Menu
In the fast-paced world of semiconductor manufacturing, the demand for precision and perfection has never been higher. At the heart of this industry’s evolution is the atomic layer deposition (ALD) technique, a critical process for achieving conformal coatings essential for the latest semiconductor technologies. While traditional thermal-ALD has its merits, it’s often limited by high deposition temperatures requirement and challenges in achieving all desired materials making plasma enhanced ALD (PE-ALD) an alternative option. PE-ALD, the game-changing alternative utilizes plasma as one of the reactants and consists of different species such as ions, radicals, electrons or photons. This innovative approach not only allows for low-temperature deposition but also enhances film quality, making it an increasingly preferred choice for industry professionals.
Even with its advantages, the chance of radical recombination in PE-ALD can negatively impact the conformality of the thin film growth deposited by this method. This issue becomes particularly significant when the targeted substrates are high aspect ratios (HAR). In such cases, there’s an increased likelihood of collisions among the incoming atoms or between the atoms and the walls of the structure. These collisions often result in atoms recombining and being lost, affecting the quality of the deposition.
Surface recombination of the atoms reduces the number of atoms available for the targeted reaction and limits the film penetration. Although the likelihood of recombination was understood before, measuring it directly was not possible. Previously, complex and indirect methods were used to estimate this probability. However, with the introduction of PillarHall chips, it is now possible to measure recombination probability directly and experimentally from the conformality measurements.
Prof. Erwin Kessels and his team at Eindhoven University of Technology have demonstrated that PillarHall chips can be used not only to identify the process conformality but also to accurately calculate the recombination probability.
With the help of modelling, it was possible to calculate how this recombination probability affects the movement of radicals inside the trenches. By 2019, they had successfully quantified the recombination probability for oxygen atoms in the PE-ALD process for materials such as SiO2, TiO2, HfO2, and Al2O3 [1].
Through their innovative method, Prof. Erwin Kessels and his team clearly proved the effect of recombination probability on how deeply a film can grow conformally inside a cavity.
They found that the chance of recombination, measured through experiments and calculations, varied significantly across different materials: SiO2 = (6 ± 2) x 10-5, TiO2 = (7 ± 4) x 10-5, Al2O3 = (1 -10) x 10-3 and HfO2 = (0.1 -10) x 10-2. This variation shows that the recombination probability highly depends on the material.
Schematic illustration showing the impact of recombination probability on the penetration depth.
Additionally, the experiments with PillarHall chips easily identified that the recombination probability changes with deposition temperature and plasma pressure. The impact of recombination probability was directly visible in the achieved penetration depth for these processes: as recombination probability increased, the penetration depth decreased.
Both SiO2 and TiO2 showed exceptional penetration depths, reaching an aspect ratio of approximately 900, while Al2O3 and HfO2 exhibited considerably lower penetration depths of 80 and 40, respectively due to their higher recombination probability.
By further experimenting with PillarHall chips and adjusting plasma exposure times, the researchers showed that significantly increasing plasma time was necessary to marginally increase the penetration depth. This highlights the PillarHall chip’s role in optimizing the reaction conditions for a better film coverage and preferred applications.
Atmospheric spatial atomic layer deposition (s-ALD) has become a standout member of the ALD family, thanks to its benefits such as lower costs and faster growth rates compared to the traditional thermal ALD or PE-ALD.
The researchers from Eindhoven University used PillarHall chips to confirm that s-ALD excels at creating high-quality films quickly and conformally [2]. They showed that s-ALD could achieve this even with a lower dose of plasma. Notably, using a plasma exposure time of just 0.73 seconds, they successfully formed conformal films of SiO2 and TiO2 in trenches with aspect ratios of 74 and 219, respectively. This difference in the penetration depth could be linked to the varying recombination probability of oxygen radicals at atmospheric pressure, which were calculated to be 4 x 10-4 for SiO2 and 6 x 10-5 for TiO2.
Apart from calculating recombination probability, PillarHall structures have also been utilized for studying ion surface interaction during PE-ALD process. This type of analysis would not have been feasible without PillarHall structures. During PE-ALD processes, ions are found only in the opening area of the PillarHall chips and not inside the cavities. Experiments using PillarHall have shown that while the presence of ions tends to reduce the growth per cycle (GPC), it actually improves the film’s quality in such a way that the wet etch rate was decreased.
Moreover, the experiments revealed that ions affect the crystallinity of the film. It was observed that films were crystalline in the opening area where there are ions, whereas the film stayed amorphous in cavities where no ions were present.