AVS2017 Session TF+SE-ThM: Control, Characterization, and Modeling of Thin Films I

To achieve the goal of rational design of next generation of advanced materials, we have been working on developing atomic layer deposition (ALD) as a promising method to tailor size and composition of nanostructured materials for a wide range of applications. Benefiting from self-limiting surface reactions, ALD enables conformal coatings of materials on three-dimensional substrates. The atomic level control of depositions makes it attractive to precisely synthesize the size and composition of nanomaterials. The size and composition of nanomaterials play important roles in achieving high performance in many applications. In this talk, we will discuss nanomaterials synthesized using Pd ALD and their applications in catalysis and energy storage. We will also illustrate that in situ characterization techniques such as synchrotron X-ray based X-ray absorption spectroscopy (XAS), X-ray pair distribution function (PDF), and FT-IR significantly advance our understandings of ALD in terms of surface chemistry and surface dynamics.

Complex oxide heterostructures possess a wide range of functional electronic and magnetic properties including metal-insulator transitions, superconductivity, ferroelectricity and colossal magnetoresistance effects. At epitaxial interfaces formed between atomically thin complex oxide films, electronic, chemical and structural interactions can be used to effectively tune the physical properties of these materials. Using a combination of atomic-scale controlled thin film synthesis and high resolution synchrotron diffraction based imaging, we show that structural distortions at the interfaces between polar La_1-xSr_xMnO₃ films and non-polar substrates can be effectively tuned by chemical modifications at these interfaces to control ferromagnetism in [001]-oriented La_1-xSr_xMnO₃ films with thickness less than 1 nm. We show that atomic-scale chemical control at polar/non polar oxide interfaces provides a powerful route to engineer novel electronic and magnetic phenomena at complex oxide interfaces.

Organoborons with short alkyl groups; trimethylboron (TMB), B(CH₃)₃, triethylboron (TEB), B(C₂H₅)₃, and tributylboron (TBB), B(C₄H₉)₃, were suggested as alternative, highly reactive, less-poisonous, non-explosive B-precursors in the mid 1990’s¹.TEB was found to exhibit the best properties for CVD of boron-carbon films, making it a popular CVD precursor². TMB and TBB were deemed not suitable as CVD precursors as no boron was found in the films deposited from these molecules. Consequently, these molecules are less investigated in CVD.

We study the gas phase chemistry of TMB in a thermal CVD process, using a combination of B-C film deposition experiments at several temperatures in both hydrogen and argon atmosphere and quantum chemical calculations for a wide range of possible gas phase reactions. We suggest that without assistance from the carrier gas, i.e. in argon ambient, TMB most likely decomposes by α-H elimination of CH₄ to form H₂CBCH₃. Methane is not highly reactive in CVD at deposition temperatures below 1000 °C, meaning that the H₂CBCH₃ species is the major film forming species. This correlates well with the B/C ratio of about 0.5 observed for films deposited in Ar at 700-900 °C. At higher temperatures, the B/C ratio of films increases as attributed to further decomposition to H₂BCH₃.

With assistance from the hydrogen carrier gas, TMB can also decompose to HB(CH₃)₂ that can further decompose to H₂BCH₃ and finally to BH₃, all with negative Gibbs free energies, albeit with some high energy barriers. This in combination with the unimolecular α-H elimination with a somewhat lower energy barrier, can explain the higher B/C ratios of films deposited in H₂.

Furthermore, we note that the onset of film deposition from TMB is 700 °C and at then at a very low deposition rate. Interestingly the film thickness does not increase with longer deposition time at 700 °C, indicating a surface poisoning effect. As this is seen both in Ar and H₂, we speculate that this is caused by CH₄ or H₂CBCH₃, which is currently the subject of our further investigations.

¹J. S. Lewis et al. Chemical vapor deposition of boron-carbon films using organometallic reagents. Mater. Lett. 1996, 27, 327.

²M. Imam et al. Gas phase chemical vapor deposition chemistry of triethylboron probed by boron–carbon thin film deposition and quantum chemical calculations. J. Mater. Chem. C 2015, 3, 10898.

Supported noble metal nanoparticles (NPs) are widely used in heterogeneous catalysis because of their high resistance against chemical poisoning. Atomic Layer Deposition (ALD) can be used to synthesize noble metal NPs on different high surface area supports, and offers sub-monolayer control over the metal loading (atoms per cm² of support) [1]. However, an improved understanding of how the deposition parameters influence the formation and growth of noble metal NPs is required to fully exploit the tuning potential of ALD.

We developed a synchrotron-compatible high-vacuum setup that enables in-situ monitoring during ALD [2]. Using this setup and focusing on ALD of Pt with the MeCpPtMe₃ precursor at 300 °C [3], we present an in-situ investigation of Pt NP growth on planar SiO₂ substrates by means of X-ray fluorescence (XRF) and grazing incidence small-angle X-ray scattering (GISAXS). The surface density of Pt atoms was determined by XRF. Analysis of the GISAXS patterns [4] yielded dynamic information on average real space parameters such as Pt cluster shape, size and spacing. The results indicate a diffusion-mediated particle growth regime for the standard O₂-based Pt ALD process, marked by a decreasing average areal density and formation of laterally elongated Pt clusters. Growth of the Pt NPs is thus not only governed by the adsorption of Pt precursor molecules from the gas-phase and subsequent combustion of the ligands, but is largely determined by adsorption of migrating Pt species on the surface and diffusion-driven particle coalescence [5].

Next, we have studied the influence of the reactant type (O₂ gas, O₂ plasma, N₂ plasma, NH₃ plasma [6]) on the Pt NP growth. Surprisingly, a clear difference in island growth behavior was found for the oxygen- vs. nitrogen-based processes. The latter processes were marked by a constant average particle distance during the growth process. Particle dimension analysis furthermore revealed vertically elongated NPs for the N₂ and NH₃ plasma-based Pt ALD processes. Therefore, it is concluded that atom and cluster surface diffusion phenomena are suppressed during the nitrogen-based processes. Finally, this insight provided the ground for the development of a tuning strategy that is based on combining the O₂-based and N₂ plasma-based ALD processes and offers independent control over NP size and coverage.

[1] Lu et al., Surf. Sci. Rep. 71 (2016) 410. [2] Dendooven et al., Rev. Sci. Instrum. 87 (2016) 113905. [3] Aaltonen et al., Chem. Mater. 15 (2003) 1924. [4] Schwartzkopf et al., Nanoscale 5 (2013) 5053. [5] Mackus et al., Chem. Mater. 25 (2013) 1905. [6] Longrie et al., ECS J. Solid State Sci. Technol. 1 (2012) Q123.

Scandium oxide (Sc₂O₃) thin films have been thoroughly studied for their use in microelectronic devices.^1–2 However, processes for the atomic layer deposition (ALD) of Sc₂O₃ films are scarce, and have mostly involved Sc(thd)₃,¹ ScCp₃,¹ Sc(MeCp)₃,² and Sc(amd)₃³ precursors. To date, the only mechanistic investigation has focused on the Sc(MeCp)₃/H₂O process using in-situ time-resolved quadrupole mass spectrometry to probe the Sc₂O₃ ALD chemistry.²

Herein, we have explored the Sc₂O₃ ALD using bis(methylcyclopentadienyl)3,5-dimethyl pyrazolatoscandium (Sc(MeCp)₂(Me₂pz)) with ozone and with D₂O. This precursor reacts with hydroxyl-terminated silicon, Si(111)– SiO₂–OH, at 150 °C and appears to remain thermally stable to 450 °C. Between 225 and 275 °C, there is a clear ligand exchange with ozone observed in the differential IR absorption spectra involving the formation of intermediate formate and carbonate species (1400–1600 cm^-1) after each ozone pulse. A short incubation period (≤ 5 ALD cycles) is observed at 225 °C prior to the onset of steady-state ligand exchange. The signature for the formation of Si–O–Sc bonds (1240 cm^–1) is clearly present after cycles 1–2 for the ozone process at 275 °C. The Sc₂O₃ growth is quantified by X-ray photoelectron spectroscopy (XPS) and by spectroscopic ellipsometry (SE), from which a growth rate of ~0.3–0.9 Å/cycle is extracted over the 225–275 °C temperature range.

In contrast, there is no ligand exchange observed for the D₂O process within the same temperature range, although some deposition occurs. The deposition rate for the D₂O process calculated by XPS and SE, is ~1.3 Å/cycle within the 225–275 °C window, which is higher than the non-uniform growth rate measured for the ozone process within that temperature range. The higher growth rate and lack of ligand exchange observed with D₂O is tentatively attributed to a CVD component that dominates the film growth process.

1. Putkonen, M.; Nieminen, M.; Niinist ö, J.; Niinist ö, L.; Sajavaara,T. Chem. Mater. 2001, 13, 4701−4707.

2. Han, J. H.; Nyns, L.; Delabie, A.; Franquet, A.; Van Elshocht, S.; Adelmann, C. Chem. Mater. 2014, 26, 1404−1412.

3. de Rouffignac, P.; Yousef, A. P.; Kim, K. H.; Gordon, R. G. Electrochem. Solid-State Lett. 2006, 9, F45–F48.

Wouldn’t it be a considerable gain of time to be able to check the stoichiometry of just deposited thin films in a few minutes? A recently commercially available sputter-based technique called plasma profiling time-of-flight mass spectrometry (PP-TOFMS) is capable to produce, in a few minutes, nm-scale depth resolved profiles of all elements (including light elements) of the periodic table, over a wide dynamic range (from 100% down to ppm)[1]. A simple ratio of the amount of ions detected from a given layer provides a calibration free semi-quantification.

For such fast feedback purposes a PP-TOFMS instrument (Horiba Scientific, Horiba FRANCE SAS, France) has been installed in the clean room of the CEA-LETI in close proximity to process tools.

In this paper we will present data obtained from microelectronics and nanotechnology thin films to demonstrate the performance of the technique. It will be shown that PP-TOFMS can be used for determining composition, detecting contamination, measuring doping level, and characterizing diffusion mechanisms.

For example we will show the ease of detecting, identifying, and locating in depth the presence of unexpected contamination in magnetic Iron Cobalt Boron multi-layers. Another example will show the depth profile of a Germanium Antimony Tellurium alloy deposited on silicon oxide used for phase change memories, a type of non-volatile random access memory. PP-TOFMS depth profiles agree with TOF-SIMS and STEM-EDX analyses for both the first nanometers and the in-depth composition.

[1] A. Tempez et al., J. Vac. Sci. Technol. B (2016) 34

[1] Weiss, W. and W. Ranke, Surface chemistry and catalysis on well-defined epitaxial iron-oxide layers. Progress in Surface Science, 2002. 70(1-3): p. 1-151.

[2]Parkinson, G.S., Iron oxide surfaces. Surface Science Reports, 2016. 71(1): p. 272-365.