The organic-inorganic halide perovskite solar cells (PSCs) have attracted a great deal of attention of solar cell research community due to an incredible device efficiency improvement from 3.8% to 22.1% since 2009 [1
]. The perovskite already gained much attention as a potential replacement of the silicon photovoltaic (PV) devices, which is still occupied the most dominant position in the current PV market, with record efficiency of about 26% [3
]. This small gap of solar cell efficiency attracted recent attention especially from the researchers with experience in dye-sensitized solar cells (DSSCs) or organic solar cells because some materials can be used in both PSCs and organic solar cells. The structure of PSCs also origins from the device structure of DSSCs [1
]. The perovskite materials have been demonstrated with largely tunable band gap (e.g., CH3
has a band gap from 1.5 eV to 2.3 eV) [4
] and great light absorption coefficient (higher than 104
], which is similar to other thin film solar cell materials such as CdTe [7
] and copper zinc tin sulfide (CZTS) [8
]. Its low-cost and convenient fabrication techniques also serve as the possible advantages over silicon-based devices that require complicated and costly high-vacuum deposition methods. Reports of successful cell fabrication on flexible substrates even indicated a greater possibility to the large-scale roll-to-roll manufacturing of PSCs that can be used in the industries [9
The initial meaning of “perovskite” was about the crystal structure of calcium titanate, which was discovered in 1839 by the German mineralogist Gustav Rose and was named by the Russian mineralogist Lev Perovski. Since then, the term “perovskite” was referred to all compounds with the same crystal structure as calcium titanate. The perovskite light absorption layer has a general formula of ABX3, where A is an organic cation (e.g., methyl-ammonium CH3NH3+), B is a metal cation (e.g., Pb2+) and X stands for the halide anion (e.g., I−).
The first record of perovskite-based solar cell efficiency, however, was reported by Miyasaka et al. [1
] only less than one decade ago. They reported an efficiency of 3.8% based on a DSSC structure. Due to the application of liquid electrolyte in the hole-transporting material (HTM), the stability of solar cell was very weak and did not attract much attention. Similar trial was done by Park et al. [12
] with the increased efficiency of 6.5% but stability was still the main problem because of the instability of HTM layer due to the liquid medium.
The application of solid-state HTM (2,2′,7,7′-tetrakis(N
-di-pmethoxyphenylamine) -9,9′-spirobifluorene, i.e., Spiro-OMeTAD), rather than liquid HTM, onto the highly-crystallized perovskite layer triggered the efficiency boosting during the past several years. Lee et al. [13
] reported a breakthrough device efficiency of 10.9% in 2012 with the open-circuit voltage higher than 1.1 V. Wang et al. [14
] introduced graphene into PSCs and acquired an efficiency of 15.6% in 2013 and the application of another perovskite material, formamidinium iodide (HC(NH2
) together with poly-triarylamine (PTAA) as a new HTM brought a remarkable 20.1% efficiency in 2015 [15
]. The current record efficiency of PSCs was 22.1%, created in 2016 by Seong Sik Shin et al. [16
]. They also accomplished a long-term and stable efficiency of 21.2% in another work [17
]. The perovskite-inserted tandem cell also achieved a promising efficiency of 26.7% by combining with Si cells [18
]. During this progress, various HTM and vacuum/non-vacuum fabrication methods have been developed, which would be discussed later in this review. Figure 1
compared the efficiency progress of PSCs with other 3rd generation photovoltaics up to date [19
]. The rapid improvement of the efficiency of PSCs make perovskite being expected to be comparable with the stable performance of c-Si solar cells whereas all other kinds of non-silicon solar cells suffered great barriers in further improvements. According to the theoretical calculation based on the well-known Shockley-Queisser limit, the perovskite devices, which have (CH3
), could achieve an efficiency around 25–27% [20
]. This result indicates that there is still opportunity for the improvement of PSCs.
Although laboratory scale PSCs exhibited a great progress, perovskite-based PVs still needs to overcome several barriers. In general, there are two major problems currently blocking the improvement pathway: device instability of device performance [21
] and hysteresis of J-V (current density-voltage) [23
]. At present, long-term efficiency measurements (>1000 h) is still not adequate for the commercialization of PSCs. The PSCs must pass a series of testing under harsh conditions and environments for similar duration (>1000 h) [24
]. Thus, it is very important to understand the degradation mechanism of both perovskite materials and other device components such as hole transport medium (HTM) and electron transport medium (ETM). The J-V hysteresis was discovered during cell testing when voltage sweeping routine changed. This phenomenon brings problems for standardizing the measurement protocol of PSCs. In addition, the toxicity from lead could be another problem during the manufacturing, using and recycling of perovskite [25
]. Currently several trials on applying non-toxic alternative metal ions have been reported [26
] but their device efficiency is still not promising. Detail information could be found later in this review.
Future research of PSCs, except efforts on improving the stability and reducing J-V hysteresis of PSCs, could also be focus on the large-area fabrication of PSCs (even small module area) and efforts on at least partial replacement of lead with other non-toxic metal ions inside the perovskite. Bi-facial illumination could also be considered for PSCs due to its structural advantages. Detail information could be found in the last part of this review.
It has been clear that the perovskite could be the next candidate to replace Si due to its outstanding structural, electrical and optical properties. This review, therefore, would start with the discussion from micro-scale observations on the crystal and electrical structures of perovskite materials. The next part is the discussions on device-level investigations: the evolution of device structure, the fabrication methods and their progresses and the exploration of each device component. We would then focus on the research efforts of device stability and toxicity of PSCs and finally show our suggestions for further directions of the perovskite research.
The toxicity of perovskite comes from the widely-used lead inside MAPbI3
and environmental concerns would be appeared especially on the issue of large-scale fabrication waste treatment. Although calculations already showed the possible contamination from perovskite would be relatively insignificant compared with other lead pollutions [178
] and the production of PSC could be able to use waste lead from daily waste [179
], studies on lead-free PSCs cannot be neglected. Tin was the first well-studied replacement metal cation since Sn and Pb are both carbon periodic elements, thus, MASnI3
is believed to be able to maintain the same crystal structure as MAPbI3
. The fact, as shown by Noel et al. [180
], is that Sn2+
could be easily oxidized to Sn4+
, leading a weak device performance. Other trials of introducing organic/inorganic additives to retard tin oxidation had also been reported [181
] but their device PCE was still not promising.
Due to this chemical instability of pure tin-based perovskite materials, the hybrid Sn-Pb metal cations in perovskite could be more realistic and the more advanced PCE also demonstrated this idea: Zhu et al. [183
] reported a remarkable PCE of 15.2% with a light absorber of MASn0.25
and a suitable control of DMSO additive and a PCBM:C60
electron transport layer. Another study also indicated MASn1-x
could have an electronic structure closer to MASnI3
even with few Sn replacement [184
]. All those results indicated that from the view of reducing process toxicity, tin is not a perfect candidate to totally replace lead due to its chemical instability.
Another intensive-studied candidate is the neighbor of lead: bismuth. Bi could form a stable (MA)3
(MABI) perovskite material. Its crystal structure was shown in Figure 21
]. Similar as Sn-doped MAPbI3
, MABI also showed better stability under ambient air for 1000 h [186
]. The first reported MABI-based perovskite only reached a low efficiency of 0.12% with a relatively low Voc
of 0.68 V and an extremely low Jsc
of 0.52 mA/cm2
]. At present, the Bi-based perovskite could only able to reach an efficiency of 0.42% due to low Jsc
]. A recent study investigated the absorption and recombination dynamics of excitations inside the MABI crystals: the I(5p)-Bi(6p), I(6p) excitation [189
] is localized in (BI3
units, resulting little free carriers released at the MABI/TiO2
]. Therefore, they suggested considering bulk-heterojunction structure with nano-scale MABI crystals in order to possibly enhance the low Jsc
. Some other reports focused on Bi-based halide double perovskites such as Cs2
(X=Br, Cl) [191
] and (MA)2
]. But no effective devices had been reported and deep understanding of optoelectronic properties are still suggested. Thus, Bi-based PSC is still not promising at present, even compared with Sn.
Other types of lead-free perovskites such as CsGeI3
(X: Cl, Br, I) [194
] had also been reported but those materials are either not suitable for visible light absorption due to large band gap [195
], or only showed low efficiency of less than 1%. Most of the material characterizations are still in lack. Thus, non-toxic perovskite development and corresponding fabrication of PSCs still has a long way before replacing the position of lead.
6. Discussion and Future Research Efforts
The above research efforts indicated that the PSCs will have a greater potential for commercialization if the stability of cells can be improved. The high efficiency and low-cost manufacturability to harvest terawatt levels by solar energy are very attractive with this next generation solar cell technology. The cell degradability is identified as due to primarily the exposure of perovskite layer to water vapor and heating effects, which change the active phase of lead-based perovskites. There are numerous efforts to improve the stability of these solar cells by many groups worldwide. Development of perovskite layers using other metals has been tried with poor success to address the toxic issue and stabilizing the perovskite structures. Also, cell passivation has been investigated to stabilizing the cells by prevention of perovskite layer to the ambient. Another approach is to reduce the heating effect by utilizing IR absorbance layers or external components. Also, integration of few of these technologies may improve the stability of this solar cell technology from current stability records of around six months.
Also, it is important to address the harmfulness of lead-based compounds in PSCs. While development of other metal-based perovskite is also interested in the viewpoint of environmental protection, their effectiveness does not reach the efficiency of lead-based compounds. Furthermore, better recycling methodologies are important to prevent the transfer of lead compounds into the environments. Similar environmental issues have been addressed for CdTe solar cells and thus, it is possible to utilize already existing infrastructure for recycling and environmental protection issues. The environmental protection authority regulates the lead content of drinking water below 0.015 g/L. Authors believe that these areas can be further improved by research efforts.
It is also important to address that, due to the transparent nature of some of the HTM/ETM and the electrodes, the PSCs could be fabricated with a structure that could absorb sunlight from both directions. An investigation in 2016 already found out that, with the help of transparent solution-processed silver nanowires (AgNWs), an efficiency of more than 11% and 7.53% could be observed with front and back illumination, respectively [197
]. Together with passivation of PCSs using transparent insulators such as polymers, the cell performance as well as durability may be enhanced. Thus, investigation on the bi-facial PSCs could also be another research direction for the improvements of PSCs.
The PSC has experienced a significant improvement from 3.8% to 22.1% since 2012 and the perovskite-based tandem cell has already achieved 26.7%, creating a new record in history of PV technology. Numerous research efforts on both PSC efficiency improvements and deeper understanding about perovskite materials’ outstanding electrical and optical properties, such as largely-tunable band gaps for light absorption, high absorption coefficients, large carrier diffusion lengths, great carrier mobility, have been established during the past few years. The current PSCs already combined structural advantages of both DSSCs and thin film PV since the discovery of perovskite and become a new challenger for Si-based PV dominant market share, not only due to record 22.1% efficiency for small area but also comparable larger-area device efficiency. The vast discovery and successful application of organic/inorganic charge transport materials and blocking layers also assisted the formation and crystallization of perovskites and helped charge transfers at the interfaces. Many kinds of PSC fabrication approaches have also been developed and most of them could fall into four major categories: one-step; two-step; vapor-assisted solution method; and thermal vapor deposition with a top PCE of 22.1% (current record), 20.26%, 16.48% and 17.6%, respectively. Also, numerous hybrid perovskite fabrication process was also invented, which is uncommon for other types of PVs. According to current progress, it is reasonable that the next high-efficiency PSC may be still based on solution-based approaches (e.g., spin coating) with a mixed perovskite phase, as applied in the record 22.1% and the stable 21.2% devices.
The PSCs still have great barriers for further improvements. The biggest problem comes with the natural instability of perovskite materials, especially the most widely-used MAPbI3. The phase transition within the range of solar cell operation temperature brought problems on device usage. The instability with varying temperature and pressure leads extra concerns for device fabrication. The moisture, UV light and oxygen would also bring irreversible damage to the perovskite layers, which largely reduced the device stability and commercialization of PSC. The efforts such as elemental adjusting, device sealing and extra blocking layer inside the device had been tried to solve these problems but more stability tests under harsh environments are strongly suggested for PSCs to reach the required standard.
Other drawbacks such as J-V hysteresis and toxicity of lead made it difficult to further improve the performance of PSCs. While the mechanism of hysteresis was still inconclusive, the lead toxic has attracted many research efforts on considering the non-toxic replacement pf perovskite materials. Research work found out that all candidates, from the neighborhood Sn, Bi and new candidates as Cs, Ge, suffered a great loss of Jsc, which directly leads to the huge loss in PCE. A more complicated replacement profile might be the solution of lead-free PSCs. Although efforts claimed low hysteresis in some PSCs, deep theoretical understanding and standardized testing protocol is suggested for PSCs.
Perovskite, compared with other PV techniques (thin film, organic, dye-sensitized), could be the best alternative solar absorber. As efforts on better perovskite layer formation and longer device durability, even the lead-based PSCs could be able to share a certain part of PV market. Such trend could influence further research and development (R&D) efforts towards higher-stable and non-toxic devices.