The levelized cost of electricity (LCOE) is an important measure for solar (PV) technology because it shows how much it will cost to run a system that makes energy over its whole life. It used to cost about $0.27/kWh to run a utility-scale silicon solar system, but that cost has gone down over the last ten years and will be between $0.026/kWh and $0.066/kWh in 2019. This quick drop was mostly caused by lower module prices, shown in $/W (CModule/PModule), which happened because of economies of scale and better module efficiency. The BOS costs for the inverter, wire, ground, and mounts now make up about two-thirds of a solar system. This is a direct result of the modules getting cheaper.
To keep lowering the LCOE and growing the amount of solar systems installed around the world, it is becoming more and more important to make modules last longer and/or produce more power, even if it costs a little more. Lab-scale silicon solar cells, on the other hand, have recently hit a power transfer rate of 26.7%, which is close to the theoretical Shockley–Queisser (SQ) limit of 33.7%. This basic limit for single-junction solar cells means that the current method of making silicon single-junction cells better to boost module efficiency and lower LCOE will soon reach a limit. Researchers have come up with many ways to go beyond the SQ limit because of this forecast. Multijunction solar cells are one of the most hopeful of these.
There is a tradeoff between the current and voltage in a single-junction solar cell that is built into the SQ limit. A photon must have energy greater than the bandgap in order to be absorbed, so a semiconductor with a smaller bandgap will absorb more photons and make more current. Any photon that is received, though, with energy above the bandgap will create pairs of electrons and holes that heat up and disappear near the ends of the band. This means that the extra potential energy is lost to heat. To get around this trade-off and lower thermalisation losses, multijunction solar cells divide the solar spectrum among several semiconductors, with each having a different bandgap and, as a result, collecting photons with different energies.
Two-terminal (2T) or four-terminal (4T) designs can be used to make tandem solar cells. In 2T tandems, the subcells are electronically linked in series and built on a single base using a recombination layer or tunnel junction. Electrically linked in parallel, the subcells in 4T tandems can be stacked on top of each other manually, or they can be spaced apart and visually linked with a spectral splitter or a reflective mirror arrangement. We will talk about flat-panel tandems in this chapter.
Each design has its own pros and cons, and there is still no clear winner in terms of which style will sell the most. Theoretically, 4T tandems can produce a little more energy than 2T tandems because the subcells don’t have to match the current, but in practice, they are less efficient because they need an extra layer of transparent conductive oxide (TCO). They also let the process be more flexible because each subcell can be made separately until the module is put together.
Why Perovskites?
It is best to use metal halide perovskites for tandems because they have great optical qualities, low voltage loss, and high single-junction cell yields. Chemical changes can be made to them so that their bandgaps are between 1.2 and 3.3 eV, which is perfect for high possible tandem efficiencies. You don’t need epitaxial growth to make these polycrystalline semiconductors, and they can be formed at low temperatures using cheap methods like spin coating, blade coating, slot-die coating, and thermal evaporation.
3 How to Make an Efficient Perovskite-Based
Tandem?
To make a tandem solar cell, you need to pay close attention to a lot of technical and scientific details. These include the width of the NIR/IR absorption, the reflected back electrode, how well it works with bendable surfaces, how resistant it is to Sn oxidation, and the low-bandgap bottom cell. To be resistant to Sn oxidation, these perovskite-based multijunction solar cells must have a low bandgap bottom cell with an Eg of 1.2 eV or less. Low bandgap perovskite solar cells based on Sn are the only ones that can be used. This part talks about general rules for perovskite-based multijunction solar cells and the unique problems that come up when you mix different technologies.
Low Bandgap Solar Cell
The low bandgap solar cell can be paired with a wide-bandgap perovskite. Silicon, copper indium gallium diselenide (CIGS), and a mixed Sn/Pb perovskite are the best choices because they have a bandgap of 1–1.2 eV and are very efficient in the near-infrared range.
Silicon
Silicon solar cells make up more than 90% of the photovoltaic market because they are easy to make and don’t cost much to make. Silicon technologies make it possible to make customised bottom cells with p-type or n-type wafers, back or front joints, different surface finishes, and polarity. All of these things make them perfect for the early stages of making perovskite-based tandems. But Al-back surface field (Al-BSF) and multicrystalline silicon cells are likely to go away in the next ten years. Together, researchers are now looking at ways to combine the passivated emitter and rear cell (PERC) family and silicon heterojunction (SHJ). Most solar makers now use PERC-type cells. SHJ has the highest power conversion efficiency in the world (26.7%) and is predicted to be able to make more cells and get a bigger part of the market in the next ten years.
Chalcopyrites: CIGS and CIS
A thin-film technology called CIGS could be a big option to c-Si because it is easy to make, doesn’t cost much, and can be used on a lot of different surfaces. It is possible to make bendable modules because it has a small absorber layer, a high absorption rate, and a straight bandgap. To get the bandgap you want, you can change the value of xSe2 from 1.04 eV for CuInSe2 to 1.7 eV for pure CuGaSe2. This lets us use top cells with bandgaps no wider than 1.6 eV, which are the most efficient and stable perovskites right now. CIGS solar cells only make up a small part of the world’s photovoltaic market—about 1% to 2%—mainly because their modules are less efficient. Co-evaporation methods are complicated, which makes upscaling hard and limits output and return. In the future, it might be hard to get enough indium.
Because CIGS absorbers have a rough surface and an uneven shape, most of the work on developing perovskite/CIGS tandems has been done on manually stacking 4T tandems. Some solutions have been suggested, such as making the surface smoother with thick solution-processed hole transport layers (HTLs) or cleaning a thick TCO front electrode by hand. Recently, scientists have been trying to put a recombination layer and HTL on an uneven surface in a way that makes it conformal. They have found some good solutions, such as an atomic layer deposition (ALD)-based NiOx layer and materials that self-assemble into a single layer.
Sn/Pb Low Bandgap Perovskites
The best kind of perovskite tandems are all-perovskite tandems, which can have efficiencies of more than 30% and work with a wider range of substrates, even flexible ones. This gives narrow market choices that silicon-based tandems can’t reach. When you mix Sn and Pb on the B-site of the perovskite, you can get bandgaps that don’t change in a straight line, between 1.2 and 1.4 eV. However, solid all-perovskite tandems were developed later, and their best performance is still a little lower than perovskite/silicon tandems. This is because they need high-quality Sn-based perovskite and aren’t as compatible with the process.
The photogenerated current could still be made better, but the external quantum efficiency of Sn/Pb perovskites needs to be raised to more than 80%, especially in the near-infrared range that is important for tandems. With these mixes, the absorption cross section is smaller than with pure Pb perovskites. This means that their absorption is weak near the edge of their band. We also need to make the charge transport layers and energy levels better because Sn/Pb perovskites were mostly made with PEDOT:PSS as the high-temperature layer and C60 or PCBM as the electron transport layer. It is predicted that the new HTL and ETL materials, which are made to work with lower valence band energy levels, will make all of the device’s features better.
The most difficult thing about Sn-based perovskites is that Sn2+ tends to change into Sn4+. This effect can be greatly lowered by mixing Sn and Pb at the B-site, which changes the chemical path to Sn oxidation. There has been progress in making Sn/Pb perovskite devices more steady, and the first IEC tests with low-gap perovskites were safely passed. The oxidation effect and general instability of Sn/Pb low-gap cells seem to have more than one cause. More work is needed to make wide gaps and get cells that are truly stable over time.
Recombination Junction
There is a device called a tandem solar cell that blends two cells into one shape. The cells in monolithic 2T tandems are optically and electrically linked in series. In 4T mechanically stacked tandems, on the other hand, they are only optically linked in series but electrically linked in parallel. Both types of tandem solar cells need a recombination joint that works well.
For a two-terminal tandem, you need a recombination junction. For a four-terminal tandem, you would need two clear wires and something to connect them optically. To help carriers recombine, you can also use a clear conducting oxide. Light waves with less energy than the wide-gap first cell’s bandgap should be able to pass through the link as easily as possible.
For a 2T tandem to work, the n-side of one subcell must touch the p-side of the other. In a 4T tandem, the electrical orientation of each cell is picked to maximise its own performance. In a 2T tandem, both subcells must be orientated in the same way.
To let charge carriers that are opposite each other run from both subcells and back together, a special connection is needed. To keep the series resistance as low as possible, the joint needs to be clear, not mirrored, have low horizontal conductivity, and high vertical conductivity. It must also keep the two subcells from mixing over the life of the gadget.
At the moment, TCOs are the most popular recombination junction materials for all perovskite-based tandem configurations. They can be formed using sputtering, chemical vapour deposition, or ALD, all of which are methods that can be used in industry.
Nanocrystalline Silicon Junction
Even though TCOs are very good at conducting electricity, they are not the best choice for tunnel joints because they can help move charges laterally and create shunt paths through perovskite cells. When the cells are made big enough for use in factories, this problem becomes more important. Highly doped p–n tunnel diodes can help solve this problem. They are made from hydrogenated nanocrystalline silicon that has been doped with trimethylboron or phosphine. These diodes offer very uneven electrical conductivity. Not only that, but TCOs also have problems with optics, which lead to optical interferences and parasitic absorption losses in the near-infrared range. Tuning the refractive index and widths of the nc-Si:H tunnel junction with oxygen is one way to solve this problem. This creates a wide range of optical qualities that can be fine-tuned for device optics.
Recombination Layer for All-Perovskite Tandems
To keep the first subcell from getting damaged while the second is being deposited, all-perovskite tandems need a recombination junction that is clear, doesn’t conduct electricity laterally, and is good at blocking solvents. To do this, oxide bilayers like SnO2/ITO, SnO2/ZTO/ITO, and AZO/ITO can be used. All-perovskite tandems should be made with more advanced joints in the future. These junctions should include materials whose refractive indices can be changed and intermediate mirrors. Vapor-based methods for handling the perovskite absorber, especially for the subcell that was formed twice, could be a key way to keep solvent damage from happening.
Wide-Bandgap Perovskite Solar Cell
For example, to get the most out of a wide-bandgap top cell or a sun-facing cell in a tandem, the design needs to be different from that of a single-junction device. For example, the voltage, current, and fill factor need to be maximised. For two-terminal tandems, the perovskite subcell(s) need to be worked on at low temperatures. This lets silicon heterojunction bottom cells be used. Low temps are even more important for all-perovskite tandems because none of the perovskite-based devices on the market right now are safe at temperatures above 150 °C. Putting the second subcell on top of a perovskite cell needs to be done at lower temperatures. This means that foldable solar cells on plastic sheets may become possible.
Also, be careful with the chemicals that are used to work on the perovskite absorber and its charge transport layers, since these can hurt parts below them in the tandem. This needs a recombination layer that is also a good barrier in all-perovskite tandems. The processing methods should work with rough surfaces, be scalable, and be compatible with industry standards. This will make it easy to change the absorption makeup.
For a highly efficient solar cell to work, the wide-gap perovskite subcell needs to be semitransparent. This lets as much light through as possible in the sub-bandgap spectral region. This makes strict optical requirements for charge-selective layers and clear electrodes that can’t absorb light in the short wavelength region of the top cell perovskite absorber or the longer wavelength region of the bottom cell.
The bandgap energy of the wide-gap cell has a lot to do with how much light it lets through. It is recommended that the perovskite absorber in the top cell have a bandgap that works well with that of the bottom cell. When solar cells use wide-bandgap perovskite emitters (>1.65 eV), their Voc losses are usually bigger than when they use narrow- or medium-bandgap perovskites. One reason for this shortage is the photoinduced segregation of halides in wide-bandgap perovskites. These are usually made by mixing iodine and bromine on the X-site of the ABX3 structure.
It has been shown that perovskite photostability can be improved by changing the shape of the perovskite and the quality of the film, adding things like potassium iodide to lower the number of defects, or covering the perovskite surfaces with layers like polyethylene oxide. But there is still no agreed upon way to stop halide phase segregation that lets you get all the tandem-relevant bandgaps without some photoinduced phase segregation.
Mitigating Optical Losses
Light management is a central topic in any tandem solar cell development.
Minimizing optical losses has been the driving force in the efficiency race of
perovskite/silicon tandems. In this section, we discuss an overview of a few
key optics-related generic problems that need to be considered to build efficient
tandems.
Parasitic Absorption Losses
For photovoltaic solar cells to work, light has to be absorbed by the active absorbent material, which could be electrodes or charge transport layers. In multijunction solar cells, thick and/or doped HTLs and ETLs should not be used to avoid parasitic absorption. This problem is best shown by the hole transport material spiro-OMeTAD, which was used in the early stages of developing perovskite/silicon tandems but quickly caused problems in both 2T and 4T tandems. Its dopants cause free carrier absorption losses in the near-infrared range, which lowers the currents in both the top and bottom cells.
Because of this issue, a lot of work has gone into making top charge transport layers, buffer layers, and electrodes that are very effective and clear. The best transparent top contacts for wide-bandgap perovskite solar cells have a p-i-n top cell with a C60 electron transport layer thermally evaporated on top of the perovskite. This is followed by ALD of a tin oxide/zinc-doped tin oxide composite sputter buffer layer and finally a sputtered TCO. In almost all record perovskite/silicon, perovskite/CIGS, and perovskite/perovskite tandems, this arrangement has been used.
There are also parasitic absorption losses that come from transparent electrodes, which are usually made from sprayed TCOs. To keep free carriers from being absorbed in the near-infrared range, they should be able to move around a lot and have few of them. Although thicker TCOs have better electrical qualities, they also absorb more light in the low wavelength range.
Due to the need for more light-passing electrodes, parasitic absorption losses may be the main thing that keeps us from making manually stacked 4T tandems that work better than 2T monolithic tandems.
Reflection Losses: Front, Middle, and Back
Large changes in the refractive index between neighbouring layers can cause reflection in perovskite/silicon tandems, which can lead to reflection losses of several mA/cm2. The quarter-wavelength rule can be used to find an answer: use materials that don’t reflect light, like LiF and MgF2. PDMS microtexture antireflection foils can also be used as an alternative. These are put on top of the front electrode during the measurement process. This method can greatly lessen the bad effects of reflection and interference patterns when used with flat surfaces, giving an absolute gain of about 2 mA/cm2 in Jsc.
Using nc-SiOx:H to change the refractive index of the recombination junction can make the external quantum efficiency (EQE) of a single-piece perovskite/silicon pair better. Light scattering can be made better on the back of a tandem to make it look better. For silicon-based tandems, the back can have micron-sized pyramids added to it, while the front can stay flat so that it works with solution-processed perovskite top cells. Adding a low-refractive-index material between the rear TCO and metal can help reduce plasmonic absorption losses in the rear metallisation. This can be done by increasing the interior rear reflectivity at long wavelengths.
To sum up, perovskite/silicon tandems can be made better by adding microtexture antireflection foils and antireflective coats and making light spread more. In many situations, these techniques can make a big difference in how well and how quickly something works.
Textured Substrates
Textured surfaces can help solar cells stop reflection losses by cutting down on front reflection losses and optical interferences and making it easier for light to get through. This is usually done with an etching method that makes micron-sized peaks on the chip surface for crystalline silicon solar cells. These pyramids are higher than most perovskite absorber layers, which are 500 nm thick. However, perovskites are made by processing them in a solution, which doesn’t work with surfaces that are rough or wrinkled. So, for perovskite/silicon tandems and all-perovskite tandems, flat ITO/glass plates and front-polished silicon bottom cells are often used.
In June 2018, the first test of a perovskite/silicon monolithic tandem using a chip with a rough double side was released. It showed a 25.2% efficiency and met current densities of more than 20 mA/cm2. This pair was created using a stepwise hybrid deposition method, which started with a CsPbIBr template that was evaporated at high temperatures. This was followed by spin coating and the introduction of a formamidinum iodide (FAI) solution. A perovskite film could be made in a straight line on Si pyramids that are only a few microns tall by interdiffusion. Off-axis and scattered light play a big role in how well a solar cell works generally, so this material is very important for modelling energy output.
Textured surfaces would also help all-perovskite tandems, especially to improve the responsiveness of low gap perovskites near their band edge and make it easier to make absorbers that are very thick. Thermal evaporation and chemical vapour deposition are two vapor-based deposition methods that are very interesting, but they will need more study from the community.
Current Matching Versus Power Matching
Matching the current is very important in monolithic tandems because the photocurrent of the pair is limited by the photocurrent of the smaller of the two subcells. If one subcell makes more current than the other, the extra current is lost. An increase in voltage can help make up for this loss in some ways. But for the best combination performance, it is very important to carefully optimise the width and bandgap in the wide-gap cell. It is best for the bandgap in the wide-gap cell to be as close to the best spectral splitting value as possible, as long as the voltage also goes up at the same time. It might be better to build the tandem cell so that it matches power (same Jmpp for both subcells) instead of current (same Jsc for both subcells) in some situations. For tandems to be used in the real world, the modules will be designed based on where they will be used and the solar range that will be available there. Spectrum-specific power-matching optimisation methods could make energy output a lot higher.
Toward Commercialization
Energy Yield
The word “efficiency” in solar cells generally refers to how well the cells convert light into electricity under standard test settings, like 1000 W/m2, the AM1.5G spectrum, and 25 °C. However, solar panels don’t usually work in these exact situations because the sun’s angle, strength, range, and temperature can change. Because multijunction solar cells depend on tight current distribution, changes in the spectrum affect them more. To take into account how things really are, the energy output tells us how much power a solar system makes each year. It has been studied that the energy output is not very different between a 4T tandem and a 2T tandem. It is also not very different between a 2T tandem and a single-junction cell. Standard test settings can still give a good idea of how much energy something will produce when it’s outside. However, there isn’t any outdoor data for perovskite-based tandems that is available to the public that could back up these claims. To get even more energy, the current-matching conditions could be changed to fit the end working site.
Cost
The LCOE (Long Life Cycle Cost) of a solar cell is based on three key things: how complicated the technology is, how well it works, and how long it lasts. Getting the LCOE down means cutting down on the number of steps needed to make something, using techniques that are common in industry, and using cheap materials. When compared to single-junction modules, tandem modules should have at least a 15% efficiency gain. This is because they take up less space, lower BOS costs, and allow for more complicated devices or advanced module packaging. For instance, it is possible to make perovskite-based tandems 30% efficient, which would lead to a 3% drop in LCOE. The lifetime of the part also has a big effect on the LCOE, since a longer guarantee term lowers the LCOE. Because c-Si technologies come with 25–30-year guarantees, more study needs to be done to make perovskite-based solar cells more stable. Perovskites may not be able to sell in all places if high security is not achieved.
Market Choice
At least for now, perovskite-based tandem solar cells are not likely to go after the same market. Similar to PERC-upgraded BSF cells, they can be made better with steps that don’t cost a lot of money. Residential rooftops, nonresidential roofs, and utility-scale systems are the main places where perovskite/silicon tandems are used. In these markets, stability is very important, and perovskite/silicon tandems should have the lowest hurdles to entry. All-perovskite tandems and perovskite/CIGS tandems, on the other hand, have big benefits, like lower production costs and devices that are light and versatile. With these, special markets could open up in areas like IoT, flight, drones, transportation systems, and BIPV. Long-term security isn’t as important, and making sure it lasts 5–10 years might be possible. Perovskite-based tandems will have to compete with III-V semiconductors and thin-film silicon. This makes it even more important that they are both very efficient and cheap.
Beyond Tandems: Triple?
Triple-junction solar cells built on perovskite could improve power transfer efficiency by cutting down on thermalisation losses. It is thought that the limit for single-junction cells is 33%, the limit for tandem cells is 46%, and the limit for triple-junction cells is 52%. At the moment, thin-film silicon solar cells and III-V semiconductors can be used to make triple-junction solar cells. Thin-film silicon solar cells are cheap but not very efficient, and III-V semiconductors are expensive but very efficient. Triple-junction solar cells made of perovskite could be very efficient and not cost a lot of money.
First tests have been done with all-solution-processed allperovskite triple cells and a perovskite/silicon/perovskite triple cell. But optical models show that switching from double to triple junction won’t make all-perovskite triple-junction cells more efficient. This means that the extra processing time and cost probably aren’t worth it. If the perovskite bandgap could be dropped below 1.2 eV and closer to 1–1.1 eV, triple-junction cells would be more interesting. A perovskite-based triple junction on Si shows good possible efficiency gains (about +3%abs from tandem) that would make the extra cost of making it worth it.
Concluding Remarks
Perovskite/silicon tandems should have an efficiency of over 30%, which would be the best performance ever seen in photovoltaics. This is a huge step forward, especially for tandem perovskite/CIGS and perovskite/perovskite. The long term looks bright for all-perovskite multijunction solar cells, even if they are first sold with silicon. But a lot of work needs to be done to get into this competitive market, especially to make it stable enough to meet the needs of top solar technologies. Concerns about lead poisoning need to be handled in order to follow new rules or get into large user markets like IoT uses. Future study should focus on turning lab-scale record cells into products that can be sold, getting rid of expensive materials and methods, making processes easier, and making devices more durable and increasing production output.
Leave a Reply