Abstract

Photovoltaic devices based on nanocomposites composed of conjugated polymers and inorganic nanocrystals show promise for the fabrication of low-cost third-generation thin film photovoltaics. In theory, hybrid solar cells can combine the advantages of the two classes of materials to potentially provide high power conversion efficiencies of up to 10%; however, certain limitations on the current within a hybrid solar cell must be overcome. Current limitations arise from incompatibilities among the various intradevice interfaces and the uncontrolled aggregation of nanocrystals during the step in which the nanocrystals are mixed into the polymer matrix. Both effects can lead to charge transfer and transport inefficiencies. This paper highlights potential strategies for resolving these obstacles and presents an outlook on the future directions of this field.

1. Introduction

Hybrid solar cells combine both organic and inorganic semiconductors in an active layer such that the organic or polymer semiconductor serves as the electron donor and transports photogenerated holes, whereas the inorganic semiconductor accepts and transports electrons [16]. Theoretically, the hybrid photovoltaic devices (HPVs) are expected to achieve a high power conversion efficiency (PCE) because they combine the advantageous characteristics of polymers and nanocrystals (NCs), including the flexibility, light weight, and low fabrication costs of polymer materials [79] and the high electron mobility, size-dependent optical properties [10, 11], and physical and chemical stability of inorganic NCs [12]. Unfortunately, the PCE values obtained thus far in hybrid devices have not exceeded 4% under simulated air mass (AM) 1.5 illumination [13]. The main barriers to a higher PCE are thought to be an inefficient exciton dissociation at the donor/acceptor (D/A) interface [1416], inhibition of recombination [17, 18], and poor charge transport to the electrodes [1921]. Therefore, the design of compatible surfaces, accounting for the different chemical properties of the organic and inorganic materials, and control over the phase separation of the composites are crucial for achieving rapid and high-yield charge separation at the D/A interface and for promoting charge transport and collection at the electrodes. Three distinct strategies have been explored toward improving the interface design in the nanocomposite materials to enable hybrid solar cells to achieve high PCE. The first approach, ligand exchange, uses a mix of polymers and inorganic NCs prepared via colloidal synthesis approaches. The second strategy, grafting, utilizes the grafting of a polymer from/onto the NCs, yielding polymer/NCs nanocomposites with improved grafting density. The third strategy, direct NC growth, involves the use of a molecular precursor to the inorganic semiconductor dissolved together with the polymer in a common solvent and this solution may then be used to deposit the photoactive layer.

Improving the photovoltaic efficiency requires a clear understanding of the structure-properties relationship; therefore, we focus here on one type of hybrid solar cell, hybrid bulk heterojunction solar cells. This paper provides the reader with insight into the basic principles underlying these devices and discusses the current state-of-the-art in the three synthetic strategies mentioned above. A goal in this field is to understand the crucial parameters that are responsible for HPVs performance. Motivated by the rapid growth and development of this field, this review describes the recent advances and progress toward device improvement. An outlook is provided on the future materials and technologies that are likely to guide the future directions of research.

2. Hybrid Solar Cells

2.1. Definitions

“Hybrid” refers to the association of at least two components of distinctly different chemical natures, the molecular-level distribution of which components are achieved either through simple mixing or through linking the components together via specific interactions, such as covalent, coordination, ionic, or hydrogen bonds. Each hybrid component possesses its chemical identity and can exist independently of the hybrid material [22].

A bulk heterojunction is by definition a homogeneous blend of a p-type and an N-type semiconductor (donor/acceptor). Organic photovoltaic devices (OPVs) based on blends of conjugated polymers and fullerenes form interpenetrating donor/acceptor networks and have been used in prototype bulk heterojunction geometry applications. Bulk heterojunctions in polymer-inorganic hybrid solar cells may be formed by replacing the fullerenes, which act as organic nanoparticles (NPs), with inorganic semiconductors as electron acceptors for dispersal in the polymer matrix. NCs based on metal oxides (ZnO [2325], TiO2 [2628]), group II–VI (ZnS [29, 30], ZnSe [31], CdTe [3234], CdS [3537]), group III–V (GaAs [38, 39], InP [40]), group IV–VI (PbSe [41], PbS [4244]), group IV (Si [45, 46]), CuInS2 [47, 48], CuInSe2 [49], have been tested for their utility as electron acceptors.

2.2. Device Structure and Working Principle

NCs/polymer bulk heterojunction hybrid solar cells usually have device architecture similar to those of organic solar cells (Figure 1). Anodes are often prepared by depositing indium tin oxide (ITO), which is conductive and transparent and which displays a high work function, onto a flexible plastic or glass substrate. The conducted polymer poly(3,4-alkenedioxythiophenes):poly(styrenesulfonate) (PEDOT:PSS) provides an anode buffer material that enables efficient hole extraction. Photoactive layers may be prepared by spin-coating a NC/polymer blend solution onto an ITO substrate to form a thin film 100–200 nm thick. A top metal electrode (e.g., Al, Ag, and Ca) is then vacuum-deposited onto the photoactive layer as the cathode.

As with OPVs, the conversion of light energy into electricity takes place in four main steps: photon absorption, exciton diffusion, charge transfer, and charge carrier transport and collection (Figure 2). Both organic semiconductor materials and inorganic NCs can absorb incident light and create bound electron-hole pairs called excitons. The excitons diffuse to the D/A interface and then dissociate into free-charge carriers. The excitons dissociate at the interface if the energy levels of the NCs and the polymer are properly aligned. This charge transfer process is necessary for creating a free-charger carrier. After charge separation, the electrons and holes are transported to their respective electrodes through percolating pathways. The holes are transported through the conjugated polymer, and the electrons are transported through the inorganic semiconductor (Figure 2).

The mechanism can be broken down into a number of steps, each of which may be characterized by an efficiency (), which is defined on a scale of 0 to 1. The successful operation of a photovoltaic device requires that most or all of the steps are characterized by close to 1. The overall efficiency of the conversion of incident photons to current, that is, the external quantum efficiency (EQE), can be written as where is the wavelength of the incident light and is the voltage across the cell.

is the photon absorption yield. Most polymers have a bandgap larger than 2 eV, which limits the light absorption range. As such, materials with a complementary absorption spectrum in the near-infrared range [42] or ultraviolet range [57, 58] could be conjugated to the inorganic NCs.

is the exciton diffusion yield. The fraction of excitons that reach the D/A interface is determined by the exciton diffusion length and the location at which an exciton is created with respect to the nearest dissociation center. The exciton diffusion length in both an OPV and an HPV is in the range of 10–20 nm for a conjugated polymer [5961].

is the exciton dissociation yield, which is the ratio of the number of excitons that dissociate to free charges at a D/A interface to the total number of excitons that reach the D/A interface. In a well-designed HPV, the donor and acceptor materials must have suitably been aligned with energy bands that enable exciton dissociation and provide a high overall electrochemical potential. These requirements constrain the type and range of suitable donor and acceptor materials, as well as the interactions between these materials at the D/A interface [2, 5].

is the charge transport yield, which is the ratio of the number of free charge carriers transported to the collecting electrode to the number of excitons dissociated at the heterojunction interface. Donor and acceptor materials are both required for a highly efficient percolated network that spans the entire active layer to provide efficient charge transport. Structural defects, impurities, and the crystallinity of both the donor and acceptor materials in the active layer can cause the charge carriers to become trapped and recombine, which reduces the transport efficiency [2]. Device architecture also require that each phase is continuous throughout the active layer to provide a pathway for rapid carrier transport to the respective electrodes.

is the charge collection yield. This parameter represents the ability of the charges to transfer from the photoactive layer to the electrodes. depends on the energy levels of the active layer and the electrode, as well as the interface properties between them [2].

A low photocurrent in an HPV results from limitations on the parameters and . Our research group has applied significant efforts toward exploiting the high internal surface area and nanoscale dimensions of inorganic/organic nanocomposites as a means for overcoming the limitations of current HPVs.

The power conversion efficiency is one of the most important parameters for characterizing the solar cell performance. Figure 3 shows a schematic diagram of the current density-voltage (-) characteristics of a typical hybrid solar cell in the dark and under illumination. The PCE is given by where is the maximum power point, is the incident light density, and is the fill factor, which is defined as the ratio of to the product of and : The and are two basic factors to determine the solar cell efficiency. An understanding of the physical processes governing these two parameters is needed for the design of new materials and device configurations that would yield a high conversion efficiency [3].

, the short-circuit current density, depends on the incident light intensity and the absorption spectrum of the active materials. is given by [62] where is the spectral photon flux of the incident light, is the bandgap of the active layer, and is the photon energy. Although inorganic acceptors can absorb light at certain wavelengths, the majority of light absorption usually takes place in the donor polymer. Dayal et al. [13] reported that the contribution of light absorption from CdSe in a poly[2,6-(4,4-bis-(2-ethylhexyl)-4H-cyclopenta[2,1-b;3,4-b′]dithiophene)-alt-4,7-(2,1,3-benzothiadiazole)] (PCPDTBT)/CdSe hybrid containing about 90 wt% CdSe nanotetrapod was only 34%. Similarly, TiO2 and ZnO only absorb sunlight in the UV range, which is characterized by a lower photon flux than the visible or IR ranges. The calculations may therefore be simplified by assuming that light absorption only occurs in the polymer.

is the open-circuit voltage. in polymer inorganic hybrid solar cells was found to depend on the difference between the polymer’s highest occupied molecular orbital (HOMO) and the inorganic acceptor conduction band [71]. Due to the quantum confinement effect, the bandgap of inorganic semiconductors varies as a function of particle size, leading to a shift in the conduction band energy level [72], which can also affect the . The of a hybrid solar cell can be increased by either moving the polymer HOMO farther away from the vacuum level or pushing the inorganic acceptor conduction band closer to the vacuum level, while retaining an energy offset between the polymer’s lowest unoccupied molecular orbital (LUMO) and acceptor conduction band larger than the exciton binding energy () [3]. In a bulk heterojunction solar cell, the theoretical maximum can be described as The conversion efficiency may be increased by optimizing the balance between the donor gap (which mainly determines ) and the donor conduction band (which mainly determines ).

3. State-of-the-Art in Hybrid Photovoltaic Materials

Bulk heterojunction hybrid solar cells lag behind the fullerene derivative-based OPVs with respect to device performance due to the limits of the current. Enhanced PCEs in bulk heterojunction HPVs may be achieved by increasing the D/A interface area, which improves the efficiency of exciton dissociation and charge transfer, and by creating interpenetrating bicontinuous percolating pathways for effective charge transport to the corresponding electrodes. Therefore, interfacial behavior and nanoscale morphology of an active layer are these critical performance factors for HPVs. Increase of D/A interface and control over the nanoscale morphology of composite layer are required; however, many issues must be overcome. First, blending inorganic NCs and organic conjugated polymers remains challenging. Dispersing inorganic NCs in a polymer matrix requires the presence of a capping agent that prevents particle aggregation but usually suppresses exciton dissociation and charge transport. Another issue is that charge transport through the composite phase is both highly sensitive to the NCs structure and to the presence of trap states within the NCs. Fortunately, significant progress toward enhancing the PCEs of HPVs has been made by optimizing colloid synthesis [7376] and self-assembly [77, 78] procedures for preparing the NCs, as well as by tuning the shapes of the NCs (dots [7981], rods [8284], tetrapods [13, 85, 86], hyperbranched structures [87], wires [56], etc.). These approaches seek to prepare continuous pathways for charge transport and reduce the prevalence of carrier traps on NCs. Moreover, quantum size effects [72] may be harnessed to tune the device performance by designing the relative alignment of the energy levels in the donor and acceptor materials [84, 88].

In this review, we focus on three strategies for improving the PCE in an HPV: ligand exchange, grafting, and direct NC growth. These methods have been used to improve the polymers/NC interface properties and to control the blend morphology. The characteristics of bulk heterojunction HPV devices are summarized in Table 1.

3.1. Nanocomposites Prepared by Ligand Exchange Chemistry

Two distinct routes may be taken to produce NCs: physical approaches, in which the NCs are fabricated by lithographic methods, ion implantation, or molecular beam deposition, or chemical approaches, in which the NCs are synthesized by colloidal chemistry in solution. The unique optical and electrical properties of colloidal semiconductor NCs have attracted significant interest and have been explored in a variety of applications, such as optoelectronic devices [89], sensors [90], and photovoltaics [2, 6]. Colloidal NCs synthesized in organic media (e.g., alkyl thiol, amines, phosphines, or phosphine oxides) are usually soluble in common organic solvents and can be mixed together with conjugated polymers, which tend to be soluble in the same solvents. Most organic surfactants and ligands tend to be insulated, which impedes charge transfer between the polymers and the NCs and impedes electron transport between adjacent NCs. In the absence of passivating ligands, it is difficult to control the composite morphology because the inorganic NCs tend to be poorly soluble in polymer matrices. The performances of such devices are significantly reduced.

In 1996, Greenham et al. investigated the effects of a QD capping ligand on the initial charge transfer process between a polymer and the CdSe QDs by simply and physically mixing the polymer and NCs [91]. By considering effective luminescence quenching as a manifestation of the exciton dissociation, these authors observed that the quenching of the photoluminescence (PL) of poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylenevinylene] (MEH-PPV) did not occur when 4 nm diameter CdSe NCs were capped with a long alkyl chain, such as trioctylphosphine oxide (TOPO), but that PL quenching was efficient after treatment with pyridine. The authors proposed that the lack of PL quenching was due to the ligand-covered NCs creating a barrier layer that prevented the NCs from approaching the polymer (reducing the interfacial area). The 11 Å thick TOPO alkyl barrier surrounding the CdSe NCs was sufficient to prevent charge transfer [91]. Long chain ligand-capped NCs could be exchanged with short chain ligands to enhance the interface area.

Pyridine ligand exchange is commonly used to improve the efficiency of hybrid solar cell performance. Long alkyl chain-capped NCs are generally washed with methanol several times and then refluxed in pure pyridine at the boiling point of pyridine for 24–48 h. Pyridine treatment appears to replace the insulating ligand, and the effects of pyridine exchange have been examined in the context of poly(3-hexylthiophene) (P3HT):CdSe [79, 92], PCPDTBT:CdSe [13], and MEH-PPV:CdSe [81] bulk heterojunction devices. These studies indicated that replacing the insulating ligands on the NCs with pyridine favored electron transport between NCs by reducing the insulating effects of the ligand and more intimated the contact with NCs and polymer, leading to improved , , and . Pyridine ligand exchange is a standard technique for preparing NPs suitable for polymer/CdSe solar cells. It is not universally suitable, however, because some polymers (such as P3HT) are not soluble in pyridine and the mechanism of ligand exchange process is still unclear [92].

The identity of the capping ligand strongly affects the degree of phase separation and the morphology of a NC/polymer thin film. A series of capping ligands (tributylamine, oleic acid, pyridine, stearic acid, and butylamine) [63, 93] have been tested in a study of their effects on the morphology and - characteristics of a P3HT:CdSe HPV device. The highest PCE (up to 1.8%) was obtained by using butylamine-capped CdSe NPs with a w/w mixing ratio of 12 : 1 (CdSe:P3HT) and a posttreatment temperature of 110°C. Butylamine has the advantage of providing NCs that are soluble in typical polymer solvents and of inducing the formation of a composite phase with small domains on the order of the exciton diffusion length (Figure 4). Therefore, there is considerable room for engineering ligands that can improve exciton separation, enhance the charge transfer efficiency at the polymer/NC interface, and form percolating electron transport pathways to the cathode [94].

After ligand exchange with short chain ligand, NCs tend to aggregate and precipitate out of organic solvents, which complicate the preparation of stable mixtures of NCs and polymers. Zhou et al. applied a novel postsynthetic treatment method to spherical CdSe QDs in which the NCs were washed with hexanoic acid without inducing ligand exchange [64]. The PL quenching, TEM, and dynamic light scattering (DSL) measurements suggested that the ligand sphere had been reduced during the washing step, thereby improving the photovoltaic device efficiency (Figure 5). One advantage of this approach is that the QDs retained their solubility after acid treatment, which allows a high concentration of the CdSe QDs in P3HT (increasing ). The large amount of CdSe led to the formation of efficient percolation networks during annealing of the photoactive composite film. Solar cells achieved efficiencies of 2.0%, which is the highest value yet reported for devices prepared using quasispherical CdSe NPs conjugated with a polymer.

Another strategy for reducing the insulating properties of the ligand in a polymer/NC hybrid material involves the application of a thermal treatment to remove weakly bound ligands [43, 65]. Seo et al. [65] replaced the TOPO ligand on CdSe NCs with tert-butyl N-(2-mercaptoethyl) carbamate ligands for the preparation of a P3HT:CdSe BHJ device by using a solvent exchange reaction. The mixture in a blend solvent was then coated onto the substrate. After heating above 200°C, the ligand was thermally cleaved with isobutene, and carbon dioxide evaporated from the film. The remaining 2-mercaptoethylamine ligands were considerably smaller than the original ligand, and the electrical characteristics of the device improved due to enhanced contact between the CdSe particles and the polymer (Figure 6). These processes increased and by forming better contact with the donor. The PCE could be increased from 0.21% to 0.44% by increasing the heat treatment temperature from 150°C to 250°C, which increased and . This approach was generally less successful than the replacement of TOPO with pyridine because the glass transition temperature of P3HT is below 200°C [95]. The use of ligands with a low boiling point and weak attachment properties to NCs could potentially lead to the facile preparation of multilayer devices.

The side chains of the conjugated polymers play an important role in regulating charge transfer [96]. Enhanced charge transfer at the interface of a composite may be achieved by preparing a polymer/NC blend via ligand exchange, in which the insulating capping ligand on the NC surfaces is exchanged with a functionalized polymer. This exchange process relies on the use of polymers having strongly coordinating functional groups that can anchor directly on the NC surfaces. Liu et al. prepared hybrid materials comprising NCs and a P3HT polymer matrix using an end-functional amino group [54] (Figure 7). The original surface ligands capped onto the NC surface were replaced with pyridine. The CdSe nanorods (NRs) were then mixed with either a nonfunctionalized P3HT polymer (polymer 1) or an amino end-functionalized P3HT polymer (polymer 4). Interestingly, the amine-functionalized P3HT provided a PCE of 1.5%, whereas the device fabricated using the nonfunctionalized P3HT yielded a PCE of 0.5%. These results suggested that the temporary pyridine ligands were exchanged with the polymeric ligands through covalent interactions that enhanced the miscibility of the NRs in the P3HT. Chemical linkages between the polymer and the NCs offer a route to improving the NP dispersion in solvents and the electronic interactions between the polymer and NCs.

However, the amine-terminated P3HT is not a sufficiently strong functional group to fully passivate the CdSe surface capped with TOPO without the need for pyridine as an intermediate ligand. To avoid the intermediate step, which may affect the quality of the final composite material, a strong type of functional group polymer was used to directly be exchanged with the original ligand. In fact, the hybrid TOPO-capped CdSe NPs were successfully embedded in the phosphoric acid-terminated P3HT by simply mixing the components in chloroform and permitting the reaction to proceed overnight [97]. The success of bonding between P3HT and CdSe was confirmed by 1H NMR. A total of 50 P3HT chains were estimated as being present on one CdSe NP, based on an analysis of the absorption spectrum. The PL quenching of the grafted P3HT and CdSe NPs clearly indicated the occurrence of charge transfer at the interface, suggesting that the electronic interactions between the functional components could be facilitated by the end-functionalizing polymers having strong binding groups.

In addition to the CdSe NCs, metal oxides, such as ZnO or TiO2, were tested as candidate materials in green solar cell devices. TiO2 NRs covered with TOPO ligands were synthesized for use in a hybrid solar cell. Although the PCEs of these devices were 1.14% for the pyridine ligand-capped NRs [98] or 0.03% for the dye-capped NRs [57, 99, 100], the charge separation and transport efficiency improved upon the removal of the insulating surfactant in the hybrid materials.

In 2004, Beek et al. [101] reported the performance of bulk heterojunction hybrid solar cells fabricated based on ZnO NCs and poly[2-methoxy-5-(3′,7′-dimethyloctyloxy)-1,4-phenylenevinylene] (MDMO-PPV). These cells yielded a PCE of 1.4% at AM 1.5. The ZnO NPs could be dispersed in dichloromethane, chloroform, or chlorobenzene up to a concentration of 70 mg mL−1 without the need for additional ligands or surfactants. However, the PCEs of the metal oxides remained low due to the presence of surface defects and large ZnO NC aggregates, which increased the recombination. Two years later, Beek et al. reported their work on ZnO NP:P3HT bulk heterojunction HPVs. The highest PCE of these devices was % [102]. The measured PCE was lower than that obtained for MDMO-PPV:ZnO, although the hole mobility of P3HT was expected to be higher than that of MDMO-PPV. The presence of ZnO may have influenced the crystallization of P3HT because the ZnO surface was hydrophilic and the small number of ZnO clusters could not form charge transport pathways to the electrodes [103]. Many approaches to ZnO surface modification, such as using surfactants [104106], anchoring molecules [107109], dipolar molecules [110, 111], and dyes [112], are available to improve the dispersion of ZnO in a polymer matrix. However, surface-modified ZnO does not necessarily lead to a higher PCE because the poor phase separation can hinder charge transport. Control over particle solubility and the blend morphology remains a key requirement for improving ZnO/polymer device performances.

Further extension of the absorption spectrum into the infrared regime could be achieved using PbS or PbSe NCs. Zhang et al. [113] demonstrated HPVs based on PbS and MEH-PPV. They compared the photovoltaic performance of MEH-PPV composite device containing either OA-capped or octylmaine-capped nanocrystals after postannealing the composite layer at 220°C. The improvement in the photovoltaic performance of device using octylamine ligand-capped PbS NCs showed an 200-fold increase in JSC, while devices using OA-capped NCs do not. The thermogravimetric analysis (TGA) data showed a 5% weight loss after heating to 200°C for octylamine-capped NCs, while no appreciable weight loss was observed for OA-capped NCs below 300°C. These results suggested that a certain amount of octylamine ligand is removed from the film during the annealing process (the boiling point of actylamine is 175°C), thus improving the efficiency of the charge transfer at the interface. Unfortunately, the device efficiency was very low because of the large ligand which is still surrounding the PbS NPs and unsuitable relative alignment of the energy levels in the donor and acceptor. Noone et al. [114] utilized a new polymer, poly(2,3-didecyl-quinoxaline-5,8-diyl-alt-N-octyldithieno[3,2-b:2′,3′-d]pyrrole) (PDTPQx), and the OA ligands of PbS NCs were replaced by butyl amine ligand. This device of PDTPQx:PbS (10:90 w/w) exhibited PCE of ~0.55%. More recently, the blending OA-capped PbS NCs with a low bandgap polymer, poly(2,6-(N-(1-octylnonyl)dithieno[3,2-b:20,30-d]pyrrole)-alt-4,7-(2,1,3-benzothiadiazole)) (PDTPBT) was directly exchanged with a short length cross-linker molecule, 1,2-ethanedithiol, Seo et al. significantly improved the device efficiency to a high value of 3.78% [115].

In spite of the improvement in PbS-based hybrid solar cells, the device engineering on PbSe-based HPVs has been difficult, with a low PCE of ~0.1% to date [41]. Although the efficiency of HPVs based on PbSe is low, it has been demonstrated that Pb(Se,S) NCs based quantum dot solar cells show promising efficiency up to 7% [116, 117].

3.2. Nanocomposites Prepared via a Grafting Process

Hybrid materials may be prepared by grafting macromolecules onto NC surfaces via specially designed linker ligands containing an anchor functionality (for NC binding) and an additional reactive group capable of reacting with side- or end-functionalized macromolecules. The grafting process was realized as follows: first, the original NC surface ligands were replaced with the linker ligands; the grafting reaction was then carried out according to the reactive group type in the linker ligand and in the grafting macromolecule. An example of this process was discussed by Zhang et al. [66]. CdSe NRs were grafted to vinyl group-terminated P3HT groups via a p-bromobenzyl-di-n-octylphosphine oxide (DOPO-Br) linker group (Figure 8). In the first step, the original TOPO ligands were exchanged with pyridine. The pyridine ligands were then replaced with arylbromide-functionalized phosphine oxides. The grafting reaction was then carried out via Heck coupling between the vinyl-terminated P3HT. The solid state PL measurement of the thin nanocomposite films revealed PL quenching of the P3HT, which was indicative of charge transfer between the P3HT and CdSe NRs.

Bifunctional linker ligands have recently been tested for their utility as original NC ligands. This strategy reduces the costs and time needed for preparing hybrid materials by performing the grafting in a single step without a ligand exchange step. The bifunctional ligand (i.e., DOPO-Br) contains a phosphine oxide group at one end, similar to the phosphine group on the TOPO, which anchors the linker to the QD surface, and an arylbromide is present at the other end to enable grafting to a macromolecule [118]. CdSe QDs were grown in DOPO-Br to directly yield the DOPO-Br capped CdSe QDs. The PPV derivatives were synthesized by polymerizing 1,4-divinyl benzene with 1,4-dibromobenzene derivatives. Poly(p-phenylene vinylene) (PPV) derivatives have been directly grafted onto a DOPO-Br functionalized CdSe QD surface via Pd-catalyzed Heck coupling [67] (Figure 9). By replacing PPV derivatives, the success obtained from grafting the vinyl-terminated P3HT to the DOPO-Br functionalized CdSe QDs was expected to apply for photovoltaic devices. The end group P3HT contacted the CdSe QDs and interacted with them directly to facilitate charge transfer from the P3HT to the CdSe at the interface. PL quenching and the short fluorescence lifetime of the P3HT:CdSe composite, compared to the corresponding values obtained from P3HT alone, confirmed that the charge transfer was effective. This system suffered from some unfortunate drawbacks: the DOPO copolymerization conditions were difficult to be controlled, and this strategy cannot be extended to the synthesis of CdSe NRs because the DOPO-Br capping group is not suitable for inducing the growth of CdSe NCs.

Briseno et al. reported a convenient method for directly attaching end-functionalized polymers to NCs with bare surfaces without the need for ligand exchange and/or direct bifunctional ligand growth processes [55]. P3HT and didodecylquaterthiophene (QT), terminated with phosphonic ester and phosphonic acid, respectively, were chemically grafted onto an N-type ZnO nanowire (NW, several micrometers in length and 30–100 nm in diameter) via self-assembly of the semiconductors onto the ZnO surface in the solution phase to yield an organic shell with a thickness of about 5–20 nm (Figure 10). A single NW solar cell was successfully prepared using p-n core/shell NWs. Although the PCE of the single NW solar cell was low, 0.036%, this device permitted the isolation and study of the parameters that affect bulk hybrid solar cell performance. Increasing the P3HT layer thickness on the ZnO surface significantly improved the performance of the NW device; however, the in the single wire devices ( = 0.4 V) was larger than the value reported in the literature, = 0.17 V for a P3HT:ZnO bulk NW array device, suggesting that the P3HT:ZnO interface in the grafted polymer was superior to the bulk spin-coated counterpart.

To date, the grafting approach offers a high grafting coverage density that increases the effective organic/inorganic interface to enable rapid charge separation. The grafting approach relies on the design of bifunctional ligands and functioned polymers that facilitate the controlled growth of NCs and the efficient grafting of NCs, respectively. However, the surface grafting and polymerization steps are usually complicated. Excessive grafting reduces the efficiency of electron transport between adjacent NCs; thus controlling the surface grafting processes remains a challenge.

3.3. Nanocomposites Prepared via an In Situ Process

A new strategy was developed for inducing NC growth in a polymer matrix in an effort to avoid organic capping ligands and improve charge transfer between the NCs and the organic matrix. This approach relies on the dissolution of the NC precursors in a solution containing the polymer, and the NCs are then grown in the polymer template. This strategy has been applied to a variety of semiconductor NCs, including TiO2 [119], ZnO [120, 121], PbS [44, 70], CdSe [122], and CdS [51, 69, 94].

The in situ NC growth method was first tested by preparing a TiO2:MDMO-PPV HPV, in which titanium isopropoxide was used as the precursor. A BHJ device was expected to provide a higher efficiency than a bilayer device; however, the results of the initial test did not confirm this assumption [123]. The low efficiency of the in situ blended devices could be explained by the noncrystalline structure of the TiO2 semiconductor. High-temperature annealing can produce crystalline TiO2, but the high temperatures are not compatible with organic semiconductors.

This strategy was applied to the fabrication of ZnO/polymer HPVs because crystalline ZnO can be formed at low temperatures. The work of Beek et al. provides an instructive example of this approach. The photoactive layer composed of ZnO:MDMO-PPV composites was prepared by spin-coating a mixed solution containing the precursor diethylzinc and MDMO-PPV [120]. During spin-coating, diethylzinc was exposed to moisture and formed Zn(OH)2 upon undergoing spontaneous hydrolysis. The thin film was then annealed at 110°C to yield an interpenetrating network of ZnO inside the MDMO-PPV. These devices offered a PCE of 1.1%, less than the value obtained for ZnO NPs mixed with MDMO-PPV (PCE = 1.6%); however, the reached a high voltage of 1.14 V. Such a high voltage has not previously been reported. Degradation through breakage of the transvinyl bonds in the PPV polymer upon mixing with the ZnO precursors produced large undesirable effects on the charge carrier [121]. The more stable P3HT did not display changes in the absorption spectra and hole transport properties during processing, providing a PCE of up to 1.4% for the ZnO:P3HT device. The improved photovoltaic performances mainly resulted from the large increase in the . The highest PCE yet obtained for an ZnO:P3HT HPV fabricated using this precursor method is 2% [52].

Oosterhout et al. investigated the use of an ester-functionalized P3HT-E that was more compatible toward the hydrophilic ZnO NCs [68]. The P3HT-E produced a much finer phase separation and a higher surface area upon blending with ZnO (Figure 11). increased due to the enhanced and ; however, better mixing in the ZnO:P3HT-E composite appeared to reduce the ZnO connectivity and hole mobility in the polymer due to the lower degree of crystallinity in the P3HT-E phase. Only thin devices displayed current and PCE enhancements.

HPVs fabricated using small bandgap poly(3-hexylselenophene) (P3HS) in combination with ZnO were expected to yield a higher PCE [124]; however, the PCE of ZnO:P3HS was 0.4% due to the lower . The spectrally resolved external quantum efficiency was compared with the optical absorption spectrum, revealing that the amorphous P3HS regions were in direct contact with the ZnO NCs, whereas the semicrystalline P3HS phase contributed negligibly to the photocurrent. , therefore, decreased due to a reduction in . The authors indicated that the interface in ZnO:P3HS was disordered, similar to the P3HT disorder observed in ZnO:P3HT devices. The polymers rapidly degraded when mixed with diethylzinc in solution [124].

Leventis et al. introduced another in situ method for fabricating metal sulfide NP/polymer films [69]. Metal xanthate precursors were selected because they decompose at low temperatures into a metal sulfide, generating only volatile side products (e.g., H2S, COS, C2H4). The decomposition process enabled the in situ growth of inorganic NCs in a relative fragile polymer host. The CdS centers formed an extended network rather than particles, which would have increased , had the materials not also vertically segregated, thereby reducing (Figure 12). The length scale of the phase segregation process depended strongly on the annealing temperature [125]. An increase in the annealing temperature from 105°C to 160°C led to a reduction in the sizes of CdS domains embedded in the P3HT matrix, from 100 nm to below 40 nm. A highly mesoporous structure formed to include interconnected CdS NP aggregates approximately 30–50 nm in diameter. The resulting photovoltaic devices exhibited a PCE of 2.17%. The in situ thermal decomposition of a single source precursor within a solid-state polymer showed promise except that aggregates formed and reduced the effective polymer/NC interface area.

Another approach to directly growing NCs involved using the polymer as a template for NC growth. The quality of the obtained NCs depended on the electrostatic and steric effects imposed by this polymer [51]. Cadmium acetate dihydrate and P3HT were dissolved in mixed solvents containing dichlorobenzene (DCB) and dimethyl sulfoxide (DMSO). Sulfur in DCB solution was then injected into the conjugated polymer and precursor solution at 180°C for 30 min to form CdS NCs in the P3HT matrix. The purification of CdS/polymer composite was achieved by removing excess cadmium, sulfur ions, and DMSO via adding anhydrous methanol to form the precipitate. After centrifugation, the supernatant was then removed and the composite redissolved in DCB. The solar cells with a device architecture of ITO/PEDOT:PSS/P3HT:CdS/Al with an active layer prepared following the redissolution composite solution. In this approach, the CdS single crystal NRs having different aspect ratios were easily obtained by controlling the cosolvent mixture ratio. In fact, the cosolvent ratio could be used to modify the architecture of the polymer templates because the physical conformations of the P3HT chains were affected by the solvation properties. The planar P3HT conformation provided a stacked molecular architecture (Figure 13(a)). The Cd2+ ions may have been confined within the network through dipole-dipole or ion-dipole interactions between the Cd2+ ions and the S atoms, leading to the formation of uniformly and randomly distributed CdS NCs and NRs within the P3HT matrix. At low concentrations of the Cd2+ ions, CdS NRs could not be formed (Figure 13(b)), although NR morphologies started to form at higher concentrations (Figure 13(c)). Improved PL quenching was demonstrated for CdS NRs having a high aspect ratio as a result of the high interface area, which formed pathways for electron transport. A device composed of P3HT and CdS NRs with an aspect ratio of 16 showed a PCE of 2.9%.

First reports on the synthesis of PbS NCs in the solution containing a conjugated polymer were from Watt et al. in 2004 [126]. They describe the in situ preparation of PbS NCs in the presence of a conjugated polymer, either MEH-PPV [126] or P3HT [127], by heating a solution of lead acetate, elemental sulfur, and the polymer in a solvent mixture of DMSO and toluene to 160°C for 15 min. Removing excess lead and sulfur ions by precipitation and redissolution followed by coating on substrates completes the preparation of the nanocomposite layers. An advantage of this synthesis route is that size and shape of the NPs can be tuned quite easily. A controlled formation of nanosized PbS particles in the presence of the conjugated polymer was shown and attributed to steric effects of the long polymer chains. Without the polymer, “bulk” PbS was formed in the reaction solution. Furthermore, Stavrinadis et al. [70] showed that nanorod-like structures can be formed in a MeH-PPV matrix by a post synthetic treatment (Figure 14). A precipitation of spherical PbS NPs in MEH-PPV solution using alcohol with the appropriate polarity like ethanol, propanol, or hexanol led to dipole-induced oriented strings of PbS NPs, although a precipitation in methanol led to PbS nanocubes.

The solar cell prepared PbS NPs exhibited a of 1 V, an of 0.13 mAcm−2, and a of 0.28, which leads to a PCE of 0.7% under 5 mWcm−2 illumination [128]. The moderate and were ascribed to a high series resistance. A possible reason could have a relation large particle-particle distance of the PbS NPs.

Unlike the nanocomposites crafted by ligand exchange and direct grafting, the NC size and shape, the polymer/NCs interface, and the dynamics of NC growth in the in situ-prepared hybrid materials were complex and the interactions between the polymer and the NCs are poorly understood. The in situ approach has received significant attention due to its simplicity and its potential for producing high-efficiency hybrid solar cells; however, additional work may be required to optimize the in situ growth approach.

4. Outlook and Conclusions

The polymer and inorganic components in hybrid photovoltaic devices offer complementary advantages, although the PCEs of these devices are currently low (~3.8%). The fabrication of hybrid layers that yield both a high charge generation efficiency and a high charge collection efficiency faces several challenges. The efficiency of charge generation may be improved by improving the extent of mixing between the donor and acceptor semiconductors by providing efficient exciton dissociation at the donor/acceptor interface. The controlled nanoscale morphology of the hybrid layer is required for the formation of an interpenetrating network for efficient charge transport. This review discusses three strategies for preparing HPVs: ligand exchange, grafting, and direct NC growth. These approaches are widely used to improve the compatibility of polymers and inorganic NCs of various types and dimensions in an effort to improve the photovoltaic performances of hybrid solar cells.

Modifying the surfaces of NCs or polymers is a common approach to improving the quality of the hybrid active layer in an HPV. The modifier group can help passivate the surface charge traps, suppress aggregation of the NCs in the polymer matrix, and tune the energy level offset between the semiconductor and organic layer. Nevertheless, most ligands form an insulating barrier that prevents charge transfer between the polymers and the NCs and prevents electron transport between adjacent NCs. The properties of the NC surface ligands are the principle factors limiting the efficiencies of these layers. The ligands must be removed or exchanged with more efficient surface modification molecules to improve phase separation, facilitate charge transfer, and hinder back recombination. Ligand exchange effects are not yet well understood, and precise control over certain factors, such as the exchange rate, has proven to be difficult. For the time being, the use of “small” molecules that reduce the interface surface energy has been used to replace the original ligands containing long insulating alkyl chains. The small molecules improve the efficiency of charge transfer and charge transport. For example, ZnO-modified fullerene carboxylic acid (PCBA) has been tested [108, 110]. Modifications to the ZnO NP surfaces increase the device performance from 0.59% to 1.20%. Electronic coupling between the ZnO and C60 derivatives increased to 5.39 mAcm−2 by enhancing charge separation and improving the compatibility of the materials at the interface.

Grafting strategies are promising because they offer a high grafting density that increases charge separation. The development of controlled surface grafting processes is still needed to avoid excessive grafting, which reduces the efficiency of electron transport between adjacent NCs. Although much work can be done to optimize this approach, controlling the grafting density and the location is complicated. The use of polymer/inorganic nanocomposites blended with carbon nanotubes (CNTs) [129131] or graphene [132, 133] has been extremely attractive for improving the PCEs of HPVs. Semiconducting CNTs have a high absorbance in the visible and near infrared (NIR) region and a high charge mobility (up to 105 cm2V−1s−1 in individual nanotubes [94]). CNTs doped or functionalized can form nano-assemblies with other molecules and polymers to provide the efficient photoinduced charge transfer or to tune the Fermi levels and to improve the positioning of the CNTs in the heterojunction region near the semiconductor [71]. In the recent study, Derbal-Habak et al. [134] demonstrated that the incorporation of 0.2% of functionalized SWCNTs into P3HT/PCBM conventional OPV improves the PCE of ~3.7%. Fortunately, Wang et al. [135] created a P3HT-PbS QDs-MWCNT nanohybrid. This device exhibited a large enhanced conversion efficiency of 3.03% as compared with 2.57% for standard P3HT:PCBM OPV. The critical role of semiconducting CNTs may improve charge transport by forming continuous pathways directly from the NCs to the electrodes that avoid the inefficient hopping pathways along the NC networks. Interestingly, a high PCE of ~14% in CNT: Si hybrid solar cells has been reported [136]. These results suggest that flexible solar cells may potentially be prepared by replacing p-type Si wafers with amorphous Si thin films.

Finally, the direct in situ growth of NCs in a polymer matrix offers an interesting approach to obtaining blends of NCs and conjugated polymers due to its simplicity and recent advances in NC synthesis. This strategy still faces several challenges. The NP size, shape, and distribution within the polymer matrix significantly influence the dielectric, electronic, optical, and structural properties of the composite layer, and the quality of NCs synthesized in situ has been rather low. This problem can be resolved by controlling the growth of NCs within the polymer matrix. A template approach employing diblock copolymers has been considered as a means for guiding the growth of NCs [137139]. Rod-like ZnO NCs composed of ZnO NPs have been formed through templated self-assembly using a poly(3-hexylthiophene)-b-pol(zinc methacrylate acetate) (P3HT-b-PZnMAAc) diblock copolymer containing rigid thiophene chains and flexible chains [139]. The dipole-induced interactions between adjacent ZnO NPs promoted the assembly of rod-like ZnO NCs from ZnO NPs by aligning the crystal orientations and forming an interpenetrating network that offered efficient charge separation and transport. Rod-like ZnO NCs were homogeneously dispersed in the polymer matrix without the need for a surfactant, and the lengths of the rod-like ZnO NCs could be controlled by adjusting the hydrolysis time. The PCEs of the P3HT:ZnO composite devices were 0.19%, although the weight fraction of ZnO was very low (10%). A polymer having both electron accepting and electron donating building blocks has been used for the purpose of controlling the morphology in an HPV. A hybrid solar cell prepared using a liquid crystal donor/acceptor copolymer, poly[3-(6-(cyanobiphenyoxy)thiophene)-alt-4,7-(benzothiadiazole) (P3HbpT-BTD) and ZnO NPs yielded a PCE as high as 1.98% [138].

Hybrid solar cells are still lagging behind PCBM-based OPV technologies with respect to device performance and commercial applicability. HPVs are currently under development and evaluation at the basic research level, and they have the potential to undergo further significant improvements. Novel device structures, the implementation of nanostructuring methods, and the development of low bandgap materials will probably lead to PCE values on the order of 10%. Progress toward the development of organic-inorganic hybrid materials will benefit the development of hybrid solar cells as well as a variety of other applications, such as light emitting diodes, photodiodes, photodetectors, fuel cells, catalysts, and sensors.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Acknowledgments

This research was supported by the Mid-career Researcher Program and Global Research Laboratory Program through the National Research Foundation of Korea, funded by the Ministry of Science, ICT & Future Planning (No. 2010-0029244).