AVS 66 Session TF-FrM: Theory and Characterization of Thin Film Properties

The management of unwanted impurities as well as the controlled introduction of dopant species continue to be challenges in wide-band-gap (WBG) semiconductors such as AlN and Ga₂O₃. In these materials, contaminants such as carbon often incorporate during growth and subsequently act as trapping centers, which can have a detrimental impact on device performance. Moreover, establishing hole conductivity through the introduction of acceptor impurities has proven to be especially difficult in these compounds. In this talk, I will discuss the use of first-principles calculations to understand the incorporation and electronic properties of impurities in WBG semiconductors.

In many growth techniques carbon-containing precursors, such as trimethylaluminum (TMA) for AlN, are often employed. These precursors are thought to be a major source of carbon contamination during growth. Focusing on atomic-layer deposition, we have developed a model to elucidate the decomposition of TMA at the AlN surface, and subsequent incorporation of carbon into the AlN film during growth. We find that the use of H-containing plasma is crucial for scrubbing methyl species from the surface. However, the plasma also leads to atomic carbon, which opens a channel for trapping carbon impurities into the film. In light of this dual role, we propose a solution for minimizing carbon contamination into AlN.

Feature-scale simulations (FSS) represent an important tool for modeling of etching, deposition, and implantation processes routinely applied during semiconductor materials processing and device fabrication. Traditionally, FSS uses a combination of Monte Carlo methods with special algorithms for modeling gaseous and solid (in volume and at the surface) species, as well as algorithms for advancing feature profiles and materials composition as a result of reactive interactions of incoming species with solid materials. The mentioned FSS approach, at least at present time, seems to be the only feasible approach to simulate materials processing at the feature-scale level ranging from tens of nanometers to tens of micrometers, while applications of such advanced methods as quantum chemistry or molecular dynamics are still too expensive computationally in spite of highly efficient code parallelization to run on multi-processor computers. Phenomenological description of reactions (chemical and physical) represents a weakness of FSS, as it is very difficult to implement proper sets of reactions valid at different conditions. Here, we discuss an approach used in the FPS3D code [1-5], for simulation of etching, deposition, and implantation with the present focus given to simulation of ALD/CVD processes. Examples include such difficult topics as HAR etching, as well as CVD and ALD. For ALD examples we consider silicon nitride deposition [5] for different chemistries, some of them allowing deposition of a single monolayer per cycle, while others allowing only a fraction of a monolayer to be deposited per cycle.

References:

[1] P. Moroz, IEEE Trans. on Plasma Science, 39 2804 (2011).

[2] P. Moroz, D. J. Moroz, ECS Transactions, 50 61 (2013).

[3] P. Moroz, D. J. Moroz, J. Physics: CS 550 012030 (2014).

[4] P. Moroz, 15^th Int. Conf. on Atomic Layer Deposition, Portland, OR (2015).

Solution-deposited palladium on amine-terminated self-assembled monolayers (SAMs) is a well-characterized catalyst and adhesion layer combination for solution-based electroless metallization of dielectric films. A reducing agent is typically added to the deposition bath or a sensitizer such as tin is co-deposited producing relatively thick Pd layers. Thinner Pd deposits would enable barrier seed layers for filling < 10 nm wide gaps in patterned dielectric films with metal. In this work, we eliminated the reducing agent from the deposition bath and worked at pH < 2 to deposit monomeric Pd(2+) species and show that the amine groups terminating the SAM reduce Pd(2+) to Pd(0). The amount of Pd deposited depended on the coverage of the two types of amines on the SAM. The adsorption of PdCl₄^2-ions in solution on protonated amine groups (–NH₃⁺) is well known. Our data suggest that the nonprotonated amine groups (–NH₂), which coexist with –NH₃⁺, chemically reduce the Pd(2+) ion to Pd metal by oxidizing to the amine radical cation (–NH₂•⁺). Pd bonds to and covers the –NH₂•⁺groups in the process depositing around a monolayer of Pd from solution on the SAM-covered silicon oxide surface. The Pd layer served as a catalyst for solution deposition of cobalt films on the surface using a reducing agent under oxygen-free conditions. The cobalt deposited initially as islands that grew together into a closed film with good adhesion.

Dropwise condensation of steam on metallic surfaces coated with thin (~ 1µm) functional hydrophobic films has the potential to achieve remarkable heat transfer coefficients approaching ~100 kW/m²K. However, the long-term durability of these thin films has limited the application of functional coatings for the past century. Although degradation due to steam condensation has been qualitatively described as ‘blistering’, no satisfactory insight exists capable of answering two key questions: what is the mechanism of water vapor mass transfer, and what is the driving force for film delamination. Hence, scientists have been forced to abandon rational thin film development in favor of ad-hoc trial-and-error approaches. Here, we demonstrate that pinholes on hydrophobic coatings are the source of blisters so commonly seen during degradation. By creating shape-controlled pinhole-blister structures in thin deposited films, we show that blisters form in a spatially-controlled order during water vapor condensation from the ambient. The shape, initiation, and growth of the blisters was systematically investigated. Our experiments demonstrate that water vapor is mainly transferred to the blister through spatially-random pinholes which exist in the film after deposition, and the driving force for film delamination is capillary force. Based on the new insights developed here, we propose a non-dimensional pressure to determine the threshold when blistering will be initiated by a pinhole, or when discrete droplets grow up and above pinholes. To the best of our knowledge, our work represents the first quantitative description of blistering initiation. The techniques and insights presented here will inform future work on polymeric thin films to enable their rational and durable design for a variety of applications.

The need to control and manipulate magnetic spin in nonvolatile memory applications drives exploration of new magnetic materials with non-uniformly ordered magnetic phases. Of particular interest are materials with inversion asymmetry, most commonly found in non-centrosymmetric space groups. Our group has successfully grown epitaxial B20 FeGe films with 4% tensile strain on a Ge(111) substrate by Sputter Beam Epitaxy, an off-axis magnetron sputtering technique in which beam-shaping, shutter control, and QCM-guided flux control of off-axis, direct-current (DC) magnetron sputter sources are employed upon high-purity elemental Fe and Ge targets in ultra-high vacuum. QCM control is modified to relative atomic ratios, and film compositions are confirmed by energy dispersive x-ray spectroscopy (EDS) and Rutherford Backscattering (RBS). X-ray diffractometry has confirmed that the films are single-crystal and phase pure, with near-substrate-limited rocking curve (FWHM 0.07º) and strong Keissig fringes in x-ray reflectometry. We find no evidence of strain relaxation up to 110nm, and off-axis XRD and HAADF STEM confirm the B20 phase necessary for the Skyrmionic phase. A strain-enhanced T_C = 350K by SQUID magnetometry and a clear Topological Hall effect (THE) signature observed at 330K suggest potential for Skyrmionic behavior at or above room temperature in a single layer thin film. Direct observation is required for confirmation of a Skyrmion lattice phase.

The oscillator strength factor of the Si-N asymmetric stretch mode of SiN films was studied for films of varying composition and thickness. Thin films were deposited by low pressure chemical vapor deposition at 850°C from mixtures of dichlorosilane and ammonia. The oscillator strength factor for each film was determined from Fourier Transform infrared spectroscopy and ellipsometric measurements. We found that the oscillator strength factor systematically decreases with increasing silicon volume fraction from2.10x10 cm to 1.44x10 cm for compositions ranging from 0% to 25% volume fraction amorphous silicon. We believe this trend is related to charge transfer induced structural changes in the basal SiN_x tetrahedron as the volume fraction of amorphous silicon increases. For stoichiometric silicon nitride the oscillator strength factor was found to be 2.01x10¹⁹ ± 7.25x10¹⁷ cm^-2, which is consistent with a reported value of 2.07x10¹⁹ cm^-2 and a theoretical value of 1.99x10¹⁹ cm^-2 . Additionally, in the composition range investigated, we found that the oscillator strength values agree favorably with trends observed in films deposited by plasma enhanced chemical vapor deposition. This work was selected as the best graduate student presentation at the 2019 NMAVS symposium (Albuquerque-June2019).

Atomic Layer Deposition (ALD) is a vapor phase technique to deposit thin films of various metals and metal oxides on a substrate. Due to sequential and self-limiting reactions, conformal and pinhole-free thin films can be produced which have widespread applications. In this process, a precursor, which is often a metal surrounded by organic functional groups, chemisorbs on the substrate and part of the molecule subsequently desorbs from the surface after completion of the reaction. Precursor chemisorption on the substrate leads to a self-limiting process and it eventually results in films with desired thickness at the Ångström length scale. To design and conduct an ALD experiment, the precursor(s) should be chosen based on the ALD conditions (i.e., bubbler and reactor temperatures, pressure, gas flow, etc.) and likely applications of the final film. It is practically impossible to carry out a huge number of ALD experiments using numerous precursors and deposition conditions in order to find the optimum one depending on the applications of interest. In addition, only existing precursors can be tested experimentally. This study focuses on developing a computational tool for the design of novel precursor materials with enhanced properties for the ALD of metal oxides and metals.

Computer-Aided Molecular Design (CAMD) is a methodology where materials with optimal desired properties are generated from the combination of functional groups. This approach is the reverse of Group Contribution Method (GCM) in which the thermodynamic properties of a compound are estimated from the structural and functional groups comprising the molecule. For CAMD, we need the properties of the functional groups. In our previous work, we have redeveloped a new GCM for ALD effectively to predict the growth rate curve using Adsorbate Solid Solution Theory (ASST). In this work, novel precursor molecules for ALD are generated using properties of the functional groups. In order to do that, we will be using a combinatorial optimization method called Efficient Ant Colony Optimization (EACO). This is the first time CAMD is being applied to design precursor materials for ALD. In the future, novel designed precursors will be synthesized and their properties will be tested experimentally using a Kurt J. Lesker ALD150LE^TM system. Characterization of the deposited films with designed precursors will validate the proposed simulation technique and help us to optimize materials in the best possible way.

The interest in Titania films has been increasing in the past few decades. This increase can be attributed to the various applications of ultrathin films of Titanium dioxide. The thin film deposition of Amorphous Titania can be performed using a low temperature precursor, Tetrakis(dimethylamido) Titanium (TDMAT) along with the use of ozone however with oxygen, comparable growth rates have only been reported at temperatures greater than 150 C. For use with organic substrates, it is important that the reaction is performed at low temperatures to prevent denaturation and degradation. The use of Ozone has also been reported to cause degradation in some polymers, mainly elastomers. Ozone cracking, along with the high cost of manufacturing a high concentration of ozone, does not make ozone the most suitable choice and could easily be replaced by using pure molecular oxygen. In this study, growth temperatures below 160 C were used to uniformly deposit amorphous Titania using TDMAT and oxygen. The as-deposited films were characterized using Spectroscopic Ellipsometry (SE), X-ray Photoelectron Spectroscopy (XPS). The application of this reaction could lead to low temperature deposition of Titanium oxide on organic substrates for the use in biomedical implants as a protective coating, seed layer to deposit other conductive metals on organic substrates. One of the most interesting phenomena observed using SE was the appearance of a very small but non-zero extinction coefficient in the visible range. This extinction coefficient was similar to Urbach tail absorption usually observed in the UV range. This also points towards slight absorption of Visible light by the deposited Titania films which opens new avenues for research in photocatalytic activity of Titania films within visible light spectrum.

Ultra-high purity (UHP) process capability is motivated by the need to produce superior, high-quality thin films and interfaces by atomic layer deposition (ALD) techniques. In particular, UHP equipment design reduces background impurity levels, including oxygen, to limit incorporation during film growth. Creating and maintaining a UHP process environment are also essential for ALD process reproducibility. This presentation will address the potential sources of background contamination, as well as system design requirements to obtain a controlled UHP process environment. Results will be presented that demonstrate the effectiveness of this technology to obtain high-quality titanium nitride thin films by plasma-enhanced ALD (PEALD) techniques.