Perovskite Solar Cell Structure and Layers

Over the past 10 years, perovskite solar cells (PSCs) have achieved record efficiencies of 26.1% single junction solar cells (as of 2023¹). These efficiencies continue to rise due to perovskite's inherently low defect densities, tuneable bandgaps (making them perfect for tandem solar cells in conjunction with silicon-PV) and impressive optical and electronic properties. Perovskite solar cells can be almost completely solution processed and are compatible with roll-to-roll processing methods.

Perovskite solar cells need several layers in order to absorb light, then separate and extract charge. In basic terms, a planar PSC needs an absorbing perovskite layer sandwiched in between a hole transport layer and an electron transport layer.

Constructing the highest performing PSCs requires the right perovskite materials selection, including choosing the right perovskite absorber layer matched with the best hole transport layer and electron transport layer. High efficiency devices also require carefully controlled processing of all layers.

Choosing Layers In A Perovskite Solar Cell

Choosing the best charge transport layers is extremely important when constructing an efficient perovskite solar cell. There are several factors to consider when making this decision, including:

• Processing limitations on substrates
• Band level alignments
• Device structure
• Wettability
• Chemical resistance of previous layers.

Good charge transport layers extract charge efficiently, while decreasing recombination losses at the transport layer/perovskite interface that may lower V_OC­².

Structurally, depending on whether the device has an inverted or regular architecture, the “bottom” transport layer must be smooth, uniform and have good wettability to ensure that a smooth perovskite layer can be deposited on top of it.

Furthermore, the “illuminated” layers — through which light will need to pass to get to the perovskite absorber — must have good transparency to allow light through to the perovskite absorbing layer.

Different Perovskite Layers

Included here are recipes for making several different perovskite films and metrics of the solar cell devices made using these materials. All devices described here have the device stack:

ITO-Coated Glass /SnO₂/perovskite material/Spiro-OMeTAD/Au

and these respective transport layers are deposited as described previously. All perovskite layers and hole transport layers were deposited in a glove box environment which maintained a level of <0.5% relative humidity and <0.5% O₂.

CsFAMAPb(I_xBr_1-x)₃

We here examine a high-performing triple cation perovskite (CsFAMAPb(I_xBr_1-x)₃) of 1.2M with a 4:1 DMF/DMSO solvent blend. Mixed halide (Iodine, I, and Bromine, Br) and mixed cation (methylammonium, MA, formamidinium, FA and Cesium, Cs) PSCs have consistently produced devices with relatively impressive stability and durability. TC devices are acheive reasonably high efficiencies ^25–27 and are very robust. Typically, these perovskites are processed in an inert environment to achieve the best performances, but nevertheless these perovskites are extremely versatile. In 2020, Bishop et al. demonstrated a fully sprayed PSC device using this TC ink, with small area efficiencies of 19.4%²⁸, and this ink has been shown to be compatible incredibly scalable techniques such as air-blading quenching²⁹. However, there are some issues with triple cation perovskites. For instance, methylammonium is extremely volatile, thermally unstable and will degrade in the presence of oxygen or moisture. Therefore, their use in PSCs can significantly lower the lifetime of the device. Nevertheless, when properly encapsulated, these PSCs have been shown to have great performance over several months.

Here, we have made devices using the aforementioned precursor, 50 µl of which was statically spin coated onto the SnO₂-coated substrate with an Ossila Spin Coater using the following spin-anneal cycle:

During perovskite deposition, the Ossila Glove Box was continually circulating N₂ compressed gas to avoid solvent build-up within the main chamber. Subsequently Spiro-OMeTAD and Au layers were deposited and devices were tested 2 days after perovskite and HTL deposition. When processing, all devices below <10% PCE or <0.5 V_OC were removed. These results are summarized below. As can be seen in Figure 2, the average PCE over 6 devices of 6 pixels was 16.4±1.8 %. The highest PCE achieved by a single pixel was 19.0%, stabilizing at 17.6% over 1 minute.

MAPbI₃

Methylammonium lead iodide (MAPbI₃) is one of the earliest formulations of perovskite used in PSCs. Here, 1.1M methylammonium iodide (MAI) was dissolved with 1.1M of PbI₂ in DMF. This is typically a high performing perovskite and can be processed at relatively low temperatures³⁰. Their primary use of methylammonium as an A-cation can lead to some intrinsic instabilities, so although they may have impressive performance initially, this may deteriorate over a matter of days, if not properly encapsulated.

Again, we fabricated some devices using this perovskite precursor, 50 µl of which is statically spin coated onto SnO₂-coated substrates using the following spin-anneal cycle:

These layers were deposited in our lab glove box. However, no circulation of N₂ was needed to get optimum device performance. After spin coating, Spiro-OMeTAD and Au layers were deposited, and devices were tested one day after perovskite and HTL deposition. When processing, all devices below <10% PCE or <0.5 V_OC were removed. The results are summarized below.

As can be seen in Figure 3, the average PCE over 3 devices of 6-pixels was 17.6±1.1 % with the champion PCE for a single pixel of 19.3%, stabilizing at 18.4% over 1 minute.

.

CsFAPbI₃

Due to the various instabilities of methylammonium-based perovskites, formamidinium base perovskites (using halide salts such as formamidinium iodide or formamidinium bromide) are an appealing alternative. FAPbI₃ has a smaller band gap than MAPbI₃, which means it can absorb a larger proportion of the sun’s spectrum. This should lead to higher photocurrents and in turn better device efficiencies³¹.

However, FAPbI₃ is phase-unstable and require high temperatures to convert fully to the black (α-) phase needed for perovskite solar cells, falling into a less absorbing, non-perovskite (β-) phase.

There has recently been a surge of formamidinium-based perovskites which show impressive device stability without the use of methylammonium cations to stabilize the α-perovskite phase. Currently (as of Nov 2021) the certified highest performance perovskite device is a doped FAPbI₃-based perovskite achieving 25.5% PCE⁸ and there have been several other extremely successful examples using FAPbI₃-based perovskites^7,32,33 including a slot die coated device with an active area of 20.77 cm² attaining PCEs of 16.6%³².

Here, we have fabricated some devices using this perovskite precursor, 50 µl of which is statically spin coated onto SnO₂-coated substrates using the following spin-anneal cycle:

During perovskite deposition, our laboratory glove box was continually circulating N₂ to avoid solvent build up within the glove box chamber. Subsequently Spiro-OMeTAD and Au layers were deposited, and devices were tested 1 day after perovskite and HTL deposition. In processing, we removed all devices <10% PCE or <0.5 V V_OC . Figure whatever shows the device metrics for all devices, along JV sweeps and fixed power tracking for the champion device.

The average device performance amongst seven devices was 14.0±2.0 % and the highest device sweep is 17.0% stabilizing at 16.0% after 1 minute.

I301

Here, we have fabricated some devices using our I301 triple cation perovskite ink. This uses dimethyl sulfoxide as the primary solvent making it a more stable perovskite solution³⁴, but otherwize has the same stoichiometry as the previously mention triple cation ink (CsFAMAPb(I_xBr_1-x)₃). 50 µl of which is statically spin coated onto SnO₂ coated substrates using the following spin-anneal cycle:

Here, all devices <10% PCE and <0.4 V V_oc have been removed. We have used two different antisolvents during quenching to determine if this will significantly affect device performance. As can be seen here in Figure 6, it does not. The average device performance of I301 devices made using ethyl acetate as an antisolvent is 17.1±0.7% with a best performance of 18.5%, while those made with anisole as the antisolvent have an average PCE of 17.3±1.3% with a champion device of 19.4%. This demonstrates that high performance perovskite solar cells can be produced with a range of antisolvents.

Perovskite Materials Mentioned

Methylammonium Iodide (MAI)

From $169

Formamidinium Iodide (FAI)

From $169

Formamidinium Bromide (FABr)

From $208

Perovskite Precursor Ink for Air Processing

From $260

Triple Cation Perovskite Precursor Ink (for Nitrogen Processing)

From $338

Comparing Perovskite Devices

Here is a table comparing these PSCs. Here we have outlined how to fabricate good perovskite layers within a glove box environment and demonstrated that good PSCs can be formed with these layers.

Electron Transport Layers

TiO₂

Devices using a mesoporous TiO₂ layer were popular in the early iterations of PSCs devices due to their use in dye-sensitized solar cells. TiO₂ has reasonably good charge extraction properties and is also very chemically resistant² so will not be affected by subsequent solvent processing.

Both compact and mesoporous TiO₂ layers are still used in high efficiency PSCs. They often used together where a compact TiO₂ layer ensures a full hole-blocking, electron-collecting layer while a mesoporous layer allows high levels of electron extraction.

However, these layers can require high temperature sintering (>400 °C) and are very time consuming to produce. Additionally, TiO₂ has been shown to be vulnerable to UV light causing degradation of the perovskite layer³, which can sabotage device performance.

TiO₂ can be doped with Ag or ruthenium to improve its performance as an ETL (by improving conduction, higher carrier mobility etc). For example, it has been shown that doping TiO₂ with Ag alters the band gap hereby improving carrier transport properties⁴. Graphene-doped TiO₂ led to a record PCE (in 2018) of 20.45% where it is determined that graphene provides a channel through the mesoporous TiO₂ which can assist with electron extraction⁵. TiO₂ has also been doped with rarer earth metals such as europium and samarium, and various other materials to improve band alignment and improve device stability².

SnO₂

SnO­₂ has a wider bandgap than TiO₂⁶ and its band alignment makes it perfect for charge extraction with methylammonium-based perovskites. Tin oxide also has high electron mobility in thin films and high chemical resistance, and can be processed using a range of low-temperature methods, such as chemical bath deposition⁷, spin coating, spray coating or slot die coating from nanoparticle suspensions⁶.

As of 2021, an SnO₂ ETL has been used to achieve the certified world record highest efficiency PSC of 25.5%⁸. Here, in combination with a chlorine-heavy perovskite precursor, a chlorine-doped SnO₂ interlayer is created between the SnO₂ ETL and perovskite layer, increasing electron extraction. However, SnO₂ layers do often require annealing at 150 °C, which could be a problem when working with some polymer substrates.

As this is a relatively low-stress ETL to fabricate, and can be made with an off-the-shelf nanoparticle solution, we have outlined a process for creating a compact SnO_2­ ETL below.

• Substrates should be placed in the UV-Ozone for at least 20 minutes before SnO₂ deposition to ensure good wettability
• 15 wt% SnO₂ nanoparticles are diluted in deionized water 4:1 and filtered immediately before use
• 50 μl of suspension is spin-coated statically at 3000 rpm for 30 s
• The layer is annealed at 150 °C for 30 minutes

ZnO

Zinc oxide is a low cost, low temperature alternative to TiO₂ which makes it more appealing for commercialization. It is also more conductive than TiO₂ and has a similar band gap.² Although its ideal anneal temperature is 400 °C, it can be deposited by a sol-gel technique, which lowers this considerably. However, ZnO is chemically unstable at high temperatures, especially in conjunction with perovskite material².

ZnO layers can also be doped with different materials to counter defects at the ZnO/perovskite interface. For example, Azmi et al. achieved an impressive PCE of 19.9% when using potassium doped ZnO in stark contrast to their 16.10% devices using only ZnO as an ETL.⁹

PCBM

Phenyl-C61-butyric acid methyl ester (PCBM) is a material that will form a stable ETL and can be easily solution processed on top of a perovskite layer, making it ideal for inverted perovskite solar cells.

For this reason, it was very popular in the early days of PSCs. However, several issues with this material ultimately limit device performance. Rough morphology of the perovskite layer can lead to an uneven layer of PCBM, creating a hot spot for non-radiative recombination. To reduce this effect, PCBM can be used in conjunction with other transport layers such as C₆₀ and LiF¹⁰.

Alternatively, PCBM can be doped to improve film morphology and to better match energy layers within the device. In 2019, Yang et al. doped pure PCBM with an n-type polymer F9TBT¹¹, which considerably improve coverage of the combined ETL, yielding an inverted MAPbI₃ devices of 20.6%.

C₆₀

C₆₀ is a fullerene material that can make a compact transport layer, with higher electron mobility than other ETLs used in inverted devices such as PCBM.¹¹ They can also be used in either regular or inverted devices. Due to their improved film morphology compared to PCBM, C₆₀-based devices exhibit low levels of hysteresis, which is an important property for inverted perovskite devices. Often, they are used in conjunction with a BCP hole-blocking layer, which allows for better charge selection at the electron transport interface.

Hole Transport Layers

Spiro-OMeTAD

Spiro-OMeTAD is one of the most used hole-transport layers for PSCs. It regularly produces high efficiency PSCs — and is used in the current world champion device⁸.

Spiro-OMeTAD can be purely solution processed and has no annealing step, meaning it can be processed easily at low temperatures. In fact, a Spiro-OMeTAD HTL has been shown to be compatible with large scale perovskite production, with Di Giacomo et al. creating >15 cm² devices with over 10% PCE¹² and Kim et al. showing that devices using Spiro-OMeTAD created with completely roll-to-roll processes have a champion PCE of 13.8%¹³.

However, Spiro-OMeTAD requires the addition of several dopants to achieve maximum conductivity, hole mobility and to ensure these dopants dissolve in chlorobenzene. Additionally, there is mounting evidence suggesting Spiro-OMeTAD may provide long term stability issues with PSC devices¹⁴. This deterioration has been linked with the aforementioned dopants, so much research is looking at replacing these dopants with Zn-TFSI, Mg-TFSI₂ and Ca-TFSI₂ and so far, promising improvements have been seen in device stability.

As Spiro-OMeTAD is the most used HTL, we have outlined our favourite recipe for creating the perfect "Spiro" layer in a regular perovskite solar cell architecture.

• 85 mg/ml of Sublimed Spiro-OMeTAD (<99.5% purity) is dissolved in chlorobenzene and left to dissolve fully for 2 hours.
• Dopants are added to this in stock solution in the following order:
• LiTFSI: 20 μl/ml from a 500 mg/ml stock (dissolved in ACN) added and vigorously dissolved for 1 minute
• tBP: 34 μl/ml from stock solvent and dissolved for 1 minute.
• Fk209 Co(III) TFSI Salt : 11 μl/ml from a 300 mg/ml stock (dissolved in ACN) added and dissolved for 1 minute.
• This layer should be deposited in an inert environment like a laboratory glove box and the solution should be filtered promptly before deposition with a 0.2 µm filter.
• 25 μl of this solution should be dynamically spin coated onto the perovskite layer at 4000 rpm for 30 s. These should then be left in a dark, dry atmosphere in air overnight before further processing to allow the Spiro-OMeTAD to oxidize fully.

PTAA

Poly[bis(4-phenyl)(2,4,6-trimethylphenyl)amine (PTAA) is another hole transport layer for PSCs. It has many of the same qualities as Spiro, in that it is a small organic molecule with which impressive perovskite solar cells have been achieved, but usually dopants are needed to achieve these high performances — the same dopants that are problematic in Spiro-OMeTAD devices i.e. LiTFSI. Therefore, more stable dopants (such as Lewis Acid Dopants) have been used to create PTAA-perovskite devices with PCE of >19% that show good device stability¹⁵.

It has also been demonstrated that dopant-free PTAA can been used in inverted devices along with PCBM as an ETL, and these devices also show good performance¹⁶.

PEDOT:PSS

PEDOT:PSS (Poly(3,4-ethylenedioxythiophene) polystyrene sulfonate) has often been used in inverted perovskite solar cell devices and was used in the first solid PSC device, achieving efficiencies of 3.8%¹⁷.

PEDOT:PSS has good wettability, good solubility in greener solvents and is quite transparent in the visible region.¹⁸ However, it suffers from low conductivity — as the PSS part of the molecule as an isolating agent in the layer- and the material suffers from mismatched energy levels with most perovskite material.

Additionally PEDOT:PSS will absorb ambient moisture, which can wreak havoc when in a perovskite stack. Doping PEDOT:PSS with various things can improve the materials charge carrier dynamics, work function and the overall film morphology, all of which lead to a better PSC. However, the challenge remains to improve all these properties at the same time.

As of 2019, the highest perovskite solar cell efficiency achieved using this HTL involves doping the PEDOT:PSS with CsI¹⁹. This lowers the work function of the HTL, reducing the energy barrier between PEDOT:PSS and the perovskite, hereby increasing hole extraction. Xhang et al. found that PEDOT:PSS doped with NiPcS₄ results in a slightly higher work function than PEDOT:PSS alone²⁰. However, this doping also significantly increases the lowest unoccupied molecular orbit (LUMO) which enables necessary electron blocking through the layer.

In turn, this leads to a higher J_SC­, and better and more stable devices.

Self-Assembled Monolayers

Self-assembled monolayers (SAMs) are the general name for 2D nanomaterials that are only one molecule thick. SAM molecules, such as MeO-2PACz will self-assemble into organized layers²¹. They consist of three parts – terminal molecule that interacts with the perovskite, an anchor molecule that attaches the SAM to the adjacent layer and a spacer molecule in between that helps enable the organization of the film.

Within perovskite solar cells, the work function of the hole and electron transport layers must be carefully optimized to enable extraction of one charge whilst blocking the movement of another. For this purpose, SAMS can be very useful as interlayers, as their electrical properties can be easily tailored²¹. They can also help improve the crystallization of the perovskite layer for a more efficient device.

Additionally, SAMs can be solution processed so can be immediately incorporated into most PSC devices. A wide range of SAMS have been used for different purposes within perovskite devices, we will outline a few here that assist with charge transport. For a more thorough review of how SAMs can improve other perovskite features, such as morphology, read on here²¹.

A SAM layer has been used to reduce the gathering of positive ions at perovskite/SnO₂ interfaces that would encourage non-radiative recombination²². SAMS have also been used as the main hole transport material in inverted PSC devices which have been shown to outperform PEDOT:PSS HTLs²³ and can certainly compete with PTAA HTLs in terms of stability²⁴.

Energy Level Diagrams in Perovskite Solar Cells

Band level alignment to extremely important in perovskite solar cells. Figure 1 shows the band gap for several hole and electron transport layers, as well as several perovskites. When the perovskite absorber captures a photon, this results in both an electron in the conduction band (CB) and a "hole" in the valence band (VB). It is important to extract these charges from their respective layers as quickly as possible to avoid recombination.

The electron that is occupying a space in the perovskite's conduction band must be able transfer easily into the conduction band of the electron transport layer. If the conduction band of a material is at a higher energy level than the perovskite's conduction band, however, there will be an energy barrier preventing this movement of electrons.

It is therefore beneficial that the conduction band of the electron transport layer is lower than that of the perovskite. This allows the electron to passively diffuse away from the perovskite.

In the hole transport layer, the energy level of the valance band should be close to or slightly higher than the perovskite's valance band so that electrons can quickly diffuse into the unoccupied "holes" in the perovskite before recombination occurs.

Perovskite Materials

View now