Applying OLEDs in a Manufacturing Process
The organic chemical compounds used in OLED display manufacturing require careful appreciation, characterization, and analysis. When they are paired with the right manufacturing processes, impressive results can be achieved.
by Kai Gilge, Ansgar Werner, and Sven Murano
SINCE the first publication of modern organic light-emitting-diode (OLED) structures by Ching Tang and Steven Van Slyke in 1987, a billion-dollar business has evolved, mainly due to the huge success of active-matrix OLED (AMOLED) displays in recent years. Along with this industrial success came significant developments and improvements in device manufacturing, especially with regard to throughput and substrate size. The dominant material deposition technology, however, remains the same as in the first fundamental work by those early Kodak researchers – thermal evaporation under high-vacuum conditions. Of course, in order to enable industrial application of this evaporation coating technique, several challenges had to be overcome with regard to tools, processes, and materials.
In this article, we will examine the material properties of the organic chemical compounds used in today’s OLED display processing landscape and highlight additional requirements expected to become critical for the next generation of displays. We will also provide insight into how these properties can be tailored during the product development process. Finally, we will offer a glimpse at what OLED manufacturing might look like 5–10 years from now.
Deposition Equipment Developments
To better understand the material requirements for organic semiconductors in OLED manufacturing, it is necessary to examine the processing details. The original thermal evaporation deposition process used small crucibles as “point sources” to deposit thin films onto static or rotating substrates under high-vacuum conditions (typically in the order of 10-7 mbar). Such sources can be described as open evaporation, i.e., the molecules in the gas phase can directly enter into the volume of the vacuum chamber. This process is not very efficient and cannot be scaled to larger substrate sizes due to a combination of reduced uniformity and a relatively low material deposition yield (typically in the range of just 5%). Therefore, improved deposition tools were developed based on linear sources in which the material is deposited across the entire width of substrates that are moved linearly along the source opening (Fig. 1). In this way, the material usage can be increased significantly and upscaling to large motherglass sizes becomes feasible. The typical substrate size used in AMOLED production is Gen 5.5 (1300 × 1500 mm); however, there are plans by Samsung Display and LG Display to ramp up Gen 8 lines for TV production in the near future. With state-of-the-art source systems, a material utilization of 15% can be achieved.
Fig. 1: At left, a simplified linear-source deposition process is depicted and at right, a point-source evaporation process. For the linear-source deposition concept, either the substrate (option 1) or the source (option 2) can be moved.
While addressing utilization and scalability challenges, modern linear sources introduce additional challenges related to the increased thermal stress imposed on the organic material. Linear sources are usually semi-open systems, which means that inside the source the pressure is greater than in the surrounding tool chamber, and the evaporated particles are distributed evenly throughout the inner volume of the source by intermolecular collisions.
Furthermore, because linear-source architectures require increased partial pressures inside the source for distributing the material homogeneously, the nozzles of these sources have to be set to a temperature well above the evaporation point in order to prevent clogging. Thus, the material reservoir is exposed to a permanent thermal stress for a duration of several days or even weeks. Figure 2 shows typical linear nozzle source equipment.
Fig. 2: This linear-nozzle-source deposition equipment is from Sunic Systems in Korea.
An important metric for the processing quality of linear-source systems is the homogeneity of the deposited layer across the substrate, which typically needs to be in the range of ± – 5%. Another important requirement is to maintain stable deposition behavior throughout the entire processing cycle, i.e., the layers have to be of the same morphology on the first and the last day of operation.
Thermal Material Requirements
Due to the adoption of linear-source architectures, the materials used in OLED manufacturing require a chemical development focus toward achieving a larger thermal stability gap. This is defined as the difference between the evaporation temperature and the thermal decomposition temperature of the considered material. In order to determine this gap, developers usually check the evaporation temperature at the onset of processes in an open evaporation source.
In order to determine the thermal-stability window of materials, different measurement approaches can be used. A fast screening method is thermal gravimetric analysis (TGA), in which the material is heated in a controlled way while the sample weight is monitored. Most frequently, this measurement is conducted in an inert-gas stream. TGA can also be carried out in high-vacuum conditions, but such instruments are less available and rather costly.
The point where the material loses 0.5 % of the original mass is defined as the decomposition temperature. This procedure has to be exercised with care since weight loss could be caused by volatile impurities, such as solvents, without the actual decomposition of the main compound. Since virtually all OLED materials for vapor deposition are purified by gradient sublimation, it is typically expected that volatile constituents have already been removed during the purification process. The presence of volatiles in the final quality of the compound indicates issues with the purification or with the stability of the material in the sublimation process.
Another source for weight loss during TGA could be, of course, sublimation itself. However, due to the low vapor pressure of most OLED materials, decomposition of the compound is typically observed before sublimation at ambient pressure. By combining TGA with DSC (differential scanning calorimetry), it is relatively straightforward to identify the physical origin of the observed mass loss and to properly identify the decomposition temperature.
If the decomposition temperature and the evaporation temperature are close, this is a strong indication for a limited thermal stability of the compound. TGA experiments are usually conducted within 1 hour or less. TGA is, therefore, particularly useful as a screening method in the early development stage of materials. However, TGA just measures fast degradation processes and does not give much information about thermal stability over an extended time period. Hence, organic compounds that can successfully pass TGA tests still require further stability investigation.
It is tempting to use TGA in a more sophisticated way to assess long-term thermal stability. One approach would be to record the TGA data for various heating rates and fit the resulting spectra to kinetics models of the decomposition process. Except for the simplest cases, the authors found the precision of this procedure insufficient to predict the stability of the material in a real condition. The main obstacles are extrapolating over more than one order of magnitude in time, and the difficulty in identifying appropriate kinetic models matched to complex decomposition processes. Consequently, a better approach is needed.
We are currently conducting long-term thermal-stability measurements in quartz ampoule tests. The materials are placed in closed evacuated quartz ampoules at a range of temperatures (e.g., Tevap, Tevap + 25 K, Tevap + 50 K, Tevap + 75 K) for a specific period of time. After the test is complete, the materials are visually inspected and chemically analyzed in order to determine the effect of the thermal stress on the materials. In addition, actual OLED devices are built with these new materials and fully tested, which provides direct verification of the material stability at a given temperature. By using this test method, it is possible to determine the thermal-stability window for a specific material.
Due to their semi-open nature, linear sources require organic materials with a thermal stability at temperatures 20 K and more above Tevap that are maintained for several days. Only in this case can it be ensured that the OLED device performance is stable even for devices fabricated after several days of consecutive processing from a given material batch.
An Electron-Transporting Material
In order to better describe these trends, the authors examine a particular example in more detail, comparing the Novaled electron-transporting material (ETL) NET-164 to a state-of-the-art ETL used in display manufacturing. Table 1 shows calorimetric data of the two materials together with the evaporation temperature measured in an open-source vacuum-deposition tool.
Table 1: Calorimetric data for two materials (NET-164 and SoA ETL) measured in an open-source vacuum-deposition tool include Tg (glass-transition temperature), Tevap (evaporation temperature), Tm (melting temperature), TGA 0.5% (temperature of 0.5% weight loss in TGA), and TGA 5% (temperature of 5% weight loss).
When comparing these two materials with TGA, the authors found that both show a sufficiently large gap between the evaporation temperature, and also a 0.5% weight loss decomposition (ΔT for NET-164 = 168 K, ΔT for SoA ETL = 132 K).
In this stability test, it can be observed that the SoA material already fails at a temperature of Tevap + 75 K, whereas NET-164 still shows good performance at Tevap + 100 K.
Table 2 shows the data generated from the two materials sealed in quartz ampoules. This table shows ampoule temperatures at which the materials were tested. The green (pass) or red (non-pass) color indicates whether the material was still indicating good quality after the test. These tests were conducted for 10 days at the temperatures indicated in the table.
Table 2: Data generated from the same two materials sealed in quartz ampoules show SoA ETL failing at a temperature of Tevap +75 K.
Figure 3 shows the pictures of two ampoules with the SoA ETL measured at 318°C (left) and 368°C (right). One can clearly see the brownish coloring of the sample exposed to the higher thermal stress. This goes along with a strongly reduced analytical quality (e.g., HPLC purity) and deteriorated OLED device performance.
Fig. 3: The ampoule on the left shows the SoA ETL material measured after exposure at 318°C and the one at right, the same material after exposure at 368°C.
For material developers, it therefore becomes increasingly important to improve the thermal stability of the materials. This can be in conflict with other optimization parameters, such as lifetime or efficiency, but, in particular, with the glass-transition temperature of the organic compounds. Almost all OLED materials form organic glasses with amorphous layers, which prevent crystallization and give rise to favorable morphologies of the organic semiconductor layers. The temperature at which these organic glasses soften determines the glass-transition temperature. This softening can lead to a breakdown of the OLED devices, e.g., via mixing of adjacent layers or due to other inter-diffusion phenomena in the device stack. For that reason, device manufacturers request materials with rather high glass-transition temperatures, which typically translates to higher evaporation temperatures. These higher evaporation temperatures finally tend to narrow down the thermal-stability window of compounds, which means, in other words, that the requirements for high thermal stability as well as for high glass-transition temperatures in many cases oppose each other.
This basic trend can also be observed when the two ETL examples, NET-164 and the SoA material, are compared. Here, NET-164 has a glass-transition temperature that is 61 K below that of the SoA material. A similar difference is observed in the evaporation temperature, which is 66 K lower for NET-164. However, it is found that NET-164 is not only more stable than the reference material relative to the evaporation temperature, but also exhibits long-term stability when compared with the SoA material at the same temperature. This indicates that a selection of robust chemical structures can help reduce the tradeoff between Tg and stability requirements.
Inorganic materials are sometimes proposed for use in OLED devices. Indeed, such materials exhibit very high thermal stability and usually do not decompose during physical vapor deposition. However, they raise other issues such as limited environmental stability, safety concerns, difficult thermal management due to very high evaporation temperature, and difficulty in measuring evaporation rate or formation of dust. Some salt-type materials exhibit decomposition by forming neutral products before volatilization, which is usually accompanied by the formation of gasses and volatiles that are difficult to control.
Other Material Requirements
In addition to the thermal properties and stability of the organic materials, many other parameters need to be optimized, such as the energy levels of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO), the charge carrier mobility for holes or electrons, the quantum efficiency for emitting materials, etc.
In many cases, the optimization of these parameters depends strongly on the environment in which a material is used. For example, the requirements for an ETL material in a blue monochrome device (such as an RGB display structure) might be completely different than the requirements in a tandem white OLED. A tandem structure is a more complex device structure in which two mono-chrome OLED units (e.g., yellow and blue) are stacked vertically to provide, in combination, white light. This structure has merits in light output and durability because it creates double the amount of photons per unit area. Furthermore, the creation of the primary colors from white OLED using color filters has advantages over using red, green, and blue OLED pixels if high pixel pitch and patterning on a large area are required.
It is well-known1 that blue OLEDs exhibit a tradeoff between lifetime on the one hand and operating voltage on the other. In order to maintain a certain operational stability in the device, as required by the target application, a compromise in operating voltage has to be accepted. The yellow unit of a tandem white structure does not exhibit the same interplay. Consequently, an ETL with a lower operating voltage can be employed. Another illustration of this principle is the case of solution-processed OLEDs using vapor-deposited ETL to optimize performance (hybrid structure). Here, the proper adjustment of LUMO level to the polymer EML and the morphological state of the interface between polymer EML and small-molecule ETL layers is essential.2
Figure 4 provides an energy-level diagram of a typical OLED device. It depicts the highest occupied molecular orbitals (HOMO) and lowest unoccupied molecular orbitals (LUMO) of the organic materials used in the layer sequence. The two orbitals can be likened to the valence band and conduction band, respectively, of inorganic semiconductors. In contrast to inorganic semiconductors, the electronic states of holes and electrons are localized on individual molecules and charge transport takes place by hopping between individual sites. In the figure, the light-emitting layer (EML) is in contact with the hole-transporting layers (left side) and electron-transport layers (right side), facilitating charge injection and transport from the electrodes to the EML. In comparison to EML materials for vapor deposition, polymer emitting materials frequently have a very high lying LUMO level; i.e., injection of electrons into the EML is more difficult. In consequence, LUMO levels of hole-transport layer (HBL) and electron-transport layer (ETL) have to be adjusted to reduce injection barriers. n-type doping of the ETL is required for efficient charge injection into the organic layers. In consequence, the development of functional parameters of new materials should be done for each specific application – there are too many variables to make one universal material.
Fig. 4: An energy-level diagram of a typical OLED device depicts the highest occupied molecular orbitals (HOMO) and lowest unoccupied molecular orbitals (LUMO) of the organic materials used in the layer sequence.
Future Developments
The current manufacturing trends for large-area glass substrates may already be pushing some materials to their thermal limits. However, the increasing price pressure of the markets demand that tact times for the tools go down further, which reduces processing times for the individual layers and therefore requires even higher deposition rates. There is certainly room for further optimization of the thermal durability of individual materials, but at a certain point, the limits of organic chemistry will prevent even higher processing temperatures. For these reasons, tool manufacturers are investigating new deposition source concepts that apply less thermal stress on the materials.
In addition, manufacturers are interested in increasing tool up-times by enhancing the production time between necessary tool maintenance. This enhances the material amounts that are needed during the production cycle, and also increases the duration of the thermal stress on the materials.
One additional challenge in this respect is the need to distribute the heat equally in the sources to avoid any hot-spot formation, which could cause localized material degradation (e.g., at nozzles, tubes, etc.). For this reason, and due to the limited filling volume of the current concepts, most tool manufacturers are developing feeder mechanisms to allow an upgrading of the material loads. This can be realized, for example, by means of revolving material compartments, which can be evaporated one after another.
Another attractive concept is the use of feeder sources, in which the material powder is loaded into the evaporation zone of the sources via spiral conveyors. However, thus far these implementations suffer from clogging or jamming of the material powders. A potential future requirement for materials might be compatibility with such feeder-source approaches; however, this approach has not been realized and the specific material properties that might be required are not fully clear at the moment. One way to approach this topic might be pelletization of the materials, which is mostly only possible through the mixing of the active compounds with caking agents. If and how this can be realized in the future remains to be seen. Other potential future material evaporation concepts might use carrier gases (e.g., already followed by the so-called OVPD process – organic vapor-phase deposition). Here, also, detailed material requirements cannot be deduced at the moment, but such approaches often require heating of tubes and shower heads to temperatures well above the evaporation temperature of the compounds, thus increasing the thermal stress on the materials.
One further interesting development is the emerging technology of printable OLEDs based on solution processing, such as ink-jet printing, spin-coating, or slot-die coating. There are very fundamental differences in the deposition processes as well as in the material requirements between evaporation and solution processing. Whereas one process is based on rather small molecules that can exist in the gas phase, the other is usually based on polymeric or oligomeric compounds that are optimized for solutions in organic solvents.
The potential move away from evaporation toward solution-processing sets very different challenges for material developers. These, however, go beyond the scope of this article.
In the meantime, a thorough understanding of the material properties of the organic chemical compounds used for OLED deposition will help developers and manufacturers refine their processes for today and tomorrow’s display requirements. A great deal of research has gone into fine-tuning these processes for industrial applications, and that research is ongoing.
References
1C. Rothe et al., “Improved Efficiency and Lifetime by Tailoring the Charge-Carrier Supply in OLEDs,” SID Symp. Digest Tech. Papers 40, 495 (2009).
2U. Denker et al., “Hybrid Polymer OLEDs with Doped Small-Molecule Electron-Transport Layers for Display Applications,” SID Symp. Digest Tech. Papers 42, 935 (2011). •