ALD/ALE 2022 Session AS-TuP: Area Selective ALD Poster Session

In our recent published work, we explore the possibility of achieving area selective ALD by applying temperature gradients on the substrate [1]. In this approach, the majority of the substrate is kept at a low temperature, which suppresses the surface chemical reaction, while a small area is heated to allow the reaction to locally proceed. Controlling the size and the position of the heating spot on the substrate allows for 'writing' on the surface, with potential applications in the bottom-up fabrication of electronic devices like solar panels and OLED displays. We study the feasibility and window of opportunity of this technique by computational modelling. We first model the control of the temperature by various illumination protocols, and then model and simulate the nucleation and growth of spatially localized spots, as well as lines, of deposited material given an inhomogeneous temperature profile. We show that the temperature profile can direct substrate deposition and control the connectivity and size of the pattern deposited on the substrate.

A practical example of this technique is the ALD process of Si₂H₆, where instead of a co-reactant the elevated temperature induced by the laser itself is used to remove the ligands [2]. For this process, it is reported that surface diffusion of deposited molecules also influences growth of the spot that is formed on the substrate. I will present new results from a model that includes this surface diffusion, and study its effect on the deposition on the substrate both at constant temperature and for the case where locally the substrate is heated by a laser. Our model suggest that the absorption rate goes up with increasing diffusivity at constant temperature.

[1] B. de Braaf, C. R. (2021). Modeling the initial monolayer formation in thermally localized surface deposition. J. Vac. Sci. Technol. B.

[2] Y. Suda, M. I. (1996). Ar+-laser-assisted subatomic-layer epitaxy of Si. Journal of Crystal Growth, 672-680.

Area-selective atomic layer deposition (AS-ALD) has been considered as a prominent technique due to the escalating demands for eliminating the edge placement errors with current top-down approaches in semiconductor processing at the sub-5 nm node.^1–3. Recently, it has been reported that anhydrous hydrazine (N₂H₄) can be employed as the reduction of the Cu oxide to metallic Cu surface.³ By employing the high reactivity of hydrazine, under the ALD environment, the metallic surface condition can be maintained, or the oxidation/reduction of the Cu surface can be repeated under the ALD environment. Eventually, area selective deposition of dielectric material (e.g., ALD-SiO_x on Si, SiN_x, TiN, AlO_x, substrates) can be achieved, whereas nucleation delay and limiting surface oxidation on Cu sample can occur.Additionally, a detailed change of Cu condition with precursor exposures will be studied using in-situ surface analysis.

In this study, the consecutive surface cleaning and AS-ALD of SiO₂ process was demonstrated. To identify the substrate dependence on ALD selectivity, Cu, Si, SiN_x, TiN, and AlO_x substrates were loaded in the ALD chamber at the same time. Prior to the ASD process, the samples were pretreated with N₂H₄ at 200 ℃. After that, the ABC-type ALD-SiO₂ was performed. In the Si precursor half-cycle, the tris(dimethylamino)silane (step A), was exposed for 0.2 s, followed by a precursor trapping time for 120 s and purging time of 180 s. In the oxygen reactant half-cycle, the O₃/O₂gas mixture (step B), was introduced for 0.2 s and captured for 120 seconds, followed by purging the chamber with a continuous flow of N₂ carrier gas for 180 seconds. After the ALD-SiO_x cycle, an additional of surface recovery step with N₂H₄ (step C) was introduced. With five supercycle ALD-SiO_x processes, growth of SiO₂ on both bare Si and SiN_xsubstrates, formation of metal-silicates (and/or SiO_x) on TiN_x and AlO_x suggest that the supercycle-based ALD-SiO_xprocess does not impact the growth of SiO_x on top of dielectric substrates. On the other hand, the deposited amount of SiO_x on Cu substrate is approximately 35% less than the AB-typed ALD-SiO_x process. Despite slight detection of SiO_x on Cu (non-growth) substrate, the feasibility of the ASD process with repeating surface oxidation and reduction was demonstrated. The detailed experimental results will be presented.

We thank Rasirc Inc. for funding this project and providing the Brute N₂H₄.

¹ P. Kapur, et al., IEEE Trans. Electron Devices,49, 590 (2002).

² M.F.J. Vos et al., J. Phys. Chem. C,122, 22519 (2018).

³ S.M. Hwang, et al., ECS Trans. 92, 265 (2019).

With continuous progress in the field of nanofabrication and nanotechnology, the semiconductor industry has greatly flourished. However, efforts for further reduction in feature sizes of electronic interconnects in search of better and fancier devices, are still ongoing. The struggle to search for better area selective deposition (ASD) processes has led researchers to manipulate deposition surfaces using different passivation tools. In this regard, surface inhibitors have gained a lot of attention. In this study, an organothiol inhibitor has been utilized for ASD on metal, oxide, and nitride surfaces, Cu, SiO₂, and TiN, respectively. The inhibitor selectively adsorbs on the Cu and SiO₂ surfaces at 400 ºC, while the TiN surface remains unaffected after exposure to the inhibitor. Upon high-temperature exposure, the organothiol inhibitor is capable of decomposing to assist the adsorption of its different parts on the Cu and SiO₂ substrates, thereby simultaneously inhibiting two surfaces through a single inhibitor. The inhibited substrates were examined for adsorption and inhibition using surface analysis tools including water contact angle (WCA) measurements, X-ray photoelectron spectroscopy (XPS), etc. Blocking results revealed promising blocking potential against HfO₂ ALD on Cu compared to SiO₂, whereas the TiN surface did not exhibit any blocking at all. Furthermore, the surface chemistry and reactivity have been explained by theoretical calculation using the Monte Carlo method and density functional theory.