Stable Perovskite Solar Cells with Atomic Layer Deposition (ALD)

Much research has gone into hybrid organic-inorganic perovskite solar cells, as they offer high conversion efficiency as well as cost-effectiveness. One major problem with these cells is their lack of environmental stability. Many scientists are working to increase the stability of these cells, while maintaining a high performance level. One promising area is introducing thin films of inorganic material into the design of the device.

ALD is crucial to this area, as it allows film growth and quality to be closely controlled. There is growing demand for tools which can be used for a variety of processing needs and conditions, as well as being capable of operating with both plasma and heat (both in research and in pilot production). For these applications, Oxford Instruments has developed the FlexAL® and OpAL® tools. This article explores the properties of ALD and how it is used in perovskite solar cells.

Atomic Layer Deposition

In atomic layer deposition (ALD), thin films are formed in cycles by exposing the substrate surface to a series of vapor-phase or gas-phase species one after the other, with alternating gaps between each dose. Each cycle sees the deposition of one submonolayer of material. Figure 1 shows a typical cycle, which has four steps.

  1. The first is a precursor dosing step. The precursor is usually a metal-halide, an inorganic metal-organic, or a metal-halide like TMA. It is often the metal center like aluminum.
  2. A purge and pump step
  3. A coreactant step which uses a small molecule like hydrogen peroxide or oxygen plasma, normally a non-metal of this sort
  4. A purge and/or pump step again

The reaction of the precursor and the reactant form a film, which is, in this case, Al2O3. It’s important that the reactant reacts with the precursor in a self-limiting manner. Moreover, the reactant and the precursors must not react with themselves or with the new surface groups formed by them. When purge and/or pump is occurring (during Step 2 or Step 4), any newly generated gaseous reaction products and residual precursor or co-reactant molecules, are removed from the ALD reactor. This is essential as the precursor and co-reactant molecules could react directly in the gaseous phase or on the substrate, which could result in unwanted chemical vapor deposition (CVD).

A schematic representation of the various steps in an ALD cycle consisting of two half-reactions. The exposures in the first half-cycle (precursor) and second half-cycle (co-reactant) are self-limiting such that the process stops when all available surface sites are occupied. The two half-cycles are separated by purge steps. The lower panels show the resulting coverage, or growth per cycle, as a function of exposure or time for that particular step. For sufficient exposure, saturated growth is obtained, while insufficient exposure results in incomplete saturation. For insufficient purging, a CVD component from mixing of the precursor and co-reactant is obtained.

Figure 1. A schematic representation of the various steps in an ALD cycle consisting of two half-reactions. The exposures in the first half-cycle (precursor) and second half-cycle (co-reactant) are self-limiting such that the process stops when all available surface sites are occupied. The two half-cycles are separated by purge steps. The lower panels show the resulting coverage, or growth per cycle, as a function of exposure or time for that particular step. For sufficient exposure, saturated growth is obtained, while insufficient exposure results in incomplete saturation. For insufficient purging, a CVD component from mixing of the precursor and co-reactant is obtained.1

The benefits of ALD include:

  • Accurate control of thickness
  • High conformal coating even if the structure has a high aspect ratio
  • Minimal pinholing, or particle generation by gas phase reactions
  • Dense and very thin films because of the lower deposition rates as against the higher deposition seen with CVD and physical vapor deposition (PVD), with the thickness being typically 1 Å/cycle
  • Ability to treat a large range of materials
  • Deposition temperature range extends down to low substrate temperatures (i.e. when plasma is used)

Figure 2 represents the key metrics for 3D coverage of a substrate:

The coverage metrics of a film on a substrate with 3D features. Coverage of the planar surface is evaluated using the uniformity while the coverage of 3D features is evaluated using the conformality. The growth control over the film thickness itself is another important metric. The ability to achieve these metrics at low temperatures is an additional important aspect.

Figure 2. The coverage metrics of a film on a substrate with 3D features. Coverage of the planar surface is evaluated using the uniformity while the coverage of 3D features is evaluated using the conformality. The growth control over the film thickness itself is another important metric. The ability to achieve these metrics at low temperatures is an additional important aspect.1

Thermal and Plasma ALD

H2O is the most commonly reported source of oxygen for metal oxide ALD2. However, it is difficult to clean from the reactor, especially at very low deposition temperatures (i.e. 100 oC or less). This is due to the strong tendency that water has of adsorbing to surfaces. Another disadvantage is the low reactivity of water at low temperatures, which leads to deposition failure. NH3 is also used for nitride growth, but takes some effort to purge from the reactor and is not reactive enough. The reactivity can be increased by making use of species that have a comparatively short lifetime as plasma species.

When ALD is performed with plasmas, the plasma species is produced either inside or close to the reaction chamber. When using plasmas, it is possible that the surface chemistry is affected by some very reactive species, such as radicals and ions3. The source gas is dissociated using radiofrequency power, which makes it possible for oxygen to be dissociated into plasma species like O2, O+ ions, and electrons. Using remote plasmas (which are used in Oxford Instruments Plasma Technology ALD platforms), both radicals and ions are present, and their levels can be tuned by changing the plasma power pressure.

When the plasma pressure is high and the power is low, for instance, ion energy becomes negligible. This is also the case regarding the flux, which means that radicals then represent the main species, e.g. O atoms in an O2 plasma, or H atoms in an H2 plasma. The operating conditions can be changed so that plasma power is high and pressure low to attain the greatest plasma reactivity. This results in the highest densities of radicals and moderate ion energies. Such conditions promote the growth of nitrides with good conductivity like TiN. Processes that use plasma are often called “plasma ALD” and others are then called “thermal ALD”.

The benefits of using plasma ALD include:

  • Better material properties like density and composition of the film and lower impurity level
  • Better range of precursors and materials
  • Lowering of the required substrate temperature
  • Potential for pretreatment of the surface, such as activation or cleaning
  • Better growth rate
  • Briefer nucleation period duration, since thermal ALD can sometimes cause a delay in growth when they are nucleating on some materials

Oxford Instruments FlexAL and OpAL ALD Tools

An ALD reactor must fulfill the following conditions:

  • Heat the reactor to the right deposition temperature
  • Inject pulses of precursors and co-reactants into the reactor volume
  • In many cases, purge reactor volume with an inert gas in between successive steps
  • The reactor volume being pumped continuously to keep the pressure around 80 mTorr

The reactor design must, therefore, keep these needs in mind. Most reactors for ALD are based on CVD reactors, but ALD is different as it can be considered a discontinuous deposition process. For this reason, filling and purging of the reactor volume is essential and must be carried out efficiently for the cycle time to be kept within reasonable limits. The precursor and co-reactant must also be pulsed automatically to keep the timing precise.

When plasma ALD is performed, the co-reactants are generally species which do not last long, and this means that the equipment used must produce the required species relatively close to the substrate4. Figure 3 is a schematic representation of Oxford Instruments’ FlexAL tool for ALD. OpAL (also from Oxford Instruments) is similar but does not have a turbopump, nor an automatic pressure control (APC), being an open-load platform.

However, both these devices are state-of-the-art plasma ALD systems with standard thermal ALD. They can perform both plasma and thermal ALD cycles within the same recipe, without any need to change the hardware when switching modes. For instance, thermal Al2O3 may be used in the beginning on a sensitive interface, and then plasma ALD may be continued with to obtain the best material properties.

Schematic of the FlexAL ALD system. The key components are indicated as well as some of their benefits.

Figure 3. Schematic of the FlexAL ALD system. The key components are indicated as well as some of their benefits.

When compared to thermal CVD, thermal ALD can yield excellent material properties at lower deposition temperatures (200o - 400 oC). Despite this, thermal ALD can suffer with respect to growth rates, cycle times and material purity at lower temperatures (e.g. between 25o and 100 oC).5 Plasma ALD is therefore suitable for lower temperatures due to its greater reactivity, while several oxides like Al2O3, TiO2 and SiO25 have been deposited at temperatures as low as room temperature.

Perovskite Solar Cells

A perovskite cell is a type of solar cell that acts as an active layer that traps solar light,and features a perovskite structure compound (e.g. an organic-inorganic lead halide material).

Materials in the perovskite category, such as methylammonium lead halide, are manufactured at no great cost, and by a simple process. The solar cell efficiency of devices made with these materials has gone up over the years, from about 3.8% in 2009 to 22.7% to the latter part of 20177. This shows how quickly solar technology is progressing in this area, compared to industry competitors.

Perovskite solar cells are being researched intensively in order to achieve higher efficiencies at lower costs. In fact, perovskite solar cells are unique in this field as they possess sufficient current and band gap regulation alongside thin film technology, which allows them to be combined in the same device with c-Si solar cells. Nonetheless, they have serious issues in that the largest is only as big as a fingernail and breaks down rapidly when exposed to most environments. It is only of late that mini-modules of larger size have been shown to be possible in a laboratory setting.8

The most common way to fabricate perovskite devices is via solution-based processing, which produces a mixed absorber layer of organo-lead that cannot withstand high temperatures. For example, a temperature of 100 oC causes the exposed perovskite to undergo layer damage if exposed for too long. Thus plasma ALD, which can operate well in low temperatures, is probably advantageous in fabricating perovskite solar cells.

 Improved efficiency and stability by thermal ALD Al2O3 on OpAL as demonstrated by the efficiency of perovskite devices with and without ALD Al2O3 as a function of storage time under varied humidity conditions (see labels top axis).

Figure 4. Improved efficiency and stability by thermal ALD Al2O3 on OpAL as demonstrated by the efficiency of perovskite devices with and without ALD Al2O3 as a function of storage time under varied humidity conditions (see labels top axis).12

Benefits of ALD in Perovskite Solar Cells

Both plasma ALD and thermal ALD are potentially beneficial to the manufacture of perovskite solar cells, and this section explains how. Some characteristics of ALD film include minimal pinholes, deposition at low temperatures, superior quality of material, very high process control, and interface engineering - all of which can be harnessed to facilitate the production of better perovskite solar cells.

Perovskite devices degrade easily, both when exposed to extrinsic factors and intrinsic factors. Extrinsic factors include oxygen, moisture, light, and temperature; while intrinsic factors comprise interfaces and species diffusion.9 To minimize the effects of moisture and oxygen on these devices, moisture barriers can be used. ALD can be utilized to produce excellent moisture barriers and has the potential to deliver similar improvements in perovskite solar cells. It is possible to use multiple materials as this type of barrier, and it is here that low temperature plasma ALD of Al2O3 has proved most valuable (as well as in low temperature plasma ALD of SiN).

Perovskites break down easily mostly due to inherent factors like species diffusion and device interface degradation. As shown in Figure 4, solar cell efficiency can be compared using cells with and without Al2O3 deposited straight on the perovskite absorber layer by thermal ALD before finishing the device. The ideal thickness of the barrier is achieved after 10 cycles of ALD (increasing the thickness would make it too much of an insulator and hinder proper hole current extraction). It is worth noting that, with Oxford Instrument tools, the ALD process can make use of partial pressures of water which do not reduce the efficiency of the perovskite layer described here. This introduces the possibility that thermal ALD processes can be used directly on the perovskite layer. Just one dose of water with these tools can boost relative humidity to about 0.1%, which is far below that of the ambient setting of the device. This is currently the only report on the direct use of ALD on the perovskite layer that has resulted in cells showing greater efficiency.

In addition to adding layers of protection, the research aimed at improving the intrinsic stability of perovskite solar cells. The key here is the inorganic layers, which act as carrier transport (generally organic layers are used in a perovskite cell). The inorganic layers should provide blocking for one carrier (e.g. holes), while conducting the other (e.g. electrons).

This arrangement should be reflected in the band diagram that emerges from the solar cell. In order to achieve sufficient blocking with high dielectric characteristics, it is better to use plasma ALD, due to the high quality of the material and the low levels of pinholing. In this setting, TiO2 plasma ALD performs just as well.8, 13 Figure 5 shows a future concept stack where the position of the perovskite (by the Pb-containing layer) is displayed alongside the stacks of plasma ALD TiO2 and MoO3.

In situations where contact with O2 plasma needs to be avoided, other plasmas such as N2 or H2 may be used as they have been proven to maintain power semiconductor devices intact while processing is occurring.14-18 The use of N2 plasma and subsequent exposure to water should produce oxide material, while still keeping the nitrogen content of the perovskite at original levels, due to the N2 plasma.

Another option is to deposit a thin layer using thermal ALD, for example by using 10 cycles of Al2O3 in the manner described above. Oxford Instrument tools can be used for thermal and plasma ALD alternately, without changing any hardware. It is also possible to combine these two processes (thermal and plasma ALD) with ease, and a few cycles of ALD are expected to strongly protect the perovskite.12

However, this might not be the case for thermal processes in general, even if it applies to Al2O3, as the surface of the perovskite material could be impacted by a delay in nucleation, or by a chemical interaction. In one case, performing thermal ALD of ZnO failed to cause growth but chemical etching of the perovskite occurred instead.9

This type of situation would probably benefit from a plasma treatment or from introducing an interlayer of Al2O3. Overall, thermal-sensitive material would greatly benefit from rapid processing, as this would avoid perovskite damage from long periods of exposure to 100 oC. Furthermore, the main application of ALD might not be directly on the perovskite, but at other interfaces in the device.

TEM image with elemental mapping acquired using Energy Dispersive Spectroscopy (EDS) of future concept of a perovskite (methylammonium lead iodide) film sandwiched between two plasmaassisted ALD layers deposited using the OpAL tool. The Pb containing layer shows the position of the perovskite material and the Ti and Mo indicate the inorganic carrier transport layers. 25 nm of TiO2 was deposited at 150 °C and the 20 nm of MoO3 on top was deposited at 50 °C.

Figure 5. TEM image with elemental mapping acquired using Energy Dispersive Spectroscopy (EDS) of future concept of a perovskite (methylammonium lead iodide) film sandwiched between two plasmaassisted ALD layers deposited using the OpAL tool. The Pb containing layer shows the position of the perovskite material and the Ti and Mo indicate the inorganic carrier transport layers. 25 nm of TiO2 was deposited at 150 °C and the 20 nm of MoO3 on top was deposited at 50 °C.9

ALD can be used to deposit a host of materials and functional layers, using: metal oxides with n-type electron-transport layers (e.g. TiO2, SnO2, and ZnO); passivation layers that have ultra-thin TiO2 (3-6 nm) films; sub-nanometer Al2O3 layers; hole-transport layers, such as p-type metal oxide layers (e.g. NiO); high work function metal oxides, like NoO3; as well as metal oxides with a wide band gap (e.g. Ga2O3 or HfO2) and tunnel junctions/interface layers. All such processes or starting recipes are available from Oxford Instruments.

Perovskite solar cells are also being analyzed for their role in indoor energy harvesting so that indoor wireless sensors and devices can be designed.13 In such a setting, the solar cells ought to show reasonable efficiency with low levels of light, whereas the efficiency of a solar cell is measured in terms of the sun levels in daylight (which is 2-3 magnitudes higher). The loss of carriers is much more obvious when the light levels are low.

To counteract this and improve efficiency, it was found that using a blocking layer created by ALD with very little pinholing was useful. At a light level of 400 lux, a flexible perovskite solar cell yielded an efficiency of 12.1% with the addition of a dense 11 nm blocking layer on PET/ITO substrates using plasma ALD TiO2 ; this was far more efficient than any other flexible solar cell technique, including a-Si, dye-sensitized organic, and CIGS, when measured under identical testing conditions.13

In a thin film perovskite solar cell, the electron transport layer can be ALD TiO2 even in the face of non-ideal band alignment. However, by using short CF4 plasma treatment, the conduction band alignment between TiO2 and the perovskite is changed for the better, providing better adhesion.17 The work described in this article was carried out on a barrel etcher, but it is expected that Oxford Instruments’ ICP etcher is capable of reproducing these results.

Conclusion

It is still difficult to obtain perovskite solar cells that are efficient while also being stable. Despite this, ALD techniques promise to help overcome these obstacles by means of their attributes, which are well-suited to this task. The important points here are the minimal pinhole levels, which are crucial in providing moisture barriers and blocking layers. The top surface of a perovskite device can be processed only if low temperature deposition can be achieved, and this is another area in which ALD can play a major role. Overall, defects can be reduced in devices and interfaces by harnessing ALD’s superior material quality, fantastic process control and the ability to engineer interfaces.

Other processes that are performed at nanoscale or even atomic scale in addition to ALD prove to be very useful in developing perovskite solar cells. One example is in etching unwanted crystal orientations or etching away a defect. Another is tuning the band position by plasma treatments or doping, as well as using other chemical treatments in the gas phase. 2D materials are not considered to be relevant at present knowledge, but it is important to explore the possible role of other sulfides. With excellent sensitivity and improved control, the range of techniques that Oxford Instruments provides can now be utilized in a variety of nanoscale applications.

This information has been sourced, reviewed and adapted from materials provided by Oxford Instruments Plasma Technology.

For more information on this source, please visit Oxford Instruments Plasma Technology.

References

  1. M.M. Kessels, Handbook of Crystal Growth, 2nd edition, p. 1101–1134 (2015).
  2. V. Miikkulainen, M. Leskelä, M. Ritala, and R.L. Puurunen, J. Appl. Phys. 113, 21301 (2013).
  3. H.B. Profijt, S.E. Potts, W.M.M. Kessels, and M.C.M. Van de Sanden, J. Vac. Sci. Technol. A 29, 50801 (2011).
  4. S.E. Potts and W.M.M. Kessels, Coord. Chem. Rev. 257, 3254 (2013).
  5. M.D. Groner, F.H. Fabreguette, J.W. Elam, and S.M. George, Chem. Mater. 16, 639 (2004).
  6. S.E. Potts, H.B. Profijt, R. Roelofs, and W.M.M. Kessels, Chem. Vap. Depos. 19, 125 (2013).
  7. NREL, https://www.nrel.gov/pv/assets/images/efficiency (n.d.).
  8. F. Di Giacomo, V. Zardetto, A. D’Epifanio, S. Pescetelli, F. Matteocci, S. Razza, A. Di Carlo, S. Licoccia, W.M.M. Kessels, M. Creatore, and T.M. Brown, Adv. Energy Mater. 5, 1401808 (2015).
  9. V. Zardetto, B.L. Williams, A. Perrotta, F. Di Giacomo, M.A. Verheijen, R. Andriessen, W.M.M. Kessels, and M. Creatore, Sustain. Energy Fuels 1, 30 (2017).
  10. W. Keuning, P. van de Weijer, H. Lifka, W.M.M. Kessels, and M. Creatore, J. Vac. Sci. Technol. A. 30, 01A131 (2012).
  11. A.-M. Andringa, A. Perrotta, K. De Peuter, H.C.M. Knoops, W.M.M. Kessels, and M. Creatore, ACS Appl. Mater. Interfaces 27, 22525 (2015).
  12. D. Koushik, W.J.H. Verhees, Y. Kuang, S. Veenstra, D. Zhang, M.A. Verheijen, M. Creatore, and R.E.I. Schropp, Energy Environ. Sci. 10, 91 (2017).
  13. G. Lucarelli, F. Di Giacomo, V. Zardetto, M. Creatore, and T.M. Brown, Nano Res. 10, 2130 (2017).
  14. S. Huang, Q. Jiang, S. Yang, C. Zhou and K.J. Chen, IEEE Electron Device Lett. 33, 516 (2012).
  15. A.J. Kerr, E. Chagarov, S. Gu, T. Kaufman-Osborn, S. Madisetti, J. Wu, P.M. Asbeck, S. Oktyabrsky, and A.C. Kummel, J. Chem. Phys. 141, 104702 (2014).
  16. J. Son, V. Chobpattana, B.M. McSkimming, and S. Stemmer, J. Vac. Sci. Technol. A Vacuum, Surfaces, Film. 33, 20602 (2015).
  17. V. Zardetto, F. Di Giacomo, H. Lifka, M.A. Verheijen, C.H.L. Weijtens, L.E. Black, S. Veenstra, W.M.M. Kessels, R. Andriessen, and M. Creatore, Adv. Mater. Interfaces 5, 1701456 (2018).

Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback
Submit