Excellent Woman Researcher Award of The Electrochemical Society of Japan
Solution Processing of Functional Thin Films and Their Device Applications
2023 Volume 91 Issue 10 Pages 101007
Details
2023 Volume 91 Issue 10 Pages 101007
Solution processing of thin films is a key energy-saving technology for sustainable development. Not only for its cost advantages, but it can achieve materials with unique structures and new functionalities through the combination of earth-abundant elements, especially by their hybridization with organic molecules. As the process relies on mild activation, such as low-temperature heating and irradiation with ultraviolet light having high photon energy, it is compatible with plastic film substrates. This compatibility makes it suitable for roll-to-roll mass production of flexible optoelectronic devices with large areas at a low cost. Herein, we present a review of our recent progress in the low-temperature solution processing of inorganic/organic hybrid thin films with various functionalities, including thin film encapsulation (TFE) to protect organic light-emitting diodes (OLEDs) from moisture for long-lasting operation, color-tunable photoluminescent nanocrystalline ZnO core-organic shell hybrids, and an energy-selective electron injection layer (EIL) for highly efficient inverted OLEDs. While general concept, methodology and challenges of solution processing are summarized, the future of the related technology is scoped for our sustainable development.
Functional thin films are essential components in all optoelectronic devices,1–6 ranging from inorganic to organic materials, and from crystalline to amorphous structures. These materials must be fabricated in the form of thin films in contact with other materials because their interfaces are the source of their functionalities. The bulky, redundant part of the layer represents a waste of material and adds unwanted weight. Thus, the methods to fabricate thin films are crucially important to achieve thin layers of the target materials precisely in desired composition, structure and thickness to maximize their functionalities. The process to obtain a thin film also needs to be compatible with the rest of the device, not to damage the materials onto which the layers are to be deposited.
Practicality and sustainability are also the emerging important issues in development of industrial technologies.7,8 Primarily, cost of the product thin films largely depends on the process energy, speed and scalability from small to large areas. Avoiding rare and toxic materials is utterly important for the technology to wide-spread and to last, while seeking for new and improved functionalities. When all these conditions are satisfied, social implementation of new technologies is made possible and the technology can benefit majority of people without restrictions in a long run.
The methods of thin film fabrication and choice of materials are also decisive factors to shape the end devices and their versatility in use. Inorganic materials are generally hard and brittle because they are typically crystalline. Therefore, they are often packaged in the form of small devices or flat, rigid panels, such as exemplified in light emitting diodes and Si solar panels. The materials are usually processed under extreme conditions, high temperature, high vacuum and with strong activation energy, in order to avoid defects and impurities. On the contrary, organic materials are generally amorphous, since organic molecules are bound together into solids by rather weak van der Waals interactions. Since their functionalities originate from the electronic structure of constituent molecules, the requirement in materials purity is generally not as high as that of inorganic materials. The weak intermolecular interaction also offers chances of low temperature solution processing of organic thin films, although it is a flip side of the fact that these organic materials are not heat resistant. Having achieved low temperature solution-based thin film processing of soft organic matters, the end device can be made elastic and light weight to find new area of applications in “flexible electronics”. Flexible organic light-emitting diodes (OLEDs), organic photovoltaics (OPVs), flexible wearable sensors and soft-robotics are the emerging new areas of applications. When the materials can be shaped in thin films by solution coating and printing onto thin plastic films, roll-to-roll (R2R) fast processing can also be achieved for significant cost reduction in mass production. Earth-abundant carbon-based organic materials are also beneficial to the sustainability of the technology.
Despite all the high expectations in organic electronics as mentioned above, the device performance is generally inferior to the established inorganic-based ones. The room of improvements should exist not only in synthetic chemical approaches for molecular design, but also in the development of thin film fabrication techniques. For example, the dissolution of and chemical reactions with the underlying organic layer during the deposition of another material from its solution can often be detrimental to achieving functional interfaces. Thus, chemical and thermal compatibility of the processing need to be carefully considered in completing the entire device. It is also important to note that various kinds of inorganic materials are needed, even when they are called “organic devices”. For example, inorganic compounds such as ZnO and MoO3 are often used for carrier injection and extraction in OLEDs9–11 and OPVs.12–14 Not only the materials for the internal device structure, but also the outer gas barrier coatings are made of inorganic materials, since the organic materials are usually sensitive to water vapor and oxygen. Therefore, methods for fabricating inorganic thin films also need to be developed to ensure compatibility with sensitive organic materials.
With the above-mentioned backgrounds and motivations to contribute to the sustainable development of flexible electronics, the author and the co-workers have conducted research on the low-temperature solution processing of inorganic thin films (including hybrid films with organic materials) for a variety of purposes, starting with the primary interest in the development of seamless glass coating as an ultrahigh gas barrier for OLEDs,15–21 color-tunable photoluminescent pattern22 and efficient electron injection layer for OLEDs in an inverted structure (iOLEDs).23 The research outcomes were quite promising due to their unique and useful functionalities, allowing us to claim that the solution-processed thin films are not just cheap alternatives with compromised performances, as compared to their dry-processed counterparts. Rather, the whole paradigm shift for materials design can be envisioned in such solution processes. In this article, successful research examples are reviewed and summarized to scope future development of solution processing as sustainable industrial technology.
General scheme for solution processing of thin films is graphically summarized in Fig. 1. Firstly, the coating solution needs to be prepared. A metal-containing precursor chemical is carefully chosen and dissolved in the appropriate solvents, often in combination with ligands to stabilize the metal against precipitation. Additives to act as dopants can also be considered. Aside from inorganic metal salts, metal-organic compounds such as metal alkoxides, complexes with amines, diketones and also polymeric precursors are typically chosen. In fact, significant chemical changes already can occur during the preparation of such mixed solutions. Not only the combination of the chemicals/solvents, but also the composition, concentration, temperature, and even the order in which the components are mixed can influence the chemical structure of the coating solution. For example, the degree of polymerization, from monomer, oligomer, up to polymer, thus significantly affecting the structure of thin films to be formed later on. The chemical structure of precursor should also change over time, during the aging process under different atmospheres, water content and agitation, for instance, with or without stirring. Although the experiences tell its importance to the end result, it is typically very difficult to grasp the exact chemical structure of the unstable coating solutions. Spectroscopic measurements can give some ideas to be linked to the observations during the process of thin film formation. The rule of thumb is that one needs to pay a full attention to the preparation of the coating solution, never treat it like a cooking recipe, thinking only about which chemical and how much of it to add to which solvent. We must keep it in our mind that conversion of solution to thin solid films with mild activation is only made possible by the instability of chemicals, so that such difficulties in understanding the chemical structures of solutions are inherent and something we have to live with.
General scheme for low-temperature solution processing of thin films.
Having succeeded in obtaining a homogeneous coating solution, the second step is to consider the methods to spread a thin layer of the solution onto the substrates, the process of printing and patterning. There are many ways to do this. In a laboratory scale work, spin-coating is the convenient and most frequently used method, which is typically suited to solutions with low viscosity to obtain rather thin (several tens to hundreds of nm) films. Screen printing and inkjet printing are also suitable for small scale, customized production, while achieving well-regulated patterning, as they are matched to thick and thin coating solutions, respectively. Gravure printing and offset printing are to be considered for fast mass production with low cost. In all cases, conditions such as atmosphere, temperature, and speed must be carefully controlled. This is particularly important because low-temperature processing relies on the chemical instability of the solution, allowing the material to continue changing during the printing process itself.
Then thirdly, the post-treatment is to be applied to obtain stable thin films for use. The freshly coated layer contains a significant amount of solvent that needs to be evaporated. As the layer dries, the precursor chemical condenses to further change its structure. It is obvious that one needs to take a full control of the atmosphere and temperature of the drying process. The dried coating then still contains a lot of organic ligands. In the traditional sol-gel processing, such organic additives are to be burned away by high temperature heat treatment to solidify and crystallize the inorganic components. In contrast to such processing of ceramic coatings, only mild activations can be applied to be compatible with the flexible devices. Simply reducing the temperature is one way, while plasma treatment,24,25 UV-ozone treatment,26,27 and irradiation of deep UV light such as vacuum ultraviolet (VUV) light15–19,28–30 can be combined to remove the organics. However, it is anticipated that some of the organic components remain after such mild treatment, as we have indeed seen in the research examples shown below. In fact, such residual organics contributed to the functionalities of the thin film materials we obtained.22 It is apparent that the transition from solutions to thin solid films is highly complex and dynamic. It is not at all simple to capture the exact chemical events taking place at each stage. In other words, however, we have a large room of control to obtain a variety of materials with various functionalities even when we start from the very same precursor chemicals. Gaining experiences with keen observations are therefore essential to sophisticate the solution processing.
Light weight and flexible OLEDs are attractive as advanced displays and lightings, such as foldable, rollable mobile displays, wallpaper-like large screen TV and to furnish the entire ceiling into a room light.31–33 However, such applications, especially as large-area devices, are currently challenging due to the lack of thin film encapsulation (TFE) technologies.
The materials used in OLEDs are highly sensitive to water vapor and oxygen.34–36 If the devices are not properly encapsulated with a high gas barrier, they quickly degrade forming non-emissive “black spots”, typically due to delamination of Al/Ca top contacts by their corrosion (Fig. 2). Oxidation and decomposition of organic molecules also result in reduced luminescence efficiencies. A very high level of encapsulation is therefore needed to make OLEDs usable in normal environments. Water vapor transmission rate (WVTR) of less than 10−5 g m−2 day−1 is needed for a practical use and achieving that becomes especially difficult for flexible OLEDs.37–39 While glass-sealed solid flat panels can be made in large areas, currently commercialized flexible OLEDs are limited to small sized ones for mobile phone applications. They employ multi-layered organic/inorganic flexible TFEs which are prepared typically by atomic layer deposition (ALD),40,41 applicable to production in small areas with low throughput, thus being very expensive.
Thin film encapsulation (TFE) for flexible OLEDs.
In order to achieve OLEDs in large flexible sheet forms, we need to develop fast and versatile solution processing of TFEs to be directly applied to the surface of the entire device. It is not only important to avoid high temperatures but also to be chemically compatible with the underlying organic device, namely, for the solvents and precursor chemicals used in the coating solution not to dissolve and attack the organic layer underneath. Although inorganic metal oxides are typically seen as candidates of TFEs, a high level of gas barrier performance can only be achieved when they are free of cracks and pin-holes, rather than being dependent on their intrinsic gas barrier properties.
As a promising candidate of TFE, we employed poly-dimethylsiloxane (PDMS) that can be coated from its solution.15,17 PDMS can be converted into inorganic SiOx by irradiation of VUV light at a low temperature, because high photon energy (λ = 172 nm, Eph = 7.2 eV) can cleave chemical bonds to remove the methyl groups of PDMS (Fig. 2). We tested a UV-curable PDMS specially developed by Shin-Etsu Chemical Co., Ltd. A blend of the polymer liquid A and cross-linker B could be coated onto OLED devices to form a PDMS layer without causing a damage to the OLED. However, spin-coating of the non-diluted PDMS liquid resulted in a very thick (tens of µm) PDMS. PDMS has a reticulated structure that is almost permeable to water vapor and oxygen, so it had only a marginal effect on the lifespan of OLEDs. When such a thick PDMS was illuminated with VUV light, the surface was converted to SiOx but with full of cracks due to significant densification during conversion of PDMS to SiOx to cause a large strain, thus not usable as a TFE.
Therefore, we attempted to reduce the thickness of PDMS by diluting the PDMS coating liquid. To achieve this, we explored suitable solvents. Although we found several candidates that could dissolve PDMS, we observed that they also dissolved the organic layers of the OLEDs. After thorough experimentation, we identified decamethylcyclopentasiloxane, known as D5, as a solvent that effectively dilutes the PDMS liquid without harming the OLED device. D5 is a liquid at room temperature with a boiling point of approximately 210 °C and is commonly used in cosmetics due to its volatility and non-irritating properties. By using D5, we were able to achieve thinner PDMS coatings, reaching down to a few tens of nanometers. To convert the PDMS into an extremely flat SiOx layer without cracks, we exposed it to VUV light under a nitrogen atmosphere. Observations through a transmission electron microscope (TEM) of a cross-section of the PDMS layer partially converted to SiOx after a short VUV irradiation showed a gradual change from PDMS to SiOx, which correlated with the penetration depth of VUV light. This gradient and the presence of elastic PDMS likely aided in relieving strain and preventing cracking. Employing a repetitive cycle of PDMS coating, UV curing, and partial conversion to SiOx through VUV irradiation, we directly formed an organic/inorganic multilayer structure (PDMS/SiOx/PDMS/…) onto the OLED device. Although we anticipated superior performance as a TFE, our stability tests showed that this coating only extended the device’s life by approximately 7 times compared to a non-encapsulated OLED device. It seemed that the PDMS-derived SiOx exhibited a microporous nature, inheriting the reticulated structure of PDMS, which limited its ability to effectively block gases.
Our challenges for the solution processing of TFEs did not end there. In fact, the dilution of PDMS with D5 to obtain a crack-free PDMS/SiOx coating became a key of success. Although it did not block gases, once if an OLED device is coated, any solvents could be applied as they did not pass through PDMS. Also, the surface conversion into SiOx improved wetting of surface to achieve overcoating various other materials with good adhesion. At first, we have employed metalorganic precursors of Zn(II), Sn(IV) and their blends to yield metal oxides in compositions as ZnO, SnO2, ZnSnO3 and Zn2SnO4 by VUV irradiation.16,18 It was crucial to employ anhydrous chemicals and dry solvents to prevent the crystallization of these oxides, as seen in the differing outcomes between Zn acetate dihydrate (ZAD) and acetate anhydrous (ZAA), resulting in nanocrystalline ZnO (nc-ZnO) and nearly amorphous ZnO, respectively. Presence of even a small amount of water thus strongly promoted crystal growth. Even though we could prepare very flat, dense ceramic coatings without pinholes and cracks on top of the PDMS/SiOx layer, they had only small effect to extend the lifetime of OLEDs probably because gaps between the grains of metal oxides permitted gas penetration.
Significant improvement has been achieved by overcoating poly-perhydrosilazane (PHPS)19 with a Si-N main chain (Fig. 2). PHPS is used as a glass coating to car bodies as it is converted into stable SiOx under air. It is diluted by dibutyl ether which is harmful to OLED materials, but can be coated when the device is protected by PDMS/SiOx as described above. By repeating the alternative coatings of PDMS, PHPS, and VUV irradiation in between, a tetrad as PDMS/SiOx/SiNy/SiOxNy that we named as PONT (Polymer/Oxide/Nitride/Ternary) structure could be fabricated as seen in the cross-section TEM image shown in Fig. 3. X-ray photoelectron spectroscopy (XPS) combined with energy dispersive x-ray spectroscopy (EDS) depth analysis indicate that the very surface of PHPS is converted to a ternary silicon oxide nitride (SiOxNy) by VUV irradiation even if the entire process is carried out in an N2-filled glovebox, probably due to presence of residual water and O2, whereas the internal volume of PHPS is converted to binary silicon nitride (SiNy). The densest SiOxNy appears as a sharp and thin dark border in the TEM image. It is interesting to note that extended VUV irradiation to PDMS in a glovebox resulted in a total loss of methyl group to fully convert it into SiOx. On the other hand, dense SiOxNy totally blocked O2 and water vapor, but only allowed H2 to escape, so that the internal volume of PHPS was converted to SiNy, not into oxides. Once the layer is coated by PHPS-derived SiOxNy, the gradient structure of PDMS/SiOx is preserved during the repetitive VUV irradiation. Such a feature of gas blockage of SiNy/SiOxNy coating already predict its high performance as a TFE. Conversion of PHPS to SiNy/SiOxNy also results in a strong thinning and increasing of refractive index as monitored by ellipsometry (Fig. 3). Once again, the gradient structure of SiNy/SiOxNy naturally occurring by VUV irradiation was helpful to reduce the strain to avoid cracks.
Low-temperature solution processing of a high performance TFE in PONT (Polymer/Oxide/Nitride/Ternary) structure; (a) X-ray photoelectron spectroscopy (XPS) depth profile waterfall graphs of a PHPS layer irradiated with VUV, showing its surface converted to SiOxNy while internal volume to SiNy. (b) Cross section TEM image and complimentary EDS elemental line profiles for Si, O, N, and C of seamless triple PONT coatings on Si and overcoated with PDMS. (c) Decrease of film thickness and increase of refractive index during conversion of PHPS on VUV irradiation, as determined by ellipsometry. (d) Water vapor transmission rate (WVTR) of single and triple PONT coatings. Reference 21 (CC-BY). (e) Lifetime test under constant current operation of OLEDs with an initial luminance of 1,000 Cd/m2 under ambient condition (25 °C, 50 %RH), for which inset are the pictures of the emissive areas of the OLED devices without and with respective encapsulations before and after accelerated degradation tests under 60 °C, 90 %RH. Reference 19 (CC-BY-NC-ND).
Three times repetitive coating of the PONT structure achieved a WVTR as low as 4.8 × 10−5 g m−2 day−1, which is the lowest ever known for solution processed TFEs to our knowledge (Fig. 3).21 Stability tests were performed for the OLEDs without encapsulation, only with PDMS/SiOx, single PONT TFE and glass sealing, for their constant current operation under air at room temperature (Fig. 3). The results were quite encouraging. While bare and PDMS/SiOx coated device quickly died (temporal increase of luminance due to delamination of Al to increase current density), the stability of the PONT device totally equaled to that of the glass-sealed device, achieving above 80 % of initial luminance longer than 700 h. Also, accelerated stability test to store the device at 60 °C, 90 % relative humidity resulted in no black spots for the device encapsulated with PONT.
Achievement of practically usable stability with the solution processed “PONT” structure is a significant advancement to “bridge” the dream and reality of flexible electronics (Pont in French means Bridge in English). This low temperature solution processing is fast, applicable to flexible devices for their R2R production. The technology should be further developed for encapsulation of not only large area flexible OLEDs but also other air sensitive flexible devices such as Perovskite solar cells.
The above-mentioned VUV conversion of metal-organic precursors yielded highly transparent uniform metal oxide layers.16,18 The high optical quality of the product thin films gave us a motivation to study their photoluminescent (PL) properties. In fact, we did monitor PL spectra to check materials change by VUV irradiation, as one of the analytical routines for the development of TFEs. Then we recognized a strong PL in visible range from fairly thin nc-ZnO thin films obtained by VUV irradiation to a Zn(II)-monoethanolamine (MEA) complex precursor (Fig. 4).22
VUV-photochemical conversion of Zn2+-monoethanolamine gel to color-tunable photoluminescent nanocrystalline ZnO/surface bound MEA core-shell thin films. The surface bound Zn-MEA complex is responsible to the observed variable PL based on charge transfer absorption/emission. Photographs of patterned thin film samples under blacklight irradiation, as prepared by irradiating VUV light for 2 (violet PL) and 20 min (green PL) through a mask. Reference 22 (CC-BY-NC-ND).
Interestingly, the color of PL widely changed by changing the VUV irradiation time, from purple (389 nm) to green (486 nm) for 1 to 20 min. Due to the broadness of the PL spectra and their occurrence in the visible range, it is unlikely that the emission originates from the band-edge emission of the ZnO core, which is expected to appear sharp and in the UV to violet range. Quantum size effect expected for nc-ZnO with variation of size also cannot account for the color change. Indeed, observation by TEM revealed about 4 nm constant crystallite size, irrespective of the VUV time. Extension of VUV irradiation simply increased the population of nc-ZnO but did not promote crystal growth. Increase of nc-ZnO with unchanged size was also supported by its UV absorption spectra to increase the peak intensity at a constant wavelength of 292 nm. Then, what should be the origin of the PL and why the emission energy changes?
We scrutinized variation of PL and PL excitation (PLE) spectra for the film samples with different VUV time and also compared them to those of the coating solution, namely, mixtures of ZAA and MEA with different ratios in 2-methoxyethanol (2-ME). The PLE maxima of the film samples occurred at clearly longer wavelengths than 292 nm, where the core of 4 nm nc-ZnO absorbs. In fact, the coating solutions showed rather similar PL and PLE spectra to those of the film samples, red-shifting the absorption, PL and PLE maxima on increasing MEA/ZAA ratio. In this case, however, the peak wavelengths of absorption and PLE matched. Although MEA is a frequently used stabilizing ligand in traditional sol-gel processing of metal oxide thin films, it is rare to find studies about photophysical properties of the coating solution itself, since people are mostly interested in the properties of the product thin films and the organic ligands are to be simply burnt away. Density functional theory (DFT) calculation indeed supports ligand to metal charge transfer (LMCT) excitation from O 2p of MEA to Zn 4s.42,43 On increasing the MEA/ZAA ratio, transition from singly coordinated Zn(MEA)(acetate) to doubly coordinated Zn(MEA)2 is expected as dominant complex species in the coating solution and that accounts for the observed red-shift of the spectra. Question remains about the changes for the film samples. FT-IR and XPS analysis of the films with different VUV time revealed preferential removal of acetate than MEA, so that the energetics of the surface complex, namely, surface bound acetate or MEA to nc-ZnO changes as their population changes under VUV irradiation. Consequently, the energies of LMCT excitation reflected in PLE as well as PL originating from the surface complex change on extension of VUV irradiation. A model to explain the mechanism and the cause of color-tunable PL is thus graphically shown in Fig. 4. Although the absence of 292 nm peak in the PLE spectra rules out contribution of photoexcitation of nc-ZnO core to the observed PL, the presence of ZnO surface should be important to stabilization of excitonic state of the surface complex to exhibit strong PL, since the as-spun thin film without VUV irradiation do not show any appreciable photoluminescence.
A unique feature of this thin film that only the VUV irradiated part result in PL in the visible range made us to test PL patterning employing a photomask (Fig. 4). By changing the time of VUV irradiation, we can also change the color of the PL pattern. It may be useful as a security printing for bank notes and other important documents. They may also be employed in light emitting diodes, ideally as diode laser because of their high optical quality and possibility to precisely control the film thickness for optical resonance.
We have explored possibilities to use solution processed metal oxide thin films as carrier transport/injection layers in OLEDs.23 Stability still is a serious problem of OLEDs, especially for flexible versions of them. Protection of the device with TFEs is a way to ensure the stability and we were successful to do this by VUV conversion of 1D polymeric metal-organic precursors as explained above. Another way of stability improvement can be expected for the OLEDs in inverted architectures (iOLEDs). Standard OLEDs employ a bottom-anode architecture for which the negative terminal at the top of the device requires metals with small workfunction such as Al and Ca. These metals are highly sensitive to O2 and moist, so that a high level of WVTR is needed for the TFE. On the contrary, stable noble metals with large workfunction such as Ag and Au can be employed as the top positive terminal in iOLEDs, and thereby significantly relax the demand to encapsulation.44–47
Achieving an efficient iOLEDs needs an optically transparent electron injection layer (EIL) at the bottom of the device, to be deposited on the substrate. ZnO has been considered as a promising candidate by many researchers because of its high optical transparency in the visible range and electron mobility.48–50 Since we have established methods to obtain nc-ZnO layers without high temperature treatments, we tested using them as EILs in iOLEDs. The research outcomes were quite surprising and in fact gave us many new insights to reconsider the role and requirements for carrier injection layers for efficient operation of OLEDs. We even recognized these carrier injection phenomena in solid-state devices to share a common ground with electrochemistry for charge transfer events at solid/solution interfaces.
Extremely simple chemical precursors were chosen for the solution processing of ZnO. Methoxyethoxide complex of Zn(II) (ZME) is commercially available and can be diluted in dry 2-ME. Thus, the coating solution only contains Zn2+ ions and 2-ME molecules, either as ligand bound to Zn ion or solvent. The complex is sensitive to water to be quickly hydrolyzed to Zn hydroxide and 2-ME. When the solution is coated and simply dried in an N2-filled glovebox, it results in an amorphous gel of ZME (ag-ZME). When the dried films are heated at a certain high temperature or VUV irradiated, they are converted to nc-ZnO. The regular components of the OLED structure are sequentially vacuum evaporated in an inverted order onto the solution processed Zn-containing layers to complete devices, namely, NBP (2,9-dinaphthalen-2-yl-4,7-diphenyl-1,10-phenanthroline) as an electron transporting layer (EIL), Ir(ppy)3 (ppy = 2-phenylpyridine) diluted at 6 wt% in CBP (4,4′-bis(N-carbazolyl)-1,1′-biphenyl) as an emission layer (EML), NPD (N,N′-di(1-naphthyl)-N,N′-diphenyl-(1,1′-biphenyl)-4,4′-diamine) as a hole transport layer (HTL), MoO3 as a hole injection layer (HIL) and Al as a top positive terminal. Full details of device preparation can be found elsewhere.23 Such iOLEDs with variation of EIL were evaluated and also compared to the OLEDs with the same emitting materials and in a standard architecture.
The iOLEDs employing ZME-derived EILs well converted to nc-ZnO in fact performed very poor. ZME could be fully converted to nc-ZnO either by heat treatment above 200 °C or VUV irradiation under N2.16,18,22 Although such nc-ZnO layers were highly transparent, flat and apparently did not have any pinholes and cracks, the devices gave no or small emission in spite of large current. We thereby simply reduced the annealing temperature down to 30 °C. The ZME layers annealed at 30, 60, 100, 130 and 200 °C were named as T30, T60, T100, T130 and T200, respectively. The current density-voltage-luminance (J-V-L) and external quantum efficiency (EQE)-current density curves of the iOLEDs employing T30-200 are shown in Figs. 5b and 5c, respectively. The T200 device initially performed poorly but showed improvement when the annealing temperature was reduced, ultimately achieving top performance for the T100 device, which is a hybrid of nc-ZnO and ag-ZME, as seen in the TEM image (Fig. 5a). The EQE of the T100 device indeed equals to that achieved by an optimized OLED in conventional bottom anode architecture and employing the same EML. It is remarkable that the T30 and T60 devices perform equally well, as they are primarily composed of ag-ZME. Organic ligands, namely, 2-ME bound to Zn ion cannot be fully removed by annealing at such low temperatures. Obviously, high conductivity of the EIL is not an issue for the efficient operation of OLEDs, contrary to the expectations of many researchers. Use of such uncooked ag-ZME with a lot of organics may raise concerns about the stability of the device. To our experience, however, these devices were fairly stable as long as they are properly sealed. Long term continuous operation of the T100 device under air at a constant current with an initial luminance of 100 cd m−2 approved more than 2000 h stability until the luminescence intensity was halved.23 These ag-ZME layers are not chemically stable when they are exposed to air but practically stable to function as EILs.
Low-temperature annealed solution processed ag-ZME/nc-ZnO as energy-selective EIL for highly efficient iOLED; (a) High resolution transmission electron microscopic (HR-TEM) and Atomic Force Microscope (AFM) images of ZME layers annealed at 100 °C under N2. (b) J-V-L curves and (c) external quantum efficiency (EQE) vs. V curves of the iOLEDs without and with ZME layers annealed from 30 to 200 °C (T30-200) as EIL. (d) Two-terminal cyclic voltammograms (CVs) measured at a slow scan rate of 20 mV s−1 up to the voltage of current onset for the iOLEDs employing T30-200 ZME EILs showing capacitive behavior of ZME and electron trapping/de-trapping at the ZME/NBP interface. (e) The profiles of DC current passing through the device obtained by averaging current for positive- and negative-going scans of the CVs to cancel out charging/discharging current. (f) The profiles of DC capacitance obtained by converting CVs to ΔQ/ΔV. Schematics to explain electron injection; (g) without EIL metallic ITO allows non-radiative direct electron transfer to hole in EML, so that large current (hole-only current) does not result in emission, (h) with nc-ZnO EILs (T200, T130 ZME) due to interfacial recombination current via intraband defect states to result in inefficient emission and (i) energy selective electron injection to LUMO of EML achieved with ag-ZME EILs (T100, T60, T30 ZME) due to their dielectric behavior and narrowly distributed state density originating from 4s orbital of Zn(II). Reference 23 (CC-BY-NC-ND).
Although precise determination of chemical structure and composition of these ag-ZME (and its hybrid with nc-ZnO) is difficult, combined analysis employing TEM, ellipsometry, XPS, Auger electron spectroscopy, UV-Vis and FT-IR supported increasing degree of conversion of ag-ZME to nc-ZnO on increasing the annealing temperature.23 The question now is why these ag-ZME, apparently insulating, could work so nicely as EILs. AC impedance measurement is a frequently used method to analyze OLED devices. However, since they have multiple interfaces, the interpretation of the impedance response is not straightforward. Instead, we performed two terminal cyclic voltammetry (CV) of the devices (Fig. 5d). The OLEDs are to be operated by DC, so that a DC measurement can be highly intuitive. Capacitive behavior of the devices is immediately recognized from the CVs. By averaging current of positive and negative going scans, profiles of DC current passing through the device can be extracted (Fig. 5e), whereas the derivative ΔQ/ΔV yields the DC capacitance curve (Fig. 5f). The DC current stays flat and almost zero, especially so for T30, and then exponentially starts rising around 2 V. The capacitance stays constant in this voltage range as well. The applied voltage only results in a dielectric polarization of the ag-ZME layer in this range. One can also recognize small humps in the CVs, which then appear as additional capacitance. They are most likely associated with electron trapping and de-trapping of the states generated at the ZME/NBP interface.51 The voltage of this electron trapping becomes smaller as the annealing temperature becomes higher and it almost disappears for T200. It is important to notice that the voltage for electron trapping and exponential rise of DC current coincide. The non-rectified ohmic DC current close to 0 V for the T200 device should be caused by leakage due to shorting at pinpoints.
These characters derived from the CV measurements clear the view to understand the role of ag-ZME for efficient operation of iOLEDs, as graphically explained in Figs. 5g–5i. If the device is fabricated without EIL, by directly depositing the rest of the OLED structure on top of the ITO substrate, holes injected from the anode and traveled through the EML will simply meet electron from ITO. This results in a large “hole only” current that does not lead to electroluminescence. Since ITO is metallic, its broad state density distribution and high electron concentration do not achieve electron injection to LUMO of EML, but allows direct interfacial recombination with the injected hole in HOMO of EML. When nc-ZnO is employed, electron injection does occur and we see luminescence owing to its band structure. However, as soon as ZnO is crystallized, intraband states are formed due to some defects or impurities, so that redundant recombination current via such states flows to reduce the efficiency and lifetime of the device due to Joule heating. Then, ag-ZME is almost electrically insulating and is without pinholes and cracks, so that current leakage is completely blocked. The applied voltage simply results in dielectric polarization of the ag-ZME layer, which we read as the DC capacitance. Electron traps at the interface are filled on further increasing the voltage, as associated with the humps in the CV. As the traps are filled, electrons are ready to “percholate” from the ag-ZME layer traveling via narrowly distributed state density of 4s orbital of Zn(II). In other words, electrons can spill over the dam of ag-ZME to be selectively injected to LUMO of EML for efficient electroluminescence. In order to test such ideas, we also employed methoxyethoxide of Ca(II) (CME) to prepare thin films of ag-CME in the same manner.23 Indeed, the iOLEDs employing ag-CME showed the same behavior as those with ag-ZME with almost the same efficiency. Ca(II) can offer the same 4s orbital for electron conduction as Zn(II). The electron conduction in these amorphous compounds of Zn(II) and Ca(II) may sound surprising but is pretty much the same as the achievement of high mobility in indium-gallium-zinc-oxide (IGZO), for which effective overlap of spherical 4s and 5s orbitals is also anticipated for electron conduction.1 It is also interesting to compare dielectric polarization of ag-ZME and ag-CME to the electrical double layer at the electrode/solution interface in electrochemical events. Formation of electrical double layer, thus charging of the interface, is essential for charge transfer reactions in electrochemistry.52
The superior performance of ag-ZME to nc-ZnO as EIL may appear surprising. However, previous studies to employ crystalline ZnO for iOLED in fact also needed organic molecules such as polyethyleneimine (PEI) to be deposited on ZnO surface to regulate the electron injection,53–55 as ZnO was too leaky otherwise. In these examples, narrow state density distribution of the organic modifier might have acted as gate for selective electron injection. These new insights evoke a paradigm shift in design of carrier injection layers. High conductivity associated with defect states is obviously not welcome. State density must only be available right above LUMO of EML for selective electron injection. Electron mobility is also not a big issue here because EILs are typically made so thin as a few tens of nm. The same arguments may also apply to carrier extraction in solar cells. Materials with narrowly distributed and suitably located state density should allow energy selective carrier extraction to maximize the voltage.
Historical development of our research for low-temperature solution processing of functional thin films has been reviewed, with the aims of application varied as TFEs, luminescence and electron injection for OLEDs. The biased view to solution processing has always been a cheap alternative to yield materials with compromised properties, compared to the dry processing that is established as industrial technology. However, these examples of recent development already show that solution processing can offer thin films with compositions and structures difficult to be achieved by the traditional methods, in some cases even leading to new and improved properties. Seamlessly interconnected organic/inorganic multilayer structure without pinholes and cracks could be obtained by a simple procedure to perform as a TFE for realization of large-area flexible OLEDs and solar cells. Core-shell hybrid structures with color-tunable efficient PL could be spontaneously formed. Metal-organic amorphous gel layers were found to perform as energy selective carrier injection contacts because of their dielectric behavior and narrowly distributed state density.
Correctly speaking, the outcomes of the research examples described herein were neither foreseen nor purposely, strategically designed. It is indeed difficult to precisely grasp the chemistry of solution and its dynamic transition to thin solid films. The resulting amorphous materials are often very difficult to be clearly identified. Nevertheless, it is valuable to consider shift of paradigm in synthetic approaches to functional thin films. The reason to study solution processing should not just be for its economical advantages and realization of flexible devices, but to explore the new frontier to seek for new materials. Materials with new structures, composition and properties can be produced even out of combination of earth abundant common elements, especially when they are hybridized with organic molecules. Understanding chemistry of materials will continue to be important to make these high expectations into reality. Having achieved that, inherent economic advantages of solution processing will keep the promises for sustainable development of new technologies to be accessible to all mankind.
The works reviewed in this article were carried out in the Innovation Center for Organic Electronics (INOEL) of Yamagata University, where L.Y. was a member of flexible OLED group as a project researcher and Restart Postdoctoral (RPD) Research Fellow of Japan Society of Promotion of Science (JSPS) for 2017–2023. The group leader and collaborator, Prof. Yoshiyuki Suzuri is gratefully acknowledged for his permanent support and encouragement throughout the research. All the internal and external collaborators of YU are acknowledged for making the research successful and enjoyable. Special thanks are given to Prof. Tsukasa Yoshida of the Graduate School of Organic Materials Science, Yamagata University, who is my partner both in science and personal life. Without his patient support, guidance and encouragement, conducting and completing the research were not possible. This work was financially supported by the JST A-STEP Program (AS2614016M), JST COI Program (MJCE1312), JSPS KAKENHI (20K05317), and JSPS Research Fellow (20J40137). This work was also partly supported by JSPS Partnerships for International Research and Education (PIRE) Grant Number JPJSJRP20221201.
Lina Yoshida: Conceptualization (Lead), Funding acquisition (Lead), Writing – original draft (Lead), Writing – review & editing (Lead)
The authors declare no conflict of interest in the manuscript.
Adaptable and Seamless Technology Transfer Program through Target-Driven R and D: AS2614016M
Center of Innovation Program: MJCE1312
Japan Society for the Promotion of Science: 20K05317
Japan Society for the Promotion of Science: 20J40137
Japan Society for the Promotion of Science: JPJSJRP20221201
L. Yoshida: ECSJ Active Member
Lina Yoshida (Assistant Professor, Research Center for Social Implementation Education, Department of Chemical Science and Engineering, National Institute of Technology, Tokyo College.)
Lina Yoshida received her Ph.D. in Engineering from Gifu University in 2011. She worked as a researcher in the Faculty of Engineering at Yamagata University from 2012 to 2013. After taking a four-years break due to maternity, she resumed her research activities in 2017 at the Innovation Center for Organic Electronics (INOEL) of Yamagata University. Since 2020, she worked as a restart postdoctoral research fellow (RPD) of JSPS. Since July 2023, she has embarked on her career as a specially appointed assistant professor at the National Institute of Technology, Tokyo College. Her research interests focus on solution-based processing of inorganic/organic functional thin films and their application to optoelectronic devices for sustainable development.
