involving BHJs of PTB7 with PC 71 BM and other fullerene have also been investigated. ICBA (Figure 5) was used by Cheng et al. [ 105 ] as an electron-cascade acceptor material in ternary blend devices of PTB7:PC 71 BM:ICBA. Due to the higher LUMO level of ICBA relative to PC 71 BM, the addition of ICBA leads to an increase in the V oc . The best performing devices were those with a PC 71 BM:ICBA ratio of 85:15 and exhibited a PCE of 8.24% compared to 7.35% for standard binary blend devices without ICBA. Very recently Ma et al. [ 106 ] added a dihydronaphthyl-based C 60 bisadduct (NCBA in Figure 5) to a normal PTB7:PC 71 BM BHJ. NCBA has a LUMO between the LUMOs of PTB7 and PC 71 BM and, therefore, plays a bridging role allowing the acceptor energy level to be tuned by changing the ratio of NCBA:PC 71 BM in the blend acceptor material. The best devices were obtained with a 15% mass ratio of NCBA and exhibited a PCE of 9.85%, compared with a PCE of 8.57% for the reference device without NCBA. Materials 2018,11, 2560 17 of 36 Table 1. Summary of the most relevant device figures of merit and efficiency results obtained with PTB7:fullerene BHJs. Acceptor D:A Wt. Solvent Additive JSC (mA cm−2)VOC (V) FF (%) PCE (%) Best (average) Device Structure Device Area (mm2)Obs. Ref PC71BM 1:1.5 CB (100 vol %) ——- 10.20 0.76 50.52 3.92 Standard ITO/PEDOT:PSS/BHJ/Ca/Al 10.0 —- [85] CB (97 vol %) DIO (3 vol %) 14.50 0.74 68.97 7.40 PC71BM 1:1.5 CB (97 vol %) DIO (3 vol %) 15.46 0.76 68 ±1 7.94 Standard ITO/PEDOT:PSS/BHJ/Ca/Al 4.5 (a) [86] PC71BM 1:1.5 o-DCB (100 vol %) ——- 18.51 0.76 60 8.50 Standard ITO/PEDOT:PSS/BHJ/Ca/Al 4.5 (b) [71] PC71BM 1:1.5 CB (97 vol %) DIO (3 vol %) 15.4 0.759 70.6 8.24 Standard ITO/PEDOT:PSS/BHJ/PFN/Ca/Al 16 (c) [87] 17.2 0.740 72.0 9.15 Inverted ITO/PFN/BHJ/MoO3/Al PC71BM 1:1.5 CB (100 vol %) ——- 11.43 0.70 46 3.99 (3.68) Standard ITO/PEDOT:PSS/BHJ/Ca/Al 7.5 (d) [61] CB (97 vol %) SH-na (3 vol %) 15.67 0.79 70 8.75 (8.42) Standard ITO/PEDOT:PSS/BHJ/Ca/Al PC71BM 1:1.5 CB (95 vol %) DPE (2 vol %) + DIO (3 vol %) 18.1 0.72 71.0 9.55 (9.25) Inverted ITO/ZnO/BHJ/MoO3/Ag —– —- [91] PC71BM 1:1.5 CB 97 vol % DIO 3 vol % 17.49 0.764 66.1 8.84 Inverted ITO/PEIE/BHJ/MoO3/Ag 9 —- [92] PC71BM 1:1.5 CB (91 vol %) FA (6 vol %) + DIO (3 vol %) 24.11 0.72 52.11 9.04 Standard ITO/PEDOT:PSS/BHJ/Li/Al 4 (e) [93] PC71BM 1:1.5 CB (97 vol %) DIO (3 vol %) 15.2 0.73 67.8 7.53 (7.40) Standard ITO/PEDOT:PSS/BHJ/PFN/Al 16 —- [94] CB (95 vol %) CBA (5 vol %) 16.7 0.75 73.0 9.11 (8.99) Standard ITO/PEDOT:PSS/BHJ/PFN/Al PC71BM 1:1.5 o-DCB (97 vol %) DIO (3 vol %) 14.2 0.74 60.0 6.30 Standard ITO/PEDOT:PSS/BHJ/Ca/Al 10 —- [98] ICBM 15.4 0.79 55.0 6.67 Fullerene 4 1:1.5 CB (97 vol %) DIO (3 vol %) 15.48 0.749 63.3 7.34 Inverted ITO/PFN/BHJ/MoOx/Al 9—- [99] Fullerene 5 14.21 0.760 67.3 7.27 Fullerene 6 14.03 0.797 61.0 6.83 PC61BM 14.29 0.740 66.5 7.03 Materials 2018,11, 2560 18 of 36 Table 1. Cont. Acceptor D:A Wt. Solvent Additive JSC (mA cm−2)VOC (V) FF (%) PCE (%) Best (average) Device Structure Device Area (mm2)Obs. Ref IC60MA-2C —– CB (97 vol %) DIO (3 vol %) 14.2 0.77 55 6.0 Inverted ITO/ZnO/BHJ/MoO3/Al 7—- [ 100 ] IC60MA-3C 12.9 0.79 50 5.1 IC60MA-4C 13.7 0.77 61 6.5 PC61BM 14.6 0.76 62 6.8 Fullerene 1e 1:1.5 CB 95 vol % DIO 5 vol % 12.3 0.825 53.3 5.4 Standard ITO/PEDOT:PSS/BHJ/Ca/Al 7.2 —- [ 101 ] PC61BM 12.1 0.760 64.4 5.9 Fullerene N3 1:1.5 o-DCB 97 vol % DIO 3 vol % 9.73 0.812 52.1 4.12 Inverted ITO/ZnO/BHJ/MoO 3 /Ag —– —- [ 102 ] Fullerene N6 9.06 0.805 50.2 3.64 DIF-ful-C60 1:1.5 o-DCB 97 vol % DIO 3 vol % 15.97 0.82 51 6.8 (6.5) Inverted ITO/ZnO/BHJ/MoO 3 /Ag —– (f) [ 103 ] PC61BM 15.40 0.70 58 6.2 (5.9) PC71BM 1:1.5 CB (97 vol %) DIO (3 vol %) 14.99 0.701 68.8 7.35 (7.23) Standard ITO/PEDOT:PSS/BHJ/Ca/Al 4—- [ 105 ] PC71BM(85%) + ICBA (15%) 16.32 0.720 69.2 8.24 (8.13) PC71BM 1:1.5 o-DCB (97 vol %) DIO (3 vol %) 17.4 0.76 64.8 8.57 Inverted ITO/ZnO/BHJ/MoO 3 /Ag 9—- [ 106 ] PC71BM (85%) + NCBA (15%) 18.6 0.78 67.9 9.85 Observations: (a) Methanol treatment; (b) PTB7 high M W of 128 kg/mol; (c) PFN is used as surface modifier; (d) With solution dipping; (e) FA is Formic acid; (f) For PTB7:DIF-ful-C60 based OPVs, the D:A mass ratio was optimized and 1:1.5 showed the best performances. Materials 2018,11, 2560 19 of 36 5. Devices Based on PTB7-Th PTB7-Th (Figure 4) possesses a conjugated backbone structure identical to PTB7, only differing on the nature of the organic ligand connected to the benzodithiophene (BDT) unit: the ether group in PTB7 is replaced by a thiophene group in PTB7-Th. Due to this chemical modification, PTB7-Th can sustain higher temperatures than PTB7 without suffering decomposition [ 107 ]. In this chapter, Section 5.1 discusses the most relevant studies performed in the optimization of the BHJ morphology and efficiency of PTB7-Th:PC 71 BM (or PC 61 BM) devices. Later, Section 5.2 addresses devices incorporating BHJs of PTB7-Th with other fullerenes. A compilation table, Table 2, is presented at the end of this chapter summarizing the most relevant and recent device efficiency results obtained with OPVs based on PTB7-Th. 5.1. PTB7-Th Devices with PC61BM and PC71BM PTB7-Th synthesis and testing in polymer solar cells was first reported by Zhang et al. [ 108 ] in 2014. A PCE of 9.0% was achieved in standard devices containing PTB7-Th:PC 71 BM in the ratio 1:1.5 and spin-coated from o-DCB solution with 3 vol % DIO. Devices processed with 1 vol % DIO and 5 vol % DIO were also tested but produced worse PCE results. Huang et al. studied the effect of the solvent additive DIO on the BHJ morphology and efficiency of PTB7-Th:PC 71 BM devices with inverted architecture, using PEIE (Polyethylenimine ethoxylated) as interfacial layer on top of ITO [ 109 ]. The device prepared from a pure CB solution, that is, without additive, shows an efficiency of 6.4% with a J sc of 16.2 mA · cm –2 , a V oc of 0.80 V and FF of 49%. Incorporating 3 vol % of DIO, the efficiency improves to 9.5%, corresponding to J sc ,V oc and FF values of 18.1 mA · cm –2 , 0.79 V and 66%, respectively. AFM and TEM studies confirm the effectiveness of DIO in improving the dispersion of PC 71 BM: the large aggregates ≈ 100 nm in diameter observed in films prepared without DIO are suppressed in films prepared with DIO [109]. A binary solvent additive made of DIO:N-methyl-pyrrolidine (NMP) (1.5%:1.5%) was employed by Wan et al. [ 110 ] to improve the PCE of PTB7-Th:PC 71 BM devices from 8.2% (without additive) and 9.5% (with 3 vol % DIO) to 10.8%. Resonant Soft x-Ray Scattering analysis showed that, although all the BHJs exhibit similar domain sizes, the BHJ processed with DIO:NMP showed higher phase purity than the BHJs with single additives or without additives, which facilitates the charge transport and reduces charge recombination leading to the higher PCE observed. Fan et al. [ 111 ] introduced a solid (melting point, m.p. = 125 ◦ C) fluorescent inhibitor (1-bromo-4-nitrobenzene) into the PTB7:PC 71 BM BHJs to improve device efficiency. All the devices were spin-coated from a CB solution with 3 vol % DIO and the best devices were obtained with 15 wt % of 1-bromo-4-nitrobenzene and exhibited a PCE of 8.95% compared to 7.58% for a reference device without 1-bromo-4-nitrobenzene. The increase in PCE was attributed to an improvement in charge transport and dissociation. The effect of different polymer:fullerene blend mass ratios (1:0.5; 1:1; 1:1.5; 1:2 and 1:3) on the photovoltaic performance of PTB7-Th:PC 71 BM devices, with standard architecture, was investigated by Komilian et al. [ 62 ]. No additives were used and the surface of the BHJ was washed with ethanol before top electrode deposition. The best devices with a PCE of 9.38% were obtained with a D:A mass ratio of 1:2. Xiao et al. [ 7 ] reported highly efficient ternary cells based on PTB7-Th:CO i 8DFIC (Acceptor A1):PC 71 BM (Acceptor A2), where COi8DFIC is a non-fullerene small molecule acceptor. The LUMO for PTB7-Th ( − 3.12 eV), PC 71 BM ( − 3.67 eV) and CO i 8DFIC ( − 3.88 eV) show a stepwise alignment and therefore PC 71 BM facilitates electron transfer from PTB7-Th to CO i 8DFIC. The mass ratio between PTB7-Th and A1+A2 was fixed to 1:1.5, while the content of A2 in acceptors was varied from 0 to 100%. The best devices were obtained with (D:A1:A2) mass ratios of (1:1.05:0.45) and produced a remarkable PCE of 14.08%. These ternary cells combine the advantage of fullerene acceptors (high µe ) and non-fullerene acceptors (strong visible or NIR absorption). Charge carrier mobilities were measured using the SCLC method. Compared with the binary blend film PTB7-Th:CO i 8DFIC (µh= 6.98 × 10 −4 , Materials 2018,11, 2560 20 of 36 µe = 3.89 × 10 −5 and µh / µe = 18), the ternary blend film PTB7-Th: CO i 8DFIC:PC 71 BM (1:1.05:0.45) showed a similar µh and a much higher µe , which resulted in a decrease in the ratio µh / µe from 18 to 1.3 ( µh = 6.35 × 10 −4 , µe = 4.80 × 10 −4 and µh / µe = 1.3). As a result of the improved electron transport and more balanced charge transport in the active layer containing PC71BM, Jsc and FF increased. Jagadamma et al. [ 112 ] investigated the effect of ex-situ thermal annealing on the BHJ morphology of PTB7-Th:PC 71 BM devices and on the corresponding efficiency. Devices with an inverted architecture were used and the thermal annealing was performed over a range of temperatures ranging from room temperature to 150 ◦ C. To exclude any negative influence from the possible degradation of charge selective layers and metal contacts, the ex-situ thermal annealing was applied only to the BHJ, that is, before metal anode deposition and not to the complete device. The unannealed devices displayed the highest PCE (9.1%) and this decreased gradually with the increase in thermal annealing temperature until it reached its lowest value (6.9%) for devices that had been annealed at 150 ◦ C. A morphological analysis using AFM showed that the BHJ morphology coarsens with increasing temperature and large scale phase separation is observed in BHJ annealed at 150 ◦ C which partially supresses exciton dissociation and increases recombination losses with the consequent drops observed in J sc ,FF and in the PCE. The effect of thickness variation of the BHJ (from 62 nm to 307 nm) on the photovoltaic performance of PTB7-Th:PC 71 BM devices, with both standard and inverted architectures, was studied by Kobori et al. [ 113 ]. In both device architectures, it was observed that FF decreased continuously with increasing the BHJ thickness. By contrast, the J sc increases with the BHJ thickness but not monotonically. Furthermore, the J sc of inverted devices were systematically higher than those of normal devices with identical thickness. In terms of overall PCE, for lower BHJ thickness (<100 nm) inverted devices perform considerably better than normal devices. However, for higher thicknesses the difference between the PCE of standard and inverted devices tends to fade away. The optimized devices were obtained in standard geometry with a BHJ thickness of 116 nm and a corresponding PCE of 9.25% and in inverted geometry with a BHJ thickness of 76 nm and a corresponding PCE of 10.4%. These results were explained by the difference of the simulated optical intensity distribution in the devices. In a similar work, Zang et al. [ 59 ] also studied the effect of BHJ thickness (70, 90, 120, 180 and 270 nm) and polymer:fullerene (D:A) mass ratio (1:1.5 and 1:3) on the performance of PTB7-Th:PC 71 BM devices with an inverted architecture. For devices with a D:A mass ratio of 1:1.5, a BHJ thickness of 90 nm produced the highest efficiency with a PCE of 9.68%, a V oc of 0.80 V, a J sc of 16.55 mA · cm −2 and a FF of 71%. Although the devices with a thicker BHJ (270 nm) displayed a considerably higher J sc (19.70 mA · cm −2 ), in agreement with the previous work by Kobori et al. [ 113 ], the corresponding PCE was lower (7.79%) due to a drastic drop in FF (50%) which offset the improved light absorption resultant from using thicker films. This decrease in FF was attributed to the less efficient charge carrier transport and dissociation in thicker films, due to the relatively low electron mobility in the BHJ. To improve the µ e of the BHJ, the authors also tested devices in which they changed the D:A mass ratio to 1:3. In this case they observed that for thick devices with a BHJ thickness of 270 nm, the PCE was higher (8.15%) than the PCE of the corresponding devices with a D:A mass ratio of 1:1.5 (7.79%). Other researchers studied the effect of different processing treatments on bulk and interfacial disorder in inverted PTB7-Th:PC 71 BM devices [ 114 ]. Devices fabricated with and without DIO showed PCEs of 8.3% and 3.8%, respectively. The lifetime stability and degradation of the PTB7-Th based solar cells has also been a topic of interest in recent years. Pearson et al. [ 115 ] measured the stability of reference PTB7-Th:PC 71 BM devices, with both standard and inverted architectures, over 70 h in an atmospheric chamber under continuous dry nitrogen flow, with typical oxygen and moisture levels of <5 ppm and <30 ppm respectively. Although devices with normal architecture exhibited a larger initial efficiency (PCE of 7.3% on average, vs. PCE of 6% for inverted cells), a PCE drop of more than 60% was observed in both device structures after only 24 h of light soaking under nitrogen. These results strongly suggest that BHJ degradation is Materials 2018,11, 2560 21 of 36 driven by light-mediated processes rather than high levels of oxygen and moisture exposure. The large PCE reduction observed in both device configurations was mostly due to a large overall reduction in J sc . Two different strategies were proposed for improving the stability of both standard and inverted devices. For standard devices, DIO was replaced by o-DCB as a co-solvent. Although the PCE of these new devices was slightly lower (6.5%), the device stability was greatly improved, which highlighted the detrimental stability effects associated with the use of DIO and previously identified by others in different BHJ systems. For inverted devices, placing a UV filter in front of the devices was also found to reduce considerably the extent of PCE degradation. Liu et al. [ 116 ] demonstrated that in encapsulated devices, completely isolated from moisture and oxygen, the absorption of high energy UV photons by the PCBM molecules leads to a degradation of exciton diffusion and charge mobility in the fullerene phase. The µe measured by the SCLC method decreased by four orders of magnitude, from 2.43 × 10 −3 to 2.16 × 10 −7 cm 2 V −1 s −1 and the µh decreased by less than one order of magnitude, after 45 h of AM 1.5G non-UV-filtered 1-sun illumination. Transient Absorption (TA) measurements combined with UV wavelength cut-off degradation experiments suggest that the burn-in may be triggered by a spin flip at the donor/acceptor interface, leading to the formation of PC 71 BM triplet anions and the accumulation of electrostatic potential energy. The release of this excess electrostatic potential energy promotes a disordering of the weakly bonded nanomorphological order, mostly in the PC 71 BM domains near the donor/acceptor interface. To circumvent this problem, the authors proposed a strategy that consisted in replacing CB by DCB as master solvent, removing the DIO from the processing solution, replacing ZnO buffer layer (an oxidant source) by PFN and applying a thermal annealing treatment, prior to light soaking, to increase the crystallinity of the PC 71 BM domains in the BHJ (Figure 8). Using this strategy, these authors were able to produce organic cells that were simultaneously highly efficient and very stable. Materials 2018, 11, x FOR PEER REVIEW 21 of 35 degradation is driven by light-mediated processes rather than high levels of oxygen and moisture exposure. The large PCE reduction observed in both device configurations was mostly due to a large overall reduction in Jsc. Two different strategies were proposed for improving the stability of both standard and inverted devices. For standard devices, DIO was replaced by o-DCB as a co-solvent. Although the PCE of these new devices was slightly lower (6.5%), the device stability was greatly improved, which highlighted the detrimental stability effects associated with the use of DIO and previously identified by others in different BHJ systems. For inverted devices, placing a UV filter in front of the devices was also found to reduce considerably the extent of PCE degradation. Liu et al. [116] demonstrated that in encapsulated devices, completely isolated from moisture and oxygen, the absorption of high energy UV photons by the PCBM molecules leads to a degradation of exciton diffusion and charge mobility in the fullerene phase. The μe measured by the SCLC method decreased by four orders of magnitude, from 2.43  10−3 to 2.16  10−7 cm2V−1s−1 and the μh decreased by less than one order of magnitude, after 45 h of AM 1.5G non-UV-filtered 1-sun illumination. Transient Absorption (TA) measurements combined with UV wavelength cut-off degradation experiments suggest that the burn-in may be triggered by a spin flip at the donor/acceptor interface, leading to the formation of PC71BM triplet anions and the accumulation of electrostatic potential energy. The release of this excess electrostatic potential energy promotes a disordering of the weakly bonded nanomorphological order, mostly in the PC71BM domains near the donor/acceptor interface. To circumvent this problem, the authors proposed a strategy that consisted in replacing CB by DCB as master solvent, removing the DIO from the processing solution, replacing ZnO buffer layer (an oxidant source) by PFN and applying a thermal annealing treatment, prior to light soaking, to increase the crystallinity of the PC71BM domains in the BHJ (Figure 8). Using this strategy, these authors were able to produce organic cells that were simultaneously highly efficient and very stable. Figure 8. Schematic of two possible paths for the nanomorphology evolution in the PC71BM acceptor phase: in the top one light induces disorder and in the bottom one light induced disorder is prevented in a highly crystalline configuration of the PC71BM molecules. Reprinted with permission from ref. [116]. 5.2. PTB7-Th Devices with Other Fullerenes Figure 8. Schematic of two possible paths for the nanomorphology evolution in the PC 71 BM acceptor phase: in the top one light induces disorder and in the bottom one light induced disorder is prevented in a highly crystalline configuration of the PC 71 BM molecules. Reprinted with permission from ref. [ 116 ]. Materials 2018,11, 2560 22 of 36 Table 2. Summary of the most relevant device figures of merit and efficiency results obtained with PTB7-Th:fullerene BHJs. Acceptor D:A (w/w)Solvent Additive JSC (mA cm−2)VOC (V) FF (%) PCE (%) Best (average) Device Structure Device Area (mm2)Obs. Ref. PC71BM 1:1.5 o-DCB (99 vol %) DIO (1 vol %) 16.03 0.786 65.12 8.21 Standard ITO/PEDOT:PSS/BHJ/Ca/Al 4—- [ 108 ] o-DCB (97 vol %) DIO (3 vol %) 16.86 0.784 68.16 9.00 o-DCB (95 vol %) DIO (5 vol %) 14.59 0.779 64.73 7.35 PC71BM 1:1.5 CB (100 vol %) ——— 16.2 0.80 49 6.4 (6.1) Inverted ITO/PEIE/BHJ/MoO3/Ag 4.5 —- [ 109 ] CB (97 vol %) DIO (3 vol %) 18.1 0.79 66 9.5 (9.3) PC71BM 1:1.5 o-DCB (100 vol %) ——— 17.0 0.83 58.1 8.2 (8.1) Standard ITO/PEDOT:PSS/BHJ/C 60 -N/Al 4(a) [ 110 ] o-DCB (97 vol %) NMP (3 vol %) 18.0 0.82 62.1 9.2 (8.9) o-DCB (97 vol %) DIO (3 vol %) 17.9 0.82 64.5 9.5 (9.2) o-DCB (97 vol %) DIO 1.5 vol % + NMP 1.5 vol % 19.1 0.82 69.1 10.8 (10.4) PC71BM 1:1.5 CB (97 vol %) DIO (3 vol %) 15.97 0.79 60.12 7.58 (7.51) ITO/PEDOT:PSS/BHJ/BCP/Ag 12.56 (b) [ 111 ] 17.74 0.784 64.11 8.95 (8.89) PC71BM 1:0.5 o-DCB (100 vol %) ——— 6.69 0.78 31 1.60 Standard ITO/PEDOT:PSS/BHJ/Ca/Al 13 (c) [62] 1:1 12.59 0.80 50 5.04 1:1.5 19.01 0.80 53 8.08 1:2 18.15 0.79 65 9.38 1:3 10.41 0.77 43 3.49 Materials 2018,11, 2560 23 of 36 Table 2. Cont. Acceptor D:A (w/w)Solvent Additive JSC (mA cm−2)VOC (V) FF (%) PCE (%) Best (average) Device Structure Device Area (mm2)Obs. Ref. PC71BM (0.45 wt) + COi8DFIC (1.05 wt) 1:1.5 CB (99 vol %) DIO (1 vol %) 28.20 0.70 71.0 14.08 Inverted ITO/ZnO/BHJ/MoO3/Ag 4 —- [7] PC71BM 1:1.5 CB (97 vol %) DIO (3 vol %) 16.6 0.77 71.3 9.1 (8.8) Inverted ITO/ZnO/BHJ/MoO3/Ag 8 (d) [ 112 ] PC71BM 1:1.5 o-DCB (97 vol %) DIO (3 vol %) 16.55 0.80 71 9.68 (9.42) Inverted ITO/ZnO/BHJ/MoO3/Ag —- (e) [59] 1:3 14.53 0.80 73 8.73 (8.51) PC71BM 1:1.8 CB (97 vol %) DIO (3 vol %) —- —- —- 9.25 Standard ITO/PEDOT:PSS/BHJ/LiF/Al 8(f) [ 113 ] —- —- —- 10.4 Inverted ITO/ZnO/BHJ/MoO3/Ag PC71BM 1:1.5 CB (100 vol %) ——— 9.7 0.83 47.5 3.8 (3.3) Inverted ITO/PEIE/BHJ/MoO3/Ag —- —- [ 114 ] CB (97 vol %) DIO (3 vol %) 16.1 0.79 64.8 8.3 (8.1) PC71BM 1:1.5 CB (97 vol %) DIO (3 vol %) Standard ITO/PEDOT:PSS/BHJ/Ca/Ag 4.5 —- [ 115 ] Inverted ITO/TiO2/BHJ/MoOx/Ag PC71BM 1:1.5 CB (97 vol %) DIO (3 vol %) 16.85 0.801 70.27 9.72 (9.49) Inverted ITO/ZnO/BHJ/MoO3/Ag 6—- [ 116 ] 1:1.5 1,2-DCB (100 vol %) ———– 14.93 0.802 65.28 7.82 Inverted ITO/PFN/BHJ/MoO3/Ag 1:2.0 16.01 0.802 72.28 9.59 (9.28) PC61BM 1:1.5 CB (97 vol %) DIO (3 vol %) 14.6 0.81 64 (7.6) Inverted ITO/PEIE/BHJ/MoO3/Ag 10 —- [ 117 ] PC71BM 17.7 0.80 66 (9.4) ICBA 13.3 1.0 53 (7.1) Materials 2018,11, 2560 24 of 36 Table 2. Cont. Acceptor D:A (w/w)Solvent Additive JSC (mA cm−2)VOC (V) FF (%) PCE (%) Best (average) Device Structure Device Area (mm2)Obs. Ref. PC61BM 1:1.5 CB (97 vol %) DIO (3 vol %) 14.0 0.80 65 7.3 Inverted ITO/ZnO/BHJ/MoOx/Ag 10.4 (g) [ 118 ] PyF5 1:2 13.7 0.84 56 6.5 FAP1 1:2 12.7 0.87 55 6.1 PC61BM 1:1.5 o-DCB (97 vol %) DIO (3 vol %) 15.14 0.84 62 7.9 (7.7) Inverted ITO/ZnO/BHJ/MoO3/Ag —- —- [ 103 ] DIF-ful-C 60 16.01 0.82 65 8.6 (8.4) PC71BM 1:1.5 CB (97 vol %) CLN (3 vol %) 14.0 0.795 48.8 5.4 Inverted ITO/ZnO/BHJ/MoO3/Ag 10 —- [ 119 ] CN-PC 71 BM 13.5 0.898 68.0 8.2 Observations: (a) NMP is N-methyl pyrrolidine. (b) Reference device (lower PCE) without 1-bromo-4-nitrobenzene; device with higher PCE contains 1-bromo-4-nitrobenzene (15 wt% active layer). (c) The BHJ was surface washed with ethanol before top electrode deposition. (d) Devices subjected to different annealing temperatures were tested (RT, 70, 100, 120 and 150 ◦ C). Best devices were obtained at RT. Only the performance of these is shown in this table. (e) For devices with D:A mass ratio 1:1.5, the following different BHJ thickness were tested: 70, 90, 120, 180 and 270 nm. Best devices were obtained with an optimized BHJ thickness of 90 nm. Only the performance of these is shown in this table. For devices with D:A ratio 1:3, the following different BHJ thickness were tested: 120, 180 and 270 nm. Best devices were obtained with an optimized BHJ thickness of 120 nm. Only the performance of these is shown in this table. (f) The performance data shown here were obtained with an optimized BHJ thickness of 116 nm in the standard devices and of 76 nm in the inverted devices. (g) The PTB7-Th:fullerene ratios were optimized and only devices with the highest efficiency ratios are shown: 1:1.5 for PC61BM and 1:2 for PyF5 and FAP1. Materials 2018,11, 2560 25 of 36 5.2. PTB7-Th Devices with Other Fullerenes Huang et al. [ 117 ] studied the BHJ morphology, photophysics and performance of inverted devices based on PTB7-Th blended with the indene-C 60 bisadduct ICBA (Figure 5) and correlated it with similar devices using the two classical fullerenes PC 61 BM and PC 71 BM. Although devices based on ICBA achieved, as expected, a higher V oc (1.0 V) than devices based on PC 61 BM and PC 71 BM (0.81 V and 0.80 V respectively) consistent with the higher LUMO of ICBA, the corresponding efficiency was lower (average PCE of 7.1% for devices with ICBA compared with 7.6% and 9.4% for devices based on PC 61 BM and PC 71 BM, respectively). Comparing the J sc values, PTB7-Th devices based on PC 71 BM have a much higher J sc (17.7 mA · cm −2 ) than similar devices based on PC 61 BM (14.6 mA · cm −2 ) and ICBA (13.3 mA · cm −2 ). This difference is largely explained by the fact that, in the visible region of the spectrum, PC 71 BM has a much higher absorption coefficient than both PC 61 BM and ICBA. ICBA BHJs are also characterized by lower polymer crystallinity, smaller domain sizes and better-mixed phases, which promotes a faster geminate recombination as revealed by transient absorption (TA) measurements. Overall this results in a poorer device performance of the ICBA-based devices. The carrier transport loss mechanism in PTB7-Th BHJs based on the amorphous fullerene acceptor PyF5, as well as on the semi-crystalline fullerene acceptors FAP1 and PC 61 BM (see chemical structures of PyF5 and FAP1 in Figure 5) were studied by Zhang et al. [ 118 ]. The PCE of optimized PyF5 and FAP1 based devices (6.5% and 6.1% respectively) is slightly lower than that of optimized PC 61 BM based devices (7.3%), owing to the slightly lower fill factor and J sc . Interestingly, optimized devices based on the fullerenes PyF5 and FAP1 require a higher fullerene mass loading (D:A = 1:2) than the optimized device based on PC 61 BM (D:A = 1:1.5). The charge carrier transport properties of pristine fullerenes and PTB7-Th:fullerene blends were determined with SCLC measurements. Although the three pristine fullerenes have very similar µe values (PyF5 = 4.0 × 10 −3 cm 2 V −1 s −1 ; FAP1 = 4.4 × 10 −3 cm 2 V −1 s −1 ; PC 61 BM = 4.5 × 10 −3 cm 2 V −1 s −1 ), the µe values of the corresponding BHJ composites are very different due to the different nanoscale morphologies. For example, the µe of PTB7-Th:PyF5 (1:1) and PTB7-Th:FAP1 (1:1) are in the order of 10 −7 –10 −6 cm 2 V −1 s −1 and these are in sharp contrast with the much higher µe of PTB7-Th:PC 61 BM which is ~10 −4 cm 2 V −1 s −1 . As the miscibility of PTB7-Th is higher with PyF5 and FAP1 than with PC 61 BM, PTB7-Th:PyF5 and PTB7-Th:FAP1 blends require significantly higher fullerene loadings to reach comparably high electron mobilities as for PTB7-Th:PC 61 BM blends. The authors, therefore, concluded that BHJ composites with good polymer–fullerene miscibility require higher fullerene loadings than composites with a tendency to phase separate. Nagarjuna et al. [ 103 ] synthesized the fullerene DIF-ful-C 60 (Figure 5) and tested it with PTB7, as previously mentioned and with PTB7-Th. The optimized PTB7-Th: DIF-ful-C 60 OPV devices showed a highest PCE of 8.6% with a J sc of 16.01 mA · cm −2 ,V oc of 0.82 V and a high FF of 65.5%, whereas the reference solar cell made from PTB7-Th:PC 61 BM blends showed a highest efficiency of 7.9%, with reduced J sc of 15.14 mA · cm −2 and FF of 62.1%. In a different work, the same authors [ 119 ] synthesized the fullerene derivative CN-PC 71 BM and optimized its application in PTB7-Th-based devices with an inverted structure. The best devices displayed a PCE of 8.2% and were obtained using a D:A mass ratio of 1:1.5, dissolved in CB and containing 3 vol % of CLN as additive. 6. Devices Based on PffBT4T-2OD The small band gap donor polymer poly[(5,6-difluoro-2,1,3-benzothiadiazol-4,7-diyl)-alt-(3,3 000 - di(2-octyldodecyl)2,2 0 ;5 0 ,2”;5”,2 000 -quaterthiophen-5,5 000 -diyl)] (PffBT4T-2OD) has recently attracted attention due to its potential to fabricate high performing OPV devices. PffBT4T-2OD, also known as PCE11, exhibits relatively high SCLC hole mobility of 1.5–3.0 × 10 −2 cm 2 V −1 s −1 [ 120 ] due to its high crystallinity. These properties, together with its tendency to form relatively pure polymer domains when blended with fullerenes, allow it to perform well in an OPV device, when used in relatively thick BHJ layers (~300 nm) with higher light absorption capabilities. PffBT4T-2OD when in solution also exhibits a peculiarly strong temperature dependent aggregation behaviour, forming a gel at room Materials 2018,11, 2560 32 of 36 37. Yan, Y.; Liu, X.; Wang, T. Conjugated-Polymer Blends for Organic Photovoltaics: Rational Control of Vertical Stratification for High Performance. Adv. Mater. 2017,29, 22. [CrossRef] 38. Chen, D.A.; Nakahara, A.; Wei, D.G.; Nordlund, D.; Russell, T.P. P3HT/PCBM Bulk Heterojunction Organic Photovoltaics: Correlating Efficiency and Morphology. Nano Lett. 2011,11, 561–567. [CrossRef] 39. Zhang, Y.; Parnell, A.J.; Pontecchiani, F.; Cooper, J.F.K.; Thompson, R.L.; Jones, R.A.L.; King, S.M.; Lidzey, D.G.; Bernardo, G. Understanding and controlling morphology evolution via DIO plasticization in PffBT4T-2OD/PC71BM devices. Sci. Rep. 2017,7, 44269. [CrossRef] 40. Hollamby, M.J. Practical applications of small-angle neutron scattering. Phys. Chem. Chem. Phys. 2013 , 15, 10566–10579. [CrossRef] 41. Elumalai, N.K.; Uddin, A. Open circuit voltage of organic solar cells: An in-depth review. Energy Environ. Sci. 2016,9, 391–410. [CrossRef] 42. Brabec, C.J.; Cravino, A.; Meissner, D.; Sariciftci, N.S.; Fromherz, T.; Rispens, M.T.; Sanchez, L.; Hummelen, J.C. Origin of the open circuit voltage of plastic solar cells. Adv. Funct. Mater. 2001 ,11, 374–380. [CrossRef] 43. Scharber, M.C.; Wuhlbacher, D.; Koppe, M.; Denk, P.; Waldauf, C.; Heeger, A.J.; Brabec, C.L. Design rules for donors in bulk-heterojunction solar cells—Towards 10% energy-conversion efficiency. Adv. Mater. 2006, 18, 789–794. [CrossRef] 44. Vandewal, K.; Tvingstedt, K.; Gadisa, A.; Inganas, O.; Manca, J.V. On the origin of the open-circuit voltage of polymer-fullerene solar cells. Nat. Mater. 2009,8, 904–909. [CrossRef] 45. Qian, D.; Zheng, Z.; Yao, H.; Tress, W.; Hopper, T.R.; Chen, S.; Li, S.; Liu, J.; Chen, S.; Zhang, J.; et al. Design rules for minimizing voltage losses in high-efficiency organic solar cells. Nat. Mater. 2018 ,17, 703–709. [CrossRef] 46. Wang, Y.M.; Qian, D.P.; Cui, Y.; Zhang, H.T.; Hou, J.H.; Vandewal, K.; Kirchartz, T.; Gao, F. Optical Gaps of Organic Solar Cells as a Reference for Comparing Voltage Losses. Adv. Energy Mater. 2018 ,8, 1801352. [CrossRef] 47. Ramirez, I.; Causa, M.; Zhong, Y.F.; Banerji, N.; Riede, M. Key Tradeoffs Limiting the Performance of Organic Photovoltaics. Adv. Energy Mater. 2018,8, 1703551. [CrossRef] 48. Lu, N.; Li, L.; Sun, P.; Liu, M. Short-circuit current model of organic solar cells. Chem. Phys. Lett. 2014 , 614, 27–30. [CrossRef] 49. Wang, J.C.; Shi, S.Q.; Leung, C.W.; Lau, S.P.; Wong, K.Y.; Chan, P.K.L. Short circuit current improvement in planar heterojunction organic solar cells by multijunction charge transfer. Appl. Phys. Lett. 2012 ,100, 053301. [CrossRef] 50. Cheng, F.; Fang, G.; Fan, X.; Liu, N.; Sun, N.; Qin, P.; Zheng, Q.; Wan, J.; Zhao, X. Enhancing the short-circuit current and efficiency of organic solar cells using MoO 3 and CuPc as buffer layers. Sol. Energy Mater. Sol. Cells 2011,95, 2914–2919. [CrossRef] 51. Banerjee, S.; Iyer, S.S.K. Short-circuit current density and spectral response modelling of bulk-heterojunction solar cells. Org. Electron. 2010,11, 2032–2036. [CrossRef] 52. Kim, M.S.; Kim, B.G.; Kim, J. Effective Variables To Control the Fill Factor of Organic Photovoltaic Cells. ACS Appl. Mater. Interfaces 2009,1, 1264–1269. [CrossRef] 53. Qi, B.Y.; Wang, J.Z. Fill factor in organic solar cells. Phys. Chem. Chem. Phys. 2013 ,15, 8972–8982. [CrossRef] 54. Gupta, D.; Mukhopadhyay, S.; Narayan, K.S. Fill factor in organic solar cells. Sol. Energy Mater. Sol. Cells 2010,94, 1309–1313. [CrossRef] 55. Jao, M.H.; Liao, H.C.; Su, W.F. Achieving a high fill factor for organic solar cells. J. Mater. Chem. A 2016 , 4, 5784–5801. [CrossRef] 56. Trukhanov, V.A.; Bruevich, V.V.; Paraschuk, D.Y. Fill factor in organic solar cells can exceed the Shockley-Queisser limit. Sci. Rep. 2015,5, 11478. [CrossRef] 57. Gebhardt, R.S.; Du, P.F.; Wodo, O.; Ganapathysubramanian, B. A data-driven identification of morphological features influencing the fill factor and efficiency of organic photovoltaic devices. Comput. Mater. Sci. 2017 , 129, 220–225. [CrossRef] 58. Tan, J.K.; Png, R.Q.; Zhao, C.; Ho, P.K.H. Ohmic transition at contacts key to maximizing fill factor and performance of organic solar cells. Nat. Commun. 2018,9, 3269. [CrossRef] 59. Zang, Y.; Xin, Q.; Zhao, J.; Lin, J. Effect of Active Layer Thickness on the Performance of Polymer Solar Cells Based on a Highly Efficient Donor Material of PTB7-Th. J. Phys. Chem. C 2018 ,122, 16532–16539. [CrossRef] Materials 2018,11, 2560 33 of 36 60. Huang, W.C.; Zhu, B.W.; Chang, S.Y.; Zhu, S.L.; Cheng, P.; Hsieh, Y.T.; Meng, L.; Wang, R.; Wang, C.C.; Zhu, C.H.; et al. High Mobility Indium Oxide Electron Transport Layer for an Efficient Charge Extraction and Optimized Nanomorphology in Organic Photovoltaics. Nano Lett. 2018,18, 5805–5811. [CrossRef] 61. Jhuo, H.J.; Liao, S.H.; Li, Y.L.; Yeh, P.N.; Chen, S.A.; Wu, W.R.; Su, C.J.; Lee, J.J.; Yamada, N.L.; Jeng, U.S. The Novel Additive 1-Naphthalenethiol Opens a New Processing Route to Efficiency-Enhanced Polymer Solar Cells. Adv. Funct. Mater. 2016,26, 3094–3104. [CrossRef] 62. Komilian, S.; Oklobia, O.; Sadat-Shafai, T. Controlling intercalations of PBDTTT-EFT side chain to initiate suitable network for charge extraction in PBDTTT-EFT:PC71BM blended bulk heterojunction solar cell. Sol. Energy Mater. Sol. Cells 2018,175, 35–40. [CrossRef] 63. Schmidt-Mende, L.; Weickert, J. Organic and Hybrid Solar Cells: An Introduction; De Gruyter: Berlin, Germany, 2016. 64. Wurfel, P. Physics of Solar Cells: From Basic Principles to Advanced Concepts; Wiley-VCH: Weinheim, Germany, 2009. 65. Cowan, S.R.; Street, R.A.; Cho, S.N.; Heeger, A.J. Transient photoconductivity in polymer bulk heterojunction solar cells: Competition between sweep-out and recombination. Phys. Rev. B 2011,83, 8. [CrossRef] 66. Cheng, P.; Zhan, X. Stability of organic solar cells: Challenges and strategies. Chem. Soc. Rev. 2016 , 45, 2544–2582. [CrossRef] 67. Xu, T.; Yu, L. How to design low bandgap polymers for highly efficient organic solar cells. Mater. Today 2014 , 17, 11–15. [CrossRef] 68. Liu, C.; Wang, K.; Gong, X.; Heeger, A.J. Low bandgap semiconducting polymers for polymeric photovoltaics. Chem. Soc. Rev. 2016,45, 4825–4846. [CrossRef] 69. Holliday, S.; Li, Y.L.; Luscombe, C.K. Recent advances in high performance donor-acceptor polymers for organic photovoltaics. Prog. Polym. Sci. 2017,70, 34–51. [CrossRef] 70. Bernardo, G.; Bucknall, D.G. Recent Progress in the Understanding and Manipulation of Morphology in Polymer: Fullerene Photovoltaic Cells; IntechOpen: Rijeka, Croatia, 2013. 71. Liu, C.; Wang, K.; Hu, X.W.; Yang, Y.L.; Hsu, C.H.; Zhang, W.; Xiao, S.; Gong, X.; Cao, Y. Molecular Weight Effect on the Efficiency of Polymer Solar Cells. ACS Appl. Mater. Interfaces 2013,5, 12163–12167. [CrossRef] 72. Lu, L.Y.; Yu, L.P. Understanding Low Bandgap Polymer PTB7 and Optimizing Polymer Solar Cells Based on It. Adv. Mater. 2014,26, 4413–4430. [CrossRef] 73. Sariciftci, N.S.; Smilowitz, L.; Heeger, A.J.; Wudl, F. Photoinduced Electron-Transfer From a Conducting Polymer to Buckminsterfullerene. Science 1992,258, 1474–1476. [CrossRef] 74. Hummelen, J.C.; Knight, B.W.; Lepeq, F.; Wudl, F.; Yao, J.; Wilkins, C.L. Preparation and Characterization of Fulleroid and Methanofullerene Derivatives. J. Org. Chem. 1995,60, 532–538. [CrossRef] 75. Yu, G.; Gao, J.; Hummelen, J.C.; Wudl, F.; Heeger, A.J. Polymer Photovoltaic Cells: Enhanced Efficiencies via a Network of Internal Donor-Acceptor Heterojunctions. Science 1995,270, 1789–1791. [CrossRef] 76. Wienk, M.M.; Kroon, J.M.; Verhees, W.J.H.; Knol, J.; Hummelen, J.C.; van Hal, P.A.; Janssen, R.A.J. Efficient methano 70 fullerene/MDMO-PPV bulk heterojunction photovoltaic cells. Angew. Chem. Int. Ed. 2003 , 42, 3371–3375. [CrossRef] [PubMed] 77. Williams, M.; Tummala, N.R.; Aziz, S.G.; Risko, C.; Bredas, J.L. Influence of Molecular Shape on Solid-State Packing in Disordered PC61BM and PC71BM Fullerenes. J. Phys. Chem. Lett. 2014 ,5, 3427–3433. [CrossRef] [PubMed] 78. Zhang, F.; Zhuo, Z.; Zhang, J.; Wang, X.; Xu, X.; Wang, Z.; Xin, Y.; Wang, J.; Wang, J.; Tang, W.; et al. Influence of PC60BM or PC70BM as electron acceptor on the performance of polymer solar cells. Sol. Energy Mater. Sol. Cells 2012,97, 71–77. [CrossRef] 79. McDowell, C.; Abdelsamie, M.; Toney, M.F.; Bazan, G.C. Solvent Additives: Key Morphology-Directing Agents for Solution-Processed Organic Solar Cells. Adv. Mater. 2018,30, 1707114. [CrossRef] [PubMed] 80. Liao, H.C.; Ho, C.C.; Chang, C.Y.; Jao, M.H.; Darling, S.B.; Su, W.F. Additives for morphology control in high-efficiency organic solar cells. Mater. Today 2013,16, 326–336. [CrossRef] 81. Lou, S.J.; Szarko, J.M.; Xu, T.; Yu, L.; Marks, T.J.; Chen, L.X. Effects of Additives on the Morphology of Solution Phase Aggregates Formed by Active Layer Components of High-Efficiency Organic Solar Cells. J. Am. Chem. Soc. 2011,133, 20661–20663. [CrossRef] Materials 2018,11, 2560 34 of 36 82. Burgués-Ceballos, I.; Machui, F.; Min, J.; Ameri, T.; Voigt, M.M.; Luponosov, Y.N.; Ponomarenko, S.A.; Lacharmoise, P.D.; Campoy-Quiles, M.; Brabec, C.J. Solubility Based Identification of Green Solvents for Small Molecule Organic Solar Cells. Adv. Funct. Mater. 2014,24, 1449–1457. [CrossRef] 83. Bernardo, G.; Washington, A.L.; Zhang, Y.; King, S.M.; Toolan, D.T.W.; Weir, M.P.; Dunbar, A.D.F.; Howse, J.R.; Dattani, R.; Fairclough, J.P.A.; et al. Does 1,8-diiodooctane affect the aggregation state of PC 71 BM in solution? R. Soc. Open Sci. 2018,5, 180937. [CrossRef] 84. Bernardo, G.; Washington, A.L.; Zhang, Y.; King, S.M.; Toolan, D.T.W.; Weir, M.P.; Dunbar, A.D.F.; Howse, J.R.; Dattani, R.; Fairclough, J.P.A.; et al. Data from: Does 1,8-Diiodooctane affect the aggregation state of PC71BM in solution? Dryad Dig. Repos. 2018. [CrossRef] 85. Liang, Y.Y.; Xu, Z.; Xia, J.B.; Tsai, S.T.; Wu, Y.; Li, G.; Ray, C.; Yu, L.P. For the Bright Future-Bulk Heterojunction Polymer Solar Cells with Power Conversion Efficiency of 7.4%. Adv. Mater. 2010 ,22, E135–E138. [CrossRef] [PubMed] 86. Zhou, H.Q.; Zhang, Y.; Seifter, J.; Collins, S.D.; Luo, C.; Bazan, G.C.; Nguyen, T.Q.; Heeger, A.J. High-Efficiency Polymer Solar Cells Enhanced by Solvent Treatment. Adv. Mater. 2013 ,25, 1646–1652. [CrossRef] [PubMed] 87. He, Z.; Zhong, C.; Su, S.; Xu, M.; Wu, H.; Cao, Y. Enhanced power-conversion efficiency in polymer solar cells using an inverted device structure. Nat. Photonics 2012,6, 591–595. [CrossRef] 88. Li, M.; Zhang, W.; Tang, X.; Jin, J.; Wang, H.; Chen, L.; Lv, W.; Chen, R.; Huang, W. Bromine-Terminated Additives for Phase-Separated Morphology Control of PTB7:PC71BM-Based Polymer Solar Cells. ACS Sustain. Chem. Eng. 2017,5, 11668–11675. [CrossRef] 89. Oseni, S.O.; Mola, G.T. The effect of uniand binary solvent additives in PTB7:PC61BM based solar cells. Sol. Energy 2017,150, 66–72. [CrossRef] 90. Zheng, Y.F.; Goh, T.; Fan, P.; Shi, W.; Yu, J.S.; Taylor, A.D. Toward Efficient Thick Active PTB7 Photovoltaic Layers Using Diphenyl Ether as a Solvent Additive. ACS Appl. Mater. Interfaces 2016 ,8, 15724–15731. [CrossRef] 91. Zheng, Y.; Wang, G.; Huang, D.; Kong, J.; Goh, T.; Huang, W.; Yu, J.; Taylor, A.D. Binary Solvent Additives Treatment Boosts the Efficiency of PTB7:PCBM Polymer Solar Cells to Over 9.5%. Sol. RRL 2018 ,2, 1700144. [CrossRef] 92. Zhao, X.; Xiang, J.; Liu, D.; Zhou, D.; Wang, G.; Zhou, G.; Alameh, K.; Ding, B.; Song, Q. Impact of alkyl chain length of 1,n-diiodoalkanes on PC71BM distribution in both bulk and air surface of PTB7:PC71BM film. Org. Electron. 2016,37, 358–365. [CrossRef] 93. Li, Y.; Xu, Z.; Zhao, S.; Huang, D.; Zhao, L.; Zhang, C.; Zhao, J.; Wang, P.; Zhu, Y. Enhanced carrier dynamics of PTB7:PC71BM based bulk heterojunction organic solar cells by the incorporation of formic acid. Org. Electron. 2016,28, 275–280. [CrossRef] 94. Chen, J.; Zhang, L.; Jiang, X.; Gao, K.; Liu, F.; Gong, X.; Chen, J.; Cao, Y. Using o-Chlorobenzaldehyde as a Fast Removable Solvent Additive during Spin-Coating PTB7-Based Active Layers: High Efficiency Thick-Film Polymer Solar Cells. Adv. Energy Mater. 2017,7, 1601344. [CrossRef] 95. Ciammaruchi, L.; Brunetti, F.; Visoly-Fisher, I. Solvent effects on the morphology and stability of PTB7:PCBM based solar cells. Sol. Energy 2016,137, 490–499. [CrossRef] 96. Dkhil, S.B.; Pfannmöller, M.; Saba, M.I.; Gaceur, M.; Heidari, H.; Videlot-Ackermann, C.; Margeat, O.; Guerrero, A.; Bisquert, J.; Garcia-Belmonte, G.; et al. Toward High-Temperature Stability of PTB7-Based Bulk Heterojunction Solar Cells: Impact of Fullerene Size and Solvent Additive. Adv. Energy Mater. 2017 , 7, 1601486. [CrossRef] 97. Bartesaghi, D.; Ye, G.; Chiechi, R.C.; Koster, L.J.A. Compatibility of PTB7 and [70]PCBM as a Key Factor for the Stability of PTB7:[70]PCBM Solar Cells. Adv. Energy Mater. 2016,6, 1502338. [CrossRef] 98. He, Y.J.; Shao, M.; Xiao, K.; Smith, S.C.; Hong, K.L. High-performance polymer photovoltaics based on rationally designed fullerene acceptors. Sol. Energy Mater. Sol. Cells 2013,118, 171–178. [CrossRef] 99. Karakawa, M.; Nagai, T.; Adachi, K.; Ie, Y.; Aso, Y. N-phenyl 60 fulleropyrrolidines: Alternative acceptor materials to PC61BM for high performance organic photovoltaic cells. J. Mater. Chem. A 2014 ,2, 20889–20895. [CrossRef] 100. Tseng, N.W.; Yu, Y.; Li, Y.K.; Zhao, J.B.; So, S.K.; Yan, H.; Ng, K.M. Isobenzofulvene-fullerene mono-adducts for organic photovoltaic applications. J. Mater. Chem. C 2015,3, 977–980. [CrossRef] Materials 2018,11, 2560 35 of 36 101. Huang, S.H.; Zhang, G.Y.; Knutson, N.S.; Fontana, M.T.; Huber, R.C.; Ferreira, A.S.; Tolbert, S.H.; Schwartz, B.J.; Rubin, Y. Beyond PCBM: Methoxylated 1,4-bisbenzyl 60 fullerene adducts for efficient organic solar cells. J. Mater. Chem. A 2016,4, 416–424. [CrossRef] 102. Nagarjuna, P.; Bagui, A.; Hou, J.H.; Singh, S.P. New Electron Acceptor Derived from Fluorene: Synthesis and Its Photovoltaic Properties. J. Phys. Chem. C 2016,120, 13390–13397. [CrossRef] 103. Nagarjuna, P.; Bagui, A.; Garg, A.; Gupta, V.; Singh, S.P. One-Step Synthesis of New Electron Acceptor for High Efficiency Solution Processable Organic Solar Cells. J. Phys. Chem. C 2017 ,121, 26615–26621. [CrossRef] 104. Roehling, J.D.; Baran, D.; Sit, J.; Kassar, T.; Ameri, T.; Unruh, T.; Brabec, C.J.; Moulé, A.J. Nanoscale Morphology of PTB7 Based Organic Photovoltaics as a Function of Fullerene Size. Sci. Rep. 2016 ,6, 30915. [CrossRef] [PubMed] 105. Cheng, P.; Li, Y.F.; Zhan, X.W. Efficient ternary blend polymer solar cells with indene-C-60 bisadduct as an electron-cascade acceptor. Energy Environ. Sci. 2014,7, 2005–2011. [CrossRef] 106. Ma, G.; Liu, Z.; Wang, N. Efficient ternary polymer solar cells with dihydronaphthyl-based C60 bisadduct as an third component material. Sol. Energy 2018,170, 164–173. [CrossRef] 107. Fernandes, L.; Gaspar, H.; Tome, J.P.C.; Figueira, F.; Bernardo, G. Thermal stability of low-bandgap copolymers PTB7 and PTB7-Th and their bulk heterojunction composites. Polym. Bull. 2018 ,75, 515–532. [CrossRef] 108. Zhang, S.Q.; Ye, L.; Zhao, W.C.; Liu, D.L.; Yao, H.F.; Hou, J.H. Side Chain Selection for Designing Highly Efficient Photovoltaic Polymers with 2D-Conjugated Structure. Macromolecules 2014 ,47, 4653–4659. [CrossRef] 109. Huang, W.C.; Gann, E.; Thomsen, L.; Dong, C.K.; Cheng, Y.B.; McNeill, C.R. Unraveling the Morphology of High Efficiency Polymer Solar Cells Based on the Donor Polymer PBDTTT-EFT. Adv. Energy Mater. 2015 , 5, 11. [CrossRef] 110. Wan, Q.; Guo, X.; Wang, Z.; Li, W.; Guo, B.; Ma, W.; Zhang, M.; Li, Y. 10.8% Efficiency Polymer Solar Cells Based on PTB7-Th and PC71BM via Binary Solvent Additives Treatment. Adv. Funct. Mater. 2016 , 26, 6635–6640. [CrossRef] 111. Fan, R.; Huai, Z.X.; Sun, Y.S.; Li, X.W.; Fu, G.S.; Huang, S.H.; Wang, L.X.; Yang, S.P. Enhanced performance of polymer solar cells based on PTB7-Th: PC71BM by doping with 1-bromo-4-nitrobenzene. J. Mater. Chem. C 2017,5, 10985–10990. [CrossRef] 112. Jagadamma, L.K.; Sajjad, M.T.; Savikhin, V.; Toney, M.F.; Samuel, I.D.W. Correlating photovoltaic properties of a PTB7-Th:PC71BM blend to photophysics and microstructure as a function of thermal annealing. J. Mater. Chem. A 2017,5, 14646–14657. [CrossRef] 113. Kobori, T.; Fukuda, T. Effect of optical intensity distribution on device performances of PTB7-Th:PC71BM-based organic photovoltaic cells. Org. Electron. 2017,51, 76–85. [CrossRef] 114. Jain, N.; Chandrasekaran, N.; Sadhanala, A.; Friend, R.H.; McNeill, C.R.; Kabra, D. Interfacial disorder in efficient polymer solar cells: The impact of donor molecular structure and solvent additives. J. Mater. Chem. A 2017,5, 24749–24757. [CrossRef] 115. Pearson, A.J.; Hopkinson, P.E.; Couderc, E.; Domanski, K.; Abdi-Jalebi, M.; Greenham, N.C. Critical light instability in CB/DIO processed PBDTTT-EFT:PC71BM organic photovoltaic devices. Org. Electron. 2016 , 30, 225–236. [CrossRef] 116. Liu, Q.; Toudert, J.; Liu, F.; Mantilla-Perez, P.; Bajo, M.M.; Russell, T.P.; Martorell, J. Circumventing UV Light Induced Nanomorphology Disorder to Achieve Long Lifetime PTB7-Th:PCBM Based Solar Cells. Adv. Energy Mater. 2017,7, 1701201. [CrossRef] 117. Huang, W.C.; Gann, E.; Chandrasekaran, N.; Prasad, S.K.K.; Chang, S.Y.; Thomsen, L.; Kabra, D.; Hodgkiss, J.M.; Cheng, Y.B.; Yang, Y.; et al. Influence of Fullerene Acceptor on the Performance, Microstructure, and Photophysics of Low Bandgap Polymer Solar Cells. Adv. Energy Mater. 2017 ,7, 10. [CrossRef] 118. Zhang, C.H.; Langner, S.; Mumyatov, A.V.; Anokhin, D.V.; Min, J.; Perea, J.D.; Gerasimov, K.L.; Osvet, A.; Ivanov, D.A.; Troshin, P.; et al. Understanding the correlation and balance between the miscibility and optoelectronic properties of polymer-fullerene solar cells. J. Mater. Chem. A 2017 ,5, 17570–17579. [CrossRef] 119. Nagarjuna, P.; Bagui, A.; Gupta, V.; Singh, S.P. A highly efficient PTB7-Th polymer donor bulk hetero-junction solar cell with increased open circuit voltage using fullerene acceptor CN-PC70BM. Org. Electron. 2017 , 43, 262–267. [CrossRef] Materials 2018,11, 2560 36 of 36 120. Liu, Y.; Zhao, J.; Li, Z.; Mu, C.; Ma, W.; Hu, H.; Jiang, K.; Lin, H.; Ade, H.; Yan, H. Aggregation and morphology control enables multiple cases of high-efficiency polymer solar cells. Nat. Commun. 2014 ,5, 5293. [CrossRef] 121. Ma, W.; Yang, G.; Jiang, K.; Carpenter Joshua, H.; Wu, Y.; Meng, X.; McAfee, T.; Zhao, J.; Zhu, C.; Wang, C.; et al. Influence of Processing Parameters and Molecular Weight on the Morphology and Properties of High-Performance PffBT4T-2OD:PC71BM Organic Solar Cells. Adv. Energy Mater. 2015 ,5, 1501400. [CrossRef] 122. Zhang, X.; Zheng, D.; Xing, S.; Wang, H.; Huang, J.; Yu, J. Precisely control the morphology and crystallization of temperature-dependent aggregation bulk heterojunction by using co-solvent system for optimized light intensity distribution and its effect on thick active layer polymer solar cells. Sol. Energy 2017 ,147, 106–112. [CrossRef] 123. Zhao, J.; Zhao, S.; Xu, Z.; Qiao, B.; Huang, D.; Zhao, L.; Li, Y.; Zhu, Y.; Wang, P. Revealing the Effect of Additives with Different Solubility on the Morphology and the Donor Crystalline Structures of Organic Solar Cells. ACS Appl. Mater. Interfaces 2016,8, 18231–18237. [CrossRef] 124. Umeyama, T.; Igarashi, K.; Sakamaki, D.; Seki, S.; Imahori, H. Unique cohesive nature of the [small beta]1-isomer of [70]PCBM fullerene on structures and photovoltaic performances of bulk heterojunction films with PffBT4T-2OD polymers. Chem. Commun. 2018,54, 405–408. [CrossRef] [PubMed] 125. Vidal, S.; Izquierdo, M.; Law, W.K.; Jiang, K.; Filippone, S.; Perles, J.; Yan, H.; Martin, N. Photochemical site-selective synthesis of [70]methanofullerenes. Chem. Commun. 2016 ,52, 12733–12736. [CrossRef] [PubMed] 126. Li, W.; Cai, J.; Cai, F.; Yan, Y.; Yi, H.; Gurney, R.S.; Liu, D.; Iraqi, A.; Wang, T. Achieving over 11% power conversion efficiency in PffBT4T-2OD-based ternary polymer solar cells with enhanced open-circuit-voltage and suppressed charge recombination. Nano Energy 2018,44, 155–163. [CrossRef] 127. Zhang, Y.; Parnell, A.J.; Blaszczyk, O.; Musser, A.J.; Samuel, I.D.W.; Lidzey, D.G.; Bernardo, G. Effect of fullerene acceptor on the performance of solar cells based on PffBT4T-2OD. Phys. Chem. Chem. Phys. 2018 , 20, 19023–19029. [CrossRef] [PubMed] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).