ALD is a thin film growth method enabling device down-scaling beyond Moore’s law¹. Device down-scaling gives boundary conditions to thin film quality in a sense of film uniformity, conformality and homogeneity, which need to be characterized even a few nanometers thin. Shrinking device dimensions and film thicknesses set new challenges to process development and quality control as thin film analysis and characterization becomes challenging. PillarHall™ LHAR test structure has proven its value in ALD conformality characterization^2-4. High aspect ratio (HAR) geometry in PillarHall has characteristically infinitely wide trench with height of 500 nm and width of 10 mm. However, in real applications the geometry and pitch size vary. The question is how variable HAR geometries correlate with PillarHall in conformality and ALD growth analyses? In this study, we used nanolaminate ALD process and set of variable high aspect ratio test structure geometries as substrates. Thermal ALD Al₂O₃-TiO₂ nanolaminate growth was studied at three different temperatures, at 110, 200 and 300 °C. Targeted film thickness was 25 nm. For nanolaminate, sublayer thicknesses were targeted to 1 nm both for Al₂O₃ and TiO₂. For comparison results were compared with LHAR chips coated with reference ALD Al₂O₃ grown at 300 °C. Five different conformality test chip designs were coated in a one run and analyzed: PillarHall LHAR4 with three different channel heights namely 100, 500 and 2000 nm. A prototype PillarHall LHAR3 test chip with narrow lateral channels widths from 2 to 78 µm (so called FIN-design) and vertical HAR (VHAR) test structure of an array of vertical pores, 1 µm in diameter and 219 µm in depth. During the coating chips were place in center of the wafer holder to minimize the conformality variation at the wafer surface level. After the coating LHAR samples were analyzed by optical microscopy, spectroscopic reflectometry, and for VHAR samples electron microscopy cross-section was used. We show that with PillarHall, we can study in more detail the film growth in the channel, geometry dependence and study the process saturation in the channel. We also study of applicability of narrow channels of FIN-design, width from 2 to 78 µm, and compare the results with the Gordon model⁵.

REFERENCES

[1] International Roadmap for Devices and Systems (IRDS™) 2021 Edition IRDS™ 2021: More Moore

[2] Gao, F. et al., J VAC SCI TECHNOL A, 33 (2015)

[3] Ylilammi, M. et al., J APPL PHYS, 123 (2018)

[4] Yim and Ylivaara, et al., Phys. Chem. Chem. Phys. 22 (2020)

[5] Gordon, R. G., Chem. Vapor Depos., 9 (2003)

Antimony (Sb) in thin film form is considered as a nano-optical material which is of interest for potential applications, such as photonic computing or high-speed holographic displays. It has been used as lens material in super-resolution near-field structures and broadband saturable absorber among others. In addition to optoelectronic properties, Sb is widely used as a dopant in tuning the transport properties of many thermoelectric materials such as PbTe, Half-Heusler compounds and so on. In this study we propose new ALD formulation for low temperature deposition of pure Sb down to 70 °C making this process applicable on the surface modification of low and medium temperature thermoelectric materials such as Mg₂Sb₃, Bi₂Te₃. Pure and conformal thin films of elemental Sb were prepared by ALD using Sb(OEt)₃ and (Et₃Si)₃Sb as precursors with a growth per cycle (GPC) of 0.18 Å at deposition temperature of 70 °C. The electrical conductivity of the films grown on mica at 70 °C (1.13´10⁴ S/cm) is slightly higher than that of reported by Pore et al. (1.07´10⁴ S/cm). NbFeSb half-Heusler particles were successfully coated with ultra-thin layer of Sb and consolidated using high pressure sintering techniques. The results show addition of few atomic layer of Sb on NbFeSb particles can significantly enhance the electrical conductivity of the studied half-Heusler compound which was reduced during the high-pressure sintering.

The self-terminating nature of atomic layer deposition (ALD) enables the preparation of uniform coatings on high-surface-area materials.¹ Our general aim is to use ALD to prepare zinc promoted heterogeneous catalysts for carbon dioxide hydrogenation into methanol.² In this work, we aimed to obtain uniform coating of zinc on mesoporous monoclinic zirconia and gamma alumina particles in a fixed-bed flow-type ALD reactor.

Zinc acetylacetonate [Zn(acac)₂], a non-toxic reactant with a good stability in air, was used as an ALD reactant. Thermogravimetric analysis confirmed Zn(acac)₂ to be a potential reactant. Zn(acac)₂ was vaporized at 120 °C and carried by N₂ to the calcined support for 3 h. For zirconia, the effect of the calcination temperature (400 to 1000 °C, typically 600 °C), process temperature (160 to 240 °C, typically 200 °C), and number of cycles (1-3, 2^nd reactant: oxygen) on zinc loading were analyzed by inductively coupled plasma-optical emission spectrometry (ICP-OES). For zirconia and alumina, the amount of Zn(acac)₂ evaporated was varied. Zinc distribution within a solid support particle was analyzed by scanning electron microscopy-energy dispersive X-ray spectroscopy (SEM-EDS) (Fig. 1 and Fig. 2 of the supplementary information). A diffusion-reaction model^3,4 was used to simulate the saturation profile of zinc within theoretical pores (Fig. 3). Electronic structure of zinc was analyzed by X-ray absorption spectroscopy (XAS). The oxidation of acac ligands on the samples was observed by in-situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) combined with mass spectrometry (MS). The zinc content on zirconia saturated to ca. 1.3 wt-% (2.2 Zn nm^-2) with increasing amount of Zn(acac)₂. The zinc content decreased from ca. 2.2 to 0.3 wt-% with increasing calcination temperature, most likely resulting from the decreasing zirconia surface area. The zinc areal number density (ca. 2.2 Zn nm^-2) remained almost constant.

This work was supported by the Academy of Finland (COOLCAT project, grant no. 329978, ALDI project, grant no. 331082, and Matter and Materials, grant No. 318913) and by R. L. Puurunen’s starting grant at Aalto university. Hannu Revitzer and Ilkka Välimaa are thanked for the ICP-OES analysis.

References

1. R. van Ommen, A. Goulas, and R. L. Puurunen, Atomic layer deposition, in Kirk-Othmer Encyclopedia of Chemical Technology, John Wiley & Sons, Inc. (2021).
2. A. Arandia et al., manuscript in preparation (2022).
3. M. Ylilammi et al., J. Appl. Phys., 123, 205301(2018).
4. J. Yim and E. Verkama et al., submitted. Preprint DOI: 10.33774/chemrxiv-2021-2j4n1.

Group 13 metal oxides (e.g. Al₂O₃, Ga₂O₃ and In₂O₃) are utilized for a broad range of electronic applications such as semiconductors, optoelectronics and sensors. Besides, they represent essential components of heterogenous catalysts. Al₂O₃ is employed industrially as support material or catalyst itself for various reactions like epoxidation of alkenes and dehydration of alcohols [1]. Moreover, Ga₂O₃ has been intensively studied for the dehydrogenation of light alkanes and In₂O₃ catalysts gained a lot of attention due to the ability of converting CO₂ into methanol [2, 3].

In heterogeneous catalysis, reactions mostly occur at active sites on the material's surface. Therefore, high surface area and homogeneous dispersion of deposited interfaces (e.g. metal oxides) are decisive for enhanced activity [4]. Yet, standard synthesis techniques like impregnation and precipitation often lead to unwanted agglomeration. One well-established technique which overcomes uncontrolled coating is atomic layer deposition (ALD). ALD leads to uniform, nanoscale films on a materials surface and is therefore progressively attracting attention as potential synthesis-tool for catalysts [5]. However, most ALD investigations target flat substrates whereas for catalyst synthesis, powders with high surface area and porosity are demanded.

In order to enable the full potential of ALD-synthesized catalysts, the deposition behavior on porous powders has to be further elucidated. Previously, we demonstrated the growth behavior of AlO_x on mesoporous silica using the ALD process of trimethylaluminum and water [6]. In the current study, we approach other catalytically active oxides of group 13 metals like gallium (GaO_x) and indium (InO_x). Both oxides were deposited on high surface area silica powder via ALD and their growth behavior was studied intensively using in-situ thermogravimetry, XRD, STEM, FTIR and NMR. Furthermore, both processes were compared to the powder ALD process of AlO_x and reference samples prepared via incipient wetness impregnation.

References

[1]Schlögl, R.; Hess, C.; Nanostructured catalysts: Selective oxidations; Royal Society of Chemistry: Cambridge, 2010, ISBN 9781847559876.

[2]Martin, O.; Pérez-Ramírez, J. et al.; Angew. Chem. Int. Ed Engl.2016, 55, 6261–6265.

[3]Castro-Fernández, P.; Müller, C.R. et al.; ACS Catal.2021, 11, 907–924.

[4]Somorjai, G.A.; Li, Y.; Introduction to surface chemistry and catalysis, 2nd ed.; Wiley: Oxford, 2010, ISBN 978-0-470-50823-7.

[5]O’Neill, B.J.; Huber, G.W. et al.; ACS Catal.2015, 5, 1804–1825.

[6]Strempel, V.E.; Rosowski, F. et al.; Nanomaterials2018, 8.

In order to achieve the uniform coating in the ALD process, a sufficient gas exposure (exposure time and gas concentration) needs to be provided. This requirement becomes particularly relevant for highly porous and high aspect-ratio substrates, where the gas transport into the substrate structure is limited by diffusion (diffusion-limited regime), or for ALD precursor systems exhibiting low surface reaction rate (reaction-limited regime). This work (Szmyt et al., Chem. Mat. 2022, 34, 1, 203-216) reports how the distinction between diffusion- and reaction-limited ALD regimes is directly quantitatively related to the width of the reaction front and the profile of chemisorption coverage in a single cycle ALD, all of them being determined by the natural length unit of the system. We introduce a new parametrization of the system based on its natural system of units, dictated by the scales of the physical phenomena governing the process. We present a range of scaling laws valid for a general porous substrate, which scale intuitively with the natural units of the system. The scaling laws describe (i) the coating depth in a diffusion-limited regime with respect to the gas exposure, (ii) the chemisorption coverage in a reaction-limited regime with respect to the gas exposure, and (iii) the width of the reaction front in the diffusion-limited regime. For the first time, the distinction between diffusion- and reaction-limited ALD regimes is directly quantitatively related to the width of the reaction front and the profile of chemisorption coverage in a single cycle ALD. The model system for the multicycle diffusion-limited coating of random fibrous mats was validated with an experiment of ALD on a forest of tortuous carbon nanotubes (CNTs).