A Review of the Improvements in the Performance and Stability of Ternary Semi-Transparent Organic Solar Cells: Material and Architectural Approaches

Abstract

In the past decade, considerable efforts have been made to develop semi-transparent organic solar cells (ST-OSCs). Different materials and architectures were examined with the aim of commercializing these devices. Among these, the use of ternary active layers demonstrated great promise for the development of efficient semi-transparent organic solar cells with the potential for future applications, including but not limited to self-powered greenhouses and powered windows. Researchers seek alternative solutions to trade-off between the power conversion efficiency (PCE) and average visible transmittance (AVT) of ST-OSCs, with photoactive materials being the key parameters that govern both (PCE) and (AVT), as well as device stability. Several new organic materials, including polymers and small molecules, were synthesized and used in conjunction with a variety of techniques to achieve semi-transparent conditions. In this review paper, we look at the working principle and key parameters of semi-transparent organic solar cells, as well as the methods that have been used to improve the performance and stability of ternary-based semi-transparent organic solar cells. The main approaches were concluded to be spectral enhancement and increments in the transparency of the active layer through band gap tuning, utilizing novel organic semi-conductors, optical engineering, and the design architecture of the active layers.

1. Introduction

In recent decades, people have been especially concerned about pollution from energy sources and greenhouse gases. In order to reduce these effects, many countries have put money and effort into the field of renewable energy. Among the renewable energy sources, solar energy is the most affordable, unpolluted, and abundant of all long-term natural resources up to date [1]. Solar photovoltaic devices are among the most efficient means of turning sunlight into direct current in solar cells [2]. Solar cell technologies can be classified into organic- and inorganic-based semi-conductors [3]. In the present day, silicon-based solar cells have been commercialized and have grown to dominate solar cell technology, despite the limitations that have hampered the device’s efficiency, such as poor absorption in the long wavelength range, cost-effectiveness, and highly complex fabrication procedures. Furthermore, the thickness and bandgap of the active layer in these devices have a vital role in modifying the optical and electrical properties of photovoltaic devices [4,5]. For instance, thick films absorb more of the sun spectrum and form photo-generated carriers, but due to the likelihood of recombination, it is difficult to collect these carriers [6,7].

Organic/polymer solar cells (OSCs) were chosen as a viable alternative for photovoltaic applications. This is because polymer solar cells (PSCs) draw vast interest due to their low cost, easy fabrication, transparency, flexibility, and capability in terms of fine-tuning the bandgap energy. For highly efficient solar cells, conjugated polymers were used as a raw material in the active layer of polymer solar cells with a bulk heterojunction (BHJ) structure. In addition, polymer solar cells have proven to be practical and cost-effective solar cells for portable electronic devices and building-integrated photovoltaic devices [8,9,10]. Nevertheless, due to intrinsic and extrinsic factors, stability and efficiency have not reached the benchmark for commercialization. Several significant endeavors have been carried out to improve stability and efficiency by means of modifying their active materials and device architecture [11,12,13]. One of the accomplishments in improving power conversion efficiency (PCE) while preserving flexibility, easy fabrication, and low cost is the ternary system for the bulk heterojunction active layer. The binary system’s narrow window of absorption spectra of organic materials reduces device efficiency, whereas the ternary system improves light harvesting by broadening the spectral response of the active layer. Hence, this approach enhances the short circuit current density (J_sc) and film morphology, and, as a result, the performance of the device is amended.

A ternary system comprises a combination of two donors and an acceptor, or a donor and two acceptors, or a donor and an acceptor with a third component, such as small molecules or dyes. Therefore, selecting the third component is a remarkable consideration for device performance. The third component determines the mechanism of charge transfer and the compatibility of film morphology, which have a direct impact on device efficiency and stability [14,15,16].

The ability of organic materials to tune their bandgap energy revealed a new field of practices for niche applications of building-integrated photovoltaics (PVs), such as self-powered greenhouses and powered windows, known as semi-transparent organic photovoltaics (ST-OPVs). Despite the fact that ST-OPVs have not yet been commercialized, the trade-off between device efficiency and average visible transparency (AVT) remains a challenge, and a number of researchers have assigned their research to increasing the efficiency to a commercial standard and optimizing (AVT) as well [17,18,19,20]. A review of the literature showed that there were useful reviews on ternary systems for bulk heterojunction structures [21,22,23,24,25,26,27,28,29,30], as well as several interesting reviews on the progress of semi-transparent organic solar cells, from materials to device performance [31,32,33,34,35,36,37]. For a semi-transparent organic solar cell, several parameters should be taken into consideration, such as film thickness, the active layer materials, and the electrodes. This means that semi-transparent solar cells and ternary blend systems rely heavily on the optical tuning of the active layer materials to facilitate such adjustments [38,39,40,41]. Semi-transparent organic solar cells have undergone significant development in recent years, and this review paper provides a thorough examination of the photovoltaic parameters of these cells, as well as an elaboration on the latest advancements, approaches, and active layer conditions that have contributed to this growth.

2. Working Mechanism of Ternary Organic Solar Cells

Organic solar cells (OSCs) are photovoltaic devices that use organic semi-conductors in their active layer to convert solar energy into electricity. The $\pi$-conjugated backbone of carbon-based materials represents a supreme organic semi-conductor for light harvesting. These materials are defined by the frontier molecular orbitals formed by a parallel overlapping of $\pi$ bonds. The uppermost $\pi$ band is known as the highest occupied molecular orbital (HOMO), whereas the lowermost $\pi *$ denotes the lowest unoccupied molecular orbital (LUMO). In disordered systems of polymers with different types of atoms, several localized states can occur in the excited state. This is due to the fact that the excited quantum states of neighboring atoms are different, and their wave functions do not fully overlap to produce an extended pathway for the excited electrons [42]. In addition, the optical energy band gap is the difference between the HOMO and LUMO levels that determine the material’s optical characteristics. Because of the absorption of electromagnetic radiation, excitons (e–h pairs) are generated. As a result of the low di-electric constant (${\epsilon }_r$) (around 2–4) and the existence of many localized states in the photo-excited state, the mutual Coulomb interaction between e–h pairs will be in the range of 0.3–1 eV (see Figure 1).

Excitons have a binding energy that is larger than the thermal energy present at ambient temperature; therefore, extra energy is needed to dissociate singlet excitons and generate sufficient free charges in the active layer. As a result, a crucial strategy for overcoming the binding energy of excitons and detaching electrons and holes is the use of a heterojunction structure, which comprises electron-donating (donor) and electron-accepting (acceptor) materials. The driving energies for exciton dissociations are provided because of the differences in ionization energy of the materials, which leads to the generation of energy to offset the interfaces. Incidentally, excitons must travel to the interface for the dissociation process, whereas organic materials possess a localized nature at excited states and a low exciton diffusion length of about 5–20 nm. Thus, a thin layer of the active medium is essential for the excitons to reach the boundary of D-A interfaces, where they can dissociate before gemination recombination occurs. Additionally, the excitons travel by random diffusion; this is because excitons are electrically neutral and are, hence, unaffected by an electric field. However, the random diffusion process incoherently derived from both Dexter energy transfer and Förster resonance energy transfer (FRET) takes place intramolecularly or intermolecularly. As a result, singlet excitons experience a short lifetime in the excited state [31,43].

Consequently, it is vital to analyze a trade-off between photon absorption and exciton dissociation in relation to semi-conductors. Bulk heterojunctions (BHJs) were introduced by Heeger’s group to overcome the short diffusion length of excitons, low charge mobility, and weak absorption rate. The BHJ is composed of a mixture of donor and acceptor elements. As previously stated, enlarging the interfacial area is a critical mechanism for exciton dissociation, so any photo-induced excitons appear to be a few nanometers away from the donor-acceptor interface at any point in the interpenetrating network. Figure 2 depicts a schematic illustration of photocurrent generation and the architecture of a bulk heterojunction organic solar cell. Charge generation and transport in the BHJ organic solar active layer may be summarized as follows: (i) excitons are created when photons are absorbed in the donor region. (ii) The excitons diffuse toward the interface of the donor and acceptor moieties, where they dissociate into free charges. (iii) Free carriers pass through the acceptor and donor domains in opposite directions. (iv) The free charges are collected by the anode and cathode electrodes, respectively.

In order to attain the exceptional device performance of the BHJ structure, the absorption efficiency of the active layer, particularly the film thickness, must be enhanced. Most of the organic semi-conductors have low charge mobility, and for efficient charge transportation and collection, the typical thickness of the active layer is about 100 nm. Based on the Beer–Lambert law, as shown in Equation (1) [37], the absorption response is in a direct relationship with film thickness.

$$A=c\epsilon d$$

(1)

where A is the absorbance, $c$ is the concentration of the solution, $\epsilon$ is the molar absorptivity, and d is the optical path length (thickness of the active layer). In order to strike a balance between absorption efficiency and charge extraction, a novel organic solar cell design called ternary organic solar cells (TOSCs) was developed. This design can improve the morphology of the active layer, the device’s stability, the recombination rate, and the voltage loss. In the ternary approach, the active layer consists of blending three components: a donor (D) and an acceptor (A) as a host binary material, plus an additional donor or acceptor. This strategy preserves the solution’s processability and large-scale production, which is a step toward commercialization [21]. Alternatively, the addition of the third component affects the degree of crystallinity and phase separation, leading to the generation of different trap and recombination sites. Thus, identifying the loss mechanism and the careful design of the ternary system are required for an efficient and stable solar cell [21,24]. The recombination process can be categorized into geminate and non-geminate recombination. In the geminate recombination process, the electron–hole pairs recombine before dissociation or during the charge transfer state (CT). However, in non-geminate recombination, free electrons and free holes are recombined when the free charges are drifted toward their corresponding electrodes. Non-geminate recombination is classified into bimolecular, trap-assist, and Auger recombination.

The materials used in the BHJ structure were categorized based on their structure, including polymers, small molecules, nanostructures, and dyes. Therefore, selecting the third component in a ternary system induces more convolution in the charge transport mechanism, whereas the contribution of the third component enhances all of the absorption profiles, exciton dissociation, charge transport, and film morphology. The literature shows that the mechanism of charge transport in ternary systems involves four main operative mechanisms [21,25,27,44]. The transport mechanism involves the energy levels and band gaps of the three components, entitled charge transfer and energy transfer mechanisms, whereas parallel alloy models, rely on the nanomorphology of the active layer. Due to the placement of the inserted components in the ternary active layer, it is remarkable that in some systems, more than one mechanism may exist simultaneously [21,27]. Thus, the location of the third component has a decisive role in defining the transport mechanism of charge carriers, as shown in Figure 3.

In the charge transport mechanism, photo-generated excitons at both donors (D1-D2-A ternary system) split apart at the interfaces between the donors and acceptors, and the electrons and holes are moved along the energy cascades toward their respective electrodes. For charge transport to occur, there are several crucial factors. First, the electronic energy levels of the three components must form an energy cascade alignment, as described in Figure 3a. In choosing the third component, whether its small molecules or dyes, thought should be given to energy levels (HOMO and LUMO) to avoid trap and exciton recombination in the active layer. Hence, it is located between the energy levels of the host binary system. Second, the position of the third component must be located at the interfaces of the host binary system (Figure 3b) and this depends on the crystallinity and the surface energy of the host materials [45,46]. Finally, the weight ratio of the third component has a strong influence on the type of charge transport, and in this model, the binary domains, which act as mediums for free carriers, should be conserved. In order to retain these routes in the binary domain and avoid disruption, a significant augmentation of the third component is required [47,48].

Energy transport mechanisms represent another comprehended transport mechanism that takes place in ternary systems, as shown in Figure 3c. The third component acts as an absorber for harvesting more photons and transferring the energy of excited electrons to the host binary system. This energy exchange between the third component and host materials has been achieved through two main transfer mechanisms called Förster resonance energy transfer (FRET) and Dexter energy transfer (DET). In FRET, the electrons are excited to an excited state through absorption processes and are bound to the nuclei of the molecules. The electrons do not transfer to other molecules, whereas their energy exchanges with the host materials via dipole–dipole interaction significantly depends on spectral overlap and intermolecular distance. However, in DET, electron exchange occurs at a short distance between molecules via a hoping process caused by molecular orbital overlaps. The position of the third component is then an essential feature due to the short intermolecular distance criterion for energy transfer.

As shown in Figure 3d,e, the third component should be in contact with a binary host material (donor or acceptor) to ensure efficient energy transfer and the dissociation of excitons [24,25,49]. As a result, various experimental techniques were used to categorize and distinguish the type of charge dynamics within ternary systems, and photoluminescence measurements (PL) were chosen as a convenient technique to distinguish between charge and energy transfer. As a result, the charge transfer occurs if the emission intensity of the host materials (donor or acceptor) is quenched by entering the third component. However, if the emission intensity of the host donor gradually drops while the emission intensity of the host acceptor steadily increases, this implies that energy transfer happens [24,25]. Nevertheless, due to experimental conditions, such as film thickness, molecular alignment, and measuring angles, some unavoidable errors are produced. Hence, additional techniques are used to determine the type of charge transfer. For instance, transient absorption (TA) spectroscopy and steady-state photo-induced absorption (PIA) were employed to probe charge dynamics [45,50].

Street and Yang proposed a new model for charge transport in BHJ solar cells called the alloy and parallel-linkage models, respectively. As shown in Figure 3f, in the alloy model, the two donors are merged, acting as one donor, and due to the comparable electronic properties, quasi-LUMO and quasi-HOMO energy levels were formed. Moreover, in the charge-transfer state, the electrons from alloy donors are transported through the percolating pathway of the acceptor, and the holes are transported through the percolating pathway of alloy donors toward the respective electrodes. For efficient transfer, the third component should have good miscibility and compatibility with respect to the host materials. Nonetheless, in the parallel linkage model (Figure 3g), the two donors do not merge to form an alloy but individually contribute to charge transfer states. Thus, the photo-generated excitons from both donors migrate toward the interfaces and then dissociate into free carriers and are transported via the donor and acceptor domains (Figure 3h). As a result, the conditions for both models are less problematic when compared to other models, and they are unaffected by the alignment of electronic energy levels and the location of the third component. [23,24,51,52].

3. Solar Cell Parameters

Photoelectric conversion and photovoltaic performance are well-known as the parameters that outline the efficiency of the solar cell. Therefore, each parameter, such as short circuit current density ($J_{sc}$), open circuit voltage ($V_{oc}$), fill factor ($FF$), and power conversion efficiency ($PCE$) describes the performance of the device. The structure and type of the device and the molecular structure of the active layer materials have a crucial impact on the performance of photovoltaic devices. Hence, solar cells can be modeled via a p-n-junction-equivalent circuit to comprehend the photo-generated carriers, recombination, and transport mechanisms.

The photoactive layer in an organic solar cell possesses low di-electric permittivity, and the localized properties of the electronic state at the optically excited state create a built-in potential that hinders exciton dissociation. Therefore, the photo-generated current density, as shown in Equation (2), depends on the absorption profile of the active layer and the solar irradiance spectrum (Figure 4) [53,54].

$$J_{ph}=\frac{q}{hc}\int A\left(\lambda \right)AM1.5G\left(\lambda \right)\lambda d\lambda$$

(2)

where $h$ is the Planck’s constant, $q$ is the electric charge, $A\left(\lambda \right)$ is the spectral response of absorbance of the active layer, $c$ is the speed of light, and $AM1.5G\left(\lambda \right)$ is the spectral irradiance of the sunlight. Hence, due to the external field, the maximum current stream out into the external load is called the short circuit current density $\left(J_{sc}\right)$. With the assumption that all the photo-generated carriers are collected by the respective electrodes without recombination, the photo-generated current density would be equal to the short circuit density [6].

Nevertheless, all photo-generated carriers cannot be migrated toward the respective electrodes due to recombination processes, either at the active layer or at the interface of the active layer/electrodes. In addition, the direction of the recombination current is opposite to the direction of the photo-generated current, and due to the decrease in barrier potential, the recombination current increases until it becomes equal to the photo-generated current. As a result, the current density at the external load becomes zero. At this point, the voltage is called the open circuit voltage $\left(V_{oc}\right)$. Accordingly, the stability between the rate of photogeneration and the rate of recombination determines the $V_{oc}$. Hence, it is extremely dependent on the band gap of the active layer. Herein, the voltage loss from the differences between the energy gap of the active layer and the voltage across the cell $V_{oc}$ is observed. This is due to the existence of radiative and non-radiative recombination and the energy offset at the interfaces between the donor and acceptor, which are required for the charge transfer energy state [55].

Both short circuit current density and open circuit voltage can be obtained experimentally from the $J-V$ characteristics of solar cells, as depicted in Figure 5, by the following equations.

$$P_m=V_m\times J_m$$

(3)

$$FF=\frac{V_m\times J_m}{V_{oc}\times J_{sc}}$$

(4)

$$\eta =\frac{P_m}{P_{in}}\times 100\%=\frac{FF\times V_{oc}\times J_{sc}}{P_{in}}\times 100\%$$

(5)

where $V_m$ and $I_m$ represent the maximum value of voltage and current delivered to the external load. FF is the fill factor, which designates the relationship between the collection and recombination of the free charge carriers. $\eta$ is the power conversion efficiency $\left(PCE\right)$, $P_m$ is the maximum output power of the cell, and $P_{in}$ is the incident light input power [9,56].

Consequently, all the parameters that are used to evaluate opaque device performance can be used for semi-transparent devices, along with several other parameters, such as average visible transmittance (AVT), color properties, quantum utilization efficiency (QUE), and light utilization efficiency (LUE). AVT is a measure of the transparency of the cell in the range of (370 nm–740 nm) with respect to the photonic response to the human eye, and it can be determined from Equation (6). In the window application, the reference value of (AVT) is considered to be 25% [35].

$$AVT=\frac{\int T\left(\lambda \right)\times V\left(\lambda \right)\times AM1.5G\left(\lambda \right)d\lambda }{\int V\left(\lambda \right)AM1.5G\left(\lambda \right)d\lambda }$$

(6)

where $V\left(\lambda \right)$ is the photonic response of the human eye, $T\left(\lambda \right)$ is the transmission spectrum of the device, and $AM1.5G\left(\lambda \right)$ is the photon flux under AM 1.5G light illumination. Moreover, for practical applications, the color properties of the device play a critical role in the visual appearance perceived by the human eye and the perception of color transparency first measured by Ameri et al. using the CIE 1931 xyz chromaticity diagram. In semi-transparent OSCs, the quantitative parameter that shows the degree of color rendering between the illuminations and transmissions of light through the device is called the color rendering index (CRI), and this can be estimated by comparing the natural source of light to the test light source [57]. In addition to the ability of the ST-OSCs in photon harvesting, conversion and visibility (a set of impartial parameters) can be utilized for device performance. The first one is called quantum utilization efficiency (QUE), which is defined by the summation of the external quantum efficiency (EQE) and transmittance (T) of the cell (Equation (7)). Thus, the value of QUE should be less than 90% over the visible region due to essential intrinsic losses. The second metric, which is crucial for calculating the performance of the cell and the best uses of light, is light utilization efficiency (LUE). This can be measured by multiplying the power conversion efficiency (PCE) by the average visible transmittance (AVT), as shown in Equation (8) [38].

$$QUE=EQE+T$$

(7)

$$LUE=PCE\times AVT$$

(8)

4. Photoactive Layer Characterization

Due to its promising uses in semi-transparent photovoltaic systems, polymer solar cells are making a significant contribution to the production of the large-area and cost-effective devices of the future. UV-Vis absorption spectroscopy may be utilized to examine the photophysical properties of the photoactive layer. It is well-established in the literature that the UV absorption bands can be attributed to the $\pi -\pi *$ and $n-\pi *$ transitions of delocalized excitons in the polymer chain, while the visible absorption bands are attributed to the intramolecular charge transfer (ICT) between the electron-rich moiety and electron-deficient moiety of the main chain [58,59]. Hence, the absorption coefficient ($\alpha$) can be determined from the absorbance data by using Equation (9), and the optical constants can be estimated using Equations (10)–(18) [60].

$$\alpha =\frac{2.303A}{t}$$

(9)

$$n=\frac{-2(R+1)\pm \sqrt{4k^2{R}^2+16R-4k^2}}{2(R-1)}$$

(10)

$$k=\frac{\alpha \lambda }{4\pi }$$

(11)

where t is the thickness of the photoactive layer, A is the absorbance, and R is the reflectance and can be calculated by using the following equation:$R=1-T-A$, where T is transmittance and is estimated from $T={10}^{-A}$. Therefore, prior to device fabrication, optical constants, such as refractive index (n) and extinction coefficient (k), as well as their derivative parameters, such as the di-electric constant ($\epsilon )$ and optical conductivity (${\sigma }_r$), must be addressed.

By analyzing the refractive index of the photoactive layer, it is possible to determine how the electromagnetic wave propagates throughout the photoactive materials and how the speed of light within the material varies relative to the vacuum. In addition, it is a complex variable, with the imaginary component representing the amount of energy lost due to the medium; this quantity is known as the extinction coefficient. In addition, the optical di-electric constant is a frequency-dependent metric that represents the electron response to the incident photon in the material, whereas the dissipation factor relates to the rate of absorption. In the meantime, the di-electric constant is a complex function, the real part of which is attributed to polarization under the influence of an electromagnetic field, while the imaginary component represents optical loss and can be characterized by the following equations:

$$\epsilon ={\epsilon }_1+i{\epsilon }_2$$

(12)

$${\epsilon }_1=n^2-k^2$$

(13)

$${\epsilon }_2=2nk$$

(14)

$$\tan \delta =\frac{{\epsilon }_2}{{\epsilon }_1}$$

(15)

where ${\epsilon }_1$ represents the real portion of the di-electric constant and ${\epsilon }_2$ represents the imaginary portion. The fundamental parameter that describes the electron’s response to photon absorption is optical conductivity. Since optical conductivity is obtained from the optical di-electric constant, it is a complex variable for which two components are defined by the following equations:

$$\sigma *={\sigma }_r+i{\sigma }_i$$

(16)

$${\sigma }_r=\omega {\epsilon }_2{\epsilon }_o$$

(17)

$${\sigma }_i=\omega {\epsilon }_1{\epsilon }_o$$

(18)

where $\omega$ is the angular frequency, ${\sigma }_r$ and ${\sigma }_i$ are the real and imaginary parts of the complex optical conductivity, respectively, and ${\epsilon }_o$ is the permittivity of free space.

5. Materials for Semi-Transparent Organic Photoactive Layer

Since the discovery of $\pi$-conjugated polymers in the early 19th century, the focus of renewable energy research has switched to the design and application of $\pi$-conjugated polymers. In comparison to their inorganic semi-conductor counterparts, the $\pi$-conjugated polymers possess a variety of advantages, such as solution processability, flexibility, and electrical and optical tunability via structural modifications. Alternatively, the side chains of the polymeric backbones improve their properties, which permits them to assemble nano-structured layers easily; additionally, due to the rich nature of the $\pi$ bonds of the conjugated polymer backbones, acceptor and donor units can be produced. Hence, the energy gap and electronic energy level of these materials demonstrate tunability and intramolecular charge transfer. To some extent, conjugated polymers have found several uses in different types of photovoltaic applications, including donor and acceptor materials, as well as hole and electron transport layers [61].

5.1. Conjugated Polymers

In organic solar cells, several architectures evolved from bilayer heterojunctions to the bulk heterojunction structures of donor and acceptor organic semi-conductor materials. Because of the low di-electric constant and low diffusion length of excitons, the research trends of the bulk heterojunction (BHJ) structure have become dominant and exhibit considerable efficiency. In this class of photovoltaic devices, the donor and acceptor materials are mixed in the solution phase and form a single junction binary heterojunction structure. Thus, the interface layer increases, and the thickness of the active layer can be reduced to around 100 nm, which is favorable for exciton diffusions and dissociations.

The conventional organic materials used were Poly(3-hexylthiophene-2,5-diyl) (P3HT), and, in particular, the fullerene [6,6]-Phenyl-C61-butyric acid methyl ester (PCBM) as the donor and acceptor materials, respectively, which showed limited efficiency and stability. Despite fullerenes’ frequent use due to high electron mobility, isotropic charge transport, good electron affinity, and good compatibility, whereas the increases in PCE were limited because of weak absorption in the near-infrared region NIR, there was a limitation on the energy level tunability, as well as the migration and aggregation of fullerene nanoparticles in the solid phase. Therefore, numerous non-fullerene acceptors were synthesized along with the synthesis of a variety of donor materials for enhancing device performance [36,62].

Furthermore, in semi-transparent organic solar cells (ST-OSCs), a trade-off is present between transparency and photon harvesting in the visible region. Therefore, the thickness and the type of materials in the active layer show a considerable effect on device performance. As such, the synthesis of new polymers/small molecules or the addition of the third component to the photoactive layer are alternative approaches toward enhancing both transparency and device performance. Semi-transparent ternary systems are generally composed of two donors and an acceptor or a donor and two acceptors, so the materials used are all polymers or polymers with small molecules or dyes. Based on the solar irradiance spectrum (Figure 5), the UV and NIR regions account for about 2% and 51% of solar power, respectively, whereas the visible region accounts for about 47%. Therefore, in the design of an active layer with high performance, there should be strong photon absorption in the UV, NIR, and IR regions and high visibility in the visible region. For that reason, the energy gap of the raw materials plays an important role in photon harvesting and charge collection. So the conventional classification for organic solar cells is based on fullerene and non-fullerene solar cells [38].

Poly(3-hexylthiophene) (P3HT) is the conventional donor material that was used in photovoltaic devices because, in the solid phase, it has a strong tendency to self-assembling into crystal phases. However, P3HT cannot be a prominent candidate for the ST-OSCs due to strong absorption in the visible region and poor hole mobility, which limits the thickness of the absorbing layer [63]. For example, Çetinkaya et al. fabricated a binary semi-transparent solar cell with the structure of (FTO/ZnO/P3HT:PCBM/MoO3/Ag/MoO3), in which they observed that the average visibility (AVT) is about 37%, whereas the efficiency of the device was around 1.77% [64]. In addition, Bliznyuk and co-workers introduced an organic dye into a P3HT:PCBM binary system and extended the absorption spectrum to near the IR region and transparent in the visible region; however, the efficiency of the ternary systems was in the range of (0.04–0.4%) for different compositions, and the efficiency was decreased compared to that of the standard binary system due to the reduction in the thickness of the active layer and traps that where introduced upon inserting the low energy gap materials [65].

Nonetheless, Xie et al. fabricated (ITO/PEDOT:PSS/PTB7-Th:PBT1-S:PC71BM/MoO3/Ag) fullerene-based ST-TOSCs, and the conventional donor was replaced by PTB7-Th by inserting a high band gap ($E_g=2.1\mathrm{e}\mathrm{V}$) PTB1-S polymer as a second donor with different weight ratios. Therefore, the optimized efficiency (10.3%) of the opaque ternary device was observed at 10 wt% of the third component, while the trade-off between AVT and PCE for the semi-transparent device was observed. The optimized AVT for the ternary system was achieved by reducing the thickness of the top electrode, but the PCE was reduced to its minimum value. For example, the device with an Ag thickness of 5 nm experienced 44.8% in terms of average visibility and 5% device efficiency [66]. Very recently, Liu and co-workers designed a new low band gap non-fullerene acceptor for ST-TOSCs, and the enhanced device performance for the active layer consisted of PM6:SN:Y6 (D:A1:A2). For this device, a PCE of 14.0% at an AVT of 20.2% was reported [67].

Additionally, the nanomorphology of the active layer is another key parameter that determines device efficiency. In the BHJ structure, the low mixing entropy and the entanglement of the polymer chain directly affect the domain size and phase separation. Hence, both molecular stacking and its orientation, followed by the crystallization and aggregation of the materials, define the charge transfer efficiency [68]. For instance, Heng and co-researchers prepared a ternary system consisting of PTFBDT-BZS:ITIC with fullerene as the third component. This study shows that by increasing the weight ratio of $\mathrm{P}{\mathrm{C}}_{71}\mathrm{B}\mathrm{M}$ up to 40%, the domain size of ITIC was reduced, and phase separation was improved. Furthermore, this study suggests that with a larger amount of the third component inhibited, there is large aggregation, and hence, the domain size was reduced [69]. In addition, Liu and co-researchers investigated the effect of domain size on device performance. In their study, the binary system of PBDB-T/N2200 was treated thermodynamically in a solution state, thereby obtaining a film with a proper domain size and a preserved degree of crystallinity [70].

5.2. Small Molecule Dyes for Semi-Transparent Solar Cells

The unique characteristics of small molecule dyes, such as electronic tunability, flexibility, cost-effectiveness, and lightweight qualities, have attracted the attention of both industry and academia. In addition, organic dyes exhibit a greater absorption coefficient and stronger exciton production than inorganic materials. Synthetic dyes are extensively employed in a variety of applications, including photovoltaic ones. For instance, several synthesized dyes, such as cyanine dye, squaraine dye, and azo dye (and its derivatives), have been employed in different structures of polymer photovoltaic devices, including semi-transparent and opaque structures [71,72,73,74,75]. Makha et al. fabricated inverted semi-transparent ternary OSCs with an AVT of 51% and a PCE of 3% by incorporating a small molecule of cyanine dye (Cy7-T) into a host binary system of PBDTTT-C:PC₇₀BM [74]. However, the expensive cost, supply limitation, heavy metal toxicity, and difficult manufacturing of these synthetic dyes eventually restrict their widespread application.

Alternatively, appealing natural dyes, such as carotene, chlorophyll, anthocyanin, and betalain dyes with their derivatives, can overcome these limitations and represent prominent candidates as one of the components of the active medium materials for light harvesting in ternary ST-OSCs. For instance, Vohra and co-workers employed a natural dye as an electron donor named β-carotene (bCar) and blended it with a fullerene acceptor to form the BHJ structure. As a result, the inverted structure of the organic solar cell experienced a fill factor of around 35%, and its efficiency was about 0.58% [76]. Moreover, Duan et al. studied aggregated chlorophyll as a p-type organic semi-conductor in both bilayer and BHJ structures and used it with the fullerene acceptor. The device performance showed power conversion efficiencies of around 5% and 3.5% for both BL and BHJ structures, respectively [77]. Furthermore, these natural dyes were used in all organic-based devices, such as perovskite solar cells, dye-sensitized solar cells (DSSCs), and organic light emitting diodes (OLEDs), but no reports focused on using these dyes in semi-transparent ternary organic solar cells [78,79,80].

5.3. Fullerene and Non-Fullerene Acceptors

Because of its noble electron transport capabilities, fullerene acceptors and their derivatives were employed in the BHJ structure, which is now the standard structural design for the manufacturing of organic solar cells. These fullerene-based materials, on the other hand, exhibit low optical absorption and limited energy level tunability. As the donor component is responsible for optical absorption, the tiny absorption window that the respective device possesses hinders the device’s efficiency.

Therefore, several approaches have been made to broaden the absorption window, which leads to enhanced photocurrent density, and, consequently, device efficiency as well as photochemical stability. Alternatively, designing a new n-type organic semi-conductor, named a non-fullerene acceptor, draws extensive attention. This is because the non-fullerene acceptor can be synthesized in a facile process, and this also preserves good electron mobility when compared to its fullerene counterparts and shows outstanding photochemical stability. Interestingly, the energy gap of these organic semi-conductors can be tuned, and therefore, the absorption window can be modified. Furthermore, several families of non-fullerene acceptors were synthesized and employed in photovoltaic applications, showing promising outcomes [27,43,81,82].

Recently, non-fullerene acceptors have been widely employed in both opaque and semi-transparent ternary BHJ organic solar cells, and the respective devices exhibited high performance. For instance, Wang et al. fabricated both ternary opaque and semi-transparent devices, which consisted of PM6:Y6:IT-M, and due to balanced charge transport and the smooth morphology of the ternary system, the photocurrent density was improved. The PCE for the opaque ternary device was around 15%, and the PCE and AVT for the semi-transparent ternary device were around 11% and 22%, respectively [83]. Additionally, Liu and co-workers employed a non-fullerene acceptor, IEICO-4,F as a host acceptor blended with PTB7-Th as a donor. By combining 20 wt% of NCBDT-4Cl, the ternary system was formed, and the optimum PCE and AVT of the ternary ST-PSCs were recorded at 10.31% and 20.6%, respectively [84]. Table 1 shows the recent developments in the materials, including the fullerene and non-fullerene acceptor and the third components for the semi-transparent ternary system.

In addition, the delayed aggregation at high temperature of the non-fullerene acceptors, when compared to that of the fullerene materials, significantly adds to an improvement in stability. Lee and Jung reported a comparison of the fullerene and non-fullerene acceptors employed in a binary BHJ structure [85]. Figure 6 shows that the bay-linked PBI derivative (di-PBI) has interestingly enhanced device stability, and 30% of the initial efficiency decreased after 3 h of operation at 100 °C. Nevertheless, the use of fullerene acceptor has led to a sharp drop in efficiency [85].

Table 1. Summary of material progress used in ST-OSCs.

Reference	Donor Materials	Acceptor Materials	Third Components	Device Efficiency
				PCE (%)	AVT (%)
[66]	PTB7-Th	PC71BM	PBT1-S	9.2	20
[86]	PM6	Y6	IHIC (20%)	12.18	27.07
[67]	PM6	Y6	SN	14	20.2
[87]	PBDB-TF	Y6	BDC-4F-C8	13.19	24.56
[88]	PM6	Y6	DIBC	14	21.6
[89]	PTB7-Th	BDTThIT-4F	IEICO-4F	9.4	24.6
[90]	PM2	Y6-BO	-	7.9	-
[84]	PTB7-Th	IEICO-4F	NCBDT-4Cl	10.31	20.6
[44]	PBDB-TF	Y6	DTNIF	13.49	22.58
[91]	D18-Cl	Y6	Y6-1O	13.02	20.2
[92]	PBDB-T	Y1	PTAA	12.1	20.1
[93]	J71	IHIC	PTB7-Th	9.37	21.4
[93]	PBDB-T	IHIC	PTB7-Th	8.76	20.6

Table 1. Summary of material progress used in ST-OSCs.

Reference	Donor Materials	Acceptor Materials	Third Components	Device Efficiency
				PCE (%)	AVT (%)
[66]	PTB7-Th	PC71BM	PBT1-S	9.2	20
[86]	PM6	Y6	IHIC (20%)	12.18	27.07
[67]	PM6	Y6	SN	14	20.2
[87]	PBDB-TF	Y6	BDC-4F-C8	13.19	24.56
[88]	PM6	Y6	DIBC	14	21.6
[89]	PTB7-Th	BDTThIT-4F	IEICO-4F	9.4	24.6
[90]	PM2	Y6-BO	-	7.9	-
[84]	PTB7-Th	IEICO-4F	NCBDT-4Cl	10.31	20.6
[44]	PBDB-TF	Y6	DTNIF	13.49	22.58
[91]	D18-Cl	Y6	Y6-1O	13.02	20.2
[92]	PBDB-T	Y1	PTAA	12.1	20.1
[93]	J71	IHIC	PTB7-Th	9.37	21.4
[93]	PBDB-T	IHIC	PTB7-Th	8.76	20.6

6. Approaches to Improve Semi-Transparent OSCs

6.1. Active Layer Strategies

As previously stated, the primary distinction between opaque and semi-transparent organic solar cells is their visible region transparency. Hence, to achieve this condition, several approaches have been proposed by researchers, as shown in Figure 7 and Table 1. For instance, Hu et al. and Cheng et al. utilized large band gap polymers to obtain semi-transparent solar cells. The broad band gap polymer D18-Cl (HOMO: −5.48 eV/LUMO: −2.75 eV), narrow band gap material Y6-1O (HOMO: −5.71 eV/LUMO: −3.84 eV), and the even narrower band gap material Y6 (HOMO: −5.65 eV/LUMO: −4.10 eV) were used by Hu and co-workers as an active layer in semi-transparent OSCs. Two approaches were utilized to convert the opaque cells into semi-transparent cells: first, adjusting the weight ratio of the donor to the acceptor, and second, reducing the thickness of the top electrode [91]. Moreover, Cheng and co-workers employed a large band gap polymer (poly[bis(4-phenyl)(2,4,6-trimethylphenyl) amine (PTAA)) as a partial substitution for the donor material and as a hole transport layer in the ST-OPVs. Thus, the inclusion of PTAA into a binary system led to enhancements in the visibility of the device in the visible region and improvements in film morphology (Figure 8) [92].

In addition, reducing the thickness of the active layer is another approach for converting an opaque structure into a semi-transparent structure. For instance, Lu et al. utilized a small molecule from DIBC as a third component in a binary PM6:Y6 host system. The stability and thickness tolerance of the active layer was demonstrated, as shown in Figure 9. The ST-OSC was exposed to continuous irradiation and thermal annealing. The device with 10% DIBC kept around 80% of its original efficiency after annealing for 300 min at 60 °C. Similarly, after 300 min of illumination, the device lost its efficiency by 33%. It was found that due to hydrogen intermolecular bonds, the aggregation of photo-sensitive materials can be enhanced. Therefore, extra channels for charge transport are facilitated because of the presence of more contact interfaces. Consequently, the incorporation of a small portion of the hydrogen bond via inserting DIBC hinders both bimolecular recombination and trap-assisted recombination [88].

On the other hand, the weight ratio adjustment of the components of the photoactive layer was used for enhancing semi-transparent solar cells. Hu and co-workers employed a narrow band gap material as a host binary material and investigated the effect of the weight ratio of the third component. In addition, IEICO-4f, as an ultra-narrow gap, forms an alloy with the host acceptor material due to similar electronic energy (HOMO and LUMO) levels [89], and the optimum weight ratio was found to be 50%. Moreover, Wang et al. balanced the trade-off between PCE and AVT upon inserting (10 wt%) of the BTTPC acceptor into the host binary blend. Because of the enhancement of the rate of hole transport, the charge recombination was reduced in the ternary blend. Additionally, further improvement was observed for AVT without affecting the PCE due to the application of a photonic reflector, as shown in Figure 10 [94].

Non-fullerene acceptors with a narrow band gap were broadly used for semi-transparent applications. For instance, Yin and co-workers utilized a non-fullerene acceptor (DTNIF) to enhance $V_{oc}$ and to balance the charge transport when 10% of the weight ratio was added to the host materials. The improvements were achieved because DTNIF has a high LUMO energy level when compared to that of the acceptor Y6, and there is an increase in the crystallinity phase in the ternary blend with decreasing π-π stacking distance [44].

Moreover, blending the fullerene acceptor with the non-fullerene acceptor has led to establishing long-term stability and enhancing the film morphology. For example, Sano et al. demonstrated that by adding a non-fullerene acceptor (ITIC) into the binary active layer (PCDTBT:PC71BM), the devices exhibited long-term operational stability. Figure 11 shows the long-term stability of opaque and semi-transparent OSCs with and without the presence of non-fullerene acceptors. Initially, the ST-OSC with incorporated ITIC experienced a drastic change in its efficiency due to the oxidation effect. However, the opaque OSC could keep around 90% of its efficiency after 45 days. This is because of the effect of the non-fullerene acceptor, which acted upon slowing down the kinetic process of degradation. Furthermore, PC71BM with ITIC yielded uniform absorption spectra [40].

Remarkably, the absorption spectra for the ternary system were used as an experimental tool to find the optimal weight ratio at which the device shows a maximum photocurrent. In addition, the third component (BDC-4F-C8) was used as a crystallization agent to enhance the crystallinity of the binary system [87]. Meanwhile, small molecule dyes and wide band gap polymers might have the potential for semi-transparent photovoltaic cells due to the ability of these materials to extend and continue the absorption spectrum in both directions (blue shift or red shift). For example, Tang et al. proposed a wavelength-selective semi-transparent OSC bulk heterojunction structure by blending a wide band gap small molecule TAPC $\left(E_g=3.54\mathrm{e}\mathrm{V}\right)$ with a non-fullerene Y6 and the optimized parameters for ST-OSCs attained an overall PCE of 3.01%, with a high AVT of 47.99% [95].

6.2. Transparent Top Electrodes

In order to achieve highly efficient semi-transparent OSCs, the prerequisites are high transmittance in the visible region, reflectance in both the UV and IR regions, and high conductivity for both electrodes, whereas the compromise between the high transmittance and high conductivity of the top electrode is prevalent. Thus, device engineering and the selection of electrode materials are very crucial for device operation. To date, several types of top electrodes have been employed in photovoltaic devices, as shown in Figure 12. However, each type of top electrode shows strengths and weaknesses.

As shown in Section 5, the most recent studies utilize thin metal layers as a top electrode. Nevertheless, a trade-off is observed between AVT and PCE, which is employed by decreasing the thickness of the top electrode. Consequently, visibility increases while efficiency decreases. Alternatively, metal nanowires, such as Ag nanowires, have the potential to overcome the compromise between the conductivity and transparency of top electrodes. Flexibility and solution processability, along with interesting optoelectronic properties, are measured as the strength of Ag nanowires, whereas poor adhesion, roughness, and device stability are considered to be drawbacks and hinder the improvement of device performance. Therefore, several studies have contributed toward improving Ag nanowires [74,96,97,98], and recently, Sannicolo et al. used Ag nanowires as a back electrode stabilized with graphene oxide sandwich layers. Concludingly, the graphene oxide layer serves to encapsulate the hole transport layer and optimize its adhesion to the top electrode, as well as protect the top electrode from the acidity of the PEDOT:PSS interface layer [99].

Another type of transparent top electrode is conductive oxides, such as indium tin oxide (ITO) and Al-doped zinc oxide (AZO), that possess excellent optoelectronic properties. Generally, ITO is used as the bottom electrode, and due to its high conductivity and transparency, it can also be used as a top electrode. However, due to the deposition process, which can be carried out by using magnetron sputtering, the photo absorber layer might be damaged, which results in poor device performance [100]. Therefore, Huang et al. proposed a laminated technique for assembling ITO as a top electrode, where the modified PEDOT:PSS was used as an adhesive layer for assembling the device [101].

For flexible semi-transparent organic solar cells, carbon nanotubes, and graphene are favorable for transparent top electrodes due to high transparency in the visible region, chemical stability, and mechanical flexibility. However, a survey from the literature showed that the device efficiency is around 1 to 4%, which is due to the low conductivity of the top electrodes but exhibits good stability and a long lifetime [35,36]. Furthermore, the promising candidate in terms of transparent top electrodes for all solution processability of semi-transparent OSCs is poly (3,4-ethylenedioythiophene)-poly-(styrenesulfonate)(PEDOT:PSS;PH1000), which exhibits high transparency in the visible range, low-temperature solution processability, and tunable conductivity.

6.3. Optical Engineering Strategies

Although the trade-off between transparency and efficiency is still the key issue hindering the commercialization of semi-transparent organic solar cells, different techniques were involved in achieving balance and perceiving a natural color for the devices. Several designs were employed as a wavelength controller of the reflection/transmission of the incoming photons, such as a distributed Bragg reflector (DBR), an organic-based reflector (cholesteric liquid crystals, CLC), 1D photonic crystals (PCs), di-electric mirrors (DMs), and microcavity (MC) structures. It has been shown that these designs can enhance the efficiency of the device by reabsorbing the transmitted photons as well as controlling the color of the active layers [102,103,104,105]. For instance, Li et al. fabricated a semi-transparent solar cell with a distributed Bragg reflector (DBR) as an infrared photonic reflector (Figure 13). The performance of the cell showed that the device with DBR exhibited a high visibility of around 29.5% and an efficiency of 7.3% [106].

In addition, due to the potential applications of ST-OSCs, particularly for power windows, the light intensity must be maintained at the required level when passing through the ST-OSC. Therefore, optical engineering on the top electrode plays a vital role in keeping the light components [93]. Xu and co-researchers prepared a binary BHJ structure of ST-OSCs with the device configuration of ITO/AuNBPs-PEDOT:PSS (40 nm)/PM6:BTP-eC9 (120 nm)/BCP (8.0 nm)/Au (1.0 nm)/Ag (10 nm). As depicted in Figure 14a, the optical coupling layer (OCL) formed of a di-electric bi-layer of LiF/MoO3 was used in their work despite the architectural method of introducing a gold nano-bi-pyramid (AuNBPs)-based hole-transporting layer to boost near-infrared absorption. As demonstrated in Figure 14b, the OCL enhances the optical path due to the scattering effect while functioning as a barrier against moisture and oxygen to affect the active layer and improve its stability. In addition, owing to OCL, the color co-ordinates changed toward equal energy white, as seen in the chromaticity diagram (Figure 14c), and Figure 14d depicts the ST-OSC photograph obtained with and without OCL [107].

7. Potential Applications and Future Perspectives

The insufficiency of energy sources relative to the amount of energy used today is the most difficult problem the humanities have ever encountered. Renewable energy sources are currently a topic of interest, and among these sources, solar energy can meet the energy demand. In addition, semi-transparent solar cells are highly desirable for a range of applications. For instance, by integrating ST-OSCs into greenhouses, they can provide light at a specific wavelength for photosynthesis in addition to providing energy [30]. Another application is the power-generating window, which may be integrated into buildings, vehicles, and wearable electronics to give clean energy and natural color while absorbing thermal energy [19]. An additional application is the semi-transparent floating photovoltaic solar cell, which has advantages over opaque solar cells, such as the reduction in water evaporation caused by a thermal barrier by lowering the surface temperature (Figure 15) [108]. However, due to the cost-effectiveness and complexity of inorganic solar cell production procedures, the trend in research is toward organic solar cells. In addition, major attempts have been made in recent decades to commercialize semi-transparent organic solar cells; thus, a variety of research has been conducted in this regard to improve both efficiency and stability.

8. Conclusions

One of the endeavors to increase power conversion efficiency (PCE) while conserving flexibility, easy fabrication, and low costs is the ternary system for a bulk heterojunction active layer. The ternary system improves light harvesting by broadening the spectral response of the active layer. A ternary system comprises a combination of three components, including a donor, an acceptor, and a third component. Therefore, selecting the third component is a remarkable consideration for device performance. The materials used in the BHJ structure were categorized based on their structures, including polymers, small molecules, nanostructures, and dyes. Hence, selecting the third component in a ternary system induces more convolution in the charge transport mechanism, whereas it can enhance the absorption profile, exciton dissociation, charge transport, and film morphology. Moreover, the ability of organic materials to tune their band gap energy revealed a new field of practices entitled semi-transparent organic photovoltaics (ST-OPVs). Additionally, in semi-transparent organic solar cells (ST-OSCs), a trade-off is presented between transparency and photon harvesting in the visible region; therefore, the thickness and the type of materials in the active layer show a considerable effect on device performance. Consequently, the synthesis of new polymers/small molecules or the addition of the third component into the photoactive layer are alternative approaches toward enhancing both transparency and device performance. Despite this progress, more research on photoactive materials is needed to fully understand the working mechanisms of ternary systems as well as device stability.