ALD Processes – Highly Homogeneous and Decorative Coatings

An example of successful transfer of knowledge and technology

On some metals, it is possible to add transparent, colorful-looking coatings due to interference effects of added layers. However, electrical field effects occur along edges or in narrow recesses during production using the PECVD process, which leads to considerable color deviations in these layers due to changes in layer thickness. After looking into this problem, experts at the Steinbeis Transfer Center for Surface and Coating Technology and Furtwangen University (HFU) have discovered that performing coating using atomic layer deposition (ALD) not only avoids these complications, but also makes it possible to deposit layers of an exact thickness, for example with layers of titanium or aluminum oxide. Applied to smooth metallic substrates, this allows highly decorative and, due to their robustness, durable colored layers to be created. The process can be used with either individual or high-volume components.

This slideshow requires JavaScript.

The process of depositing thin layers in plasma using CVD processes – or PECVD for short, which stands for plasma-enhanced chemical vapor deposition – has been known for some time and is now used in a wide range of industrial applications. It is only possible to add layers uniformly to flat components, however. One phenomenon with coatings is the development of interference colors, in which layers of varying thickness manifest themselves in visual terms in rainbow colors, a characteristic not wanted in the majority of applications. Applied to small objects, such as those used in medical technology or watchmaking can often not be coated with PECVD true to contours.

One interesting alternative in this area is atomic layer deposition (ALD), a process that allows layer thicknesses to be adjusted highly precisely down to the sub-nanometer range. ALD makes it possible to coat components in complex formats and very different dimensions, such that coatings remain completely even and reproducible. This also makes it possible to achieve entirely homogeneous colors. Even with larger batches with a high packing density the coating of all parts is the same, resulting in virtually zero rejects. Since this process is more expensive than PECVD, from an economic standpoint it is the preferred option for small components, such as watch parts.

Optical interference effects

Many are familiar with interference colors from, among other things, soap bubbles drifting around in bright sunlight. The color effects on bubbles are created by certain wavelengths of light being reinforced or canceled out. This physical relationship of interference within thin films can be described as follows [1]: A uniform light wave arrives from a medium with an optical refractive index of n₁ on a thin layer of thickness d and a refractive index of n₂. One part of the wave is reflected directly at the first interface between the media (B’), while another part is reflected at the second (lower) interface (C). Inside medium 2, the light waves travel at a different velocity. The path difference of reflected beams 1″ and 2″ is g. According to Fermat’s principle, the transit time of light traveling from A to B is the same as light traveling from A’ to B’. In addition, if n₂ > n₁ there is a phase shift of π (half a period) at the interface between medium 1 and medium 2. When the phase difference = π (corresponding to a path difference of λ/2), the light with a wavelength of λ (colors) is canceled out.

The different layer thicknesses (d) therefore result in different wavelengths (colors) being left out of the reflected light. Light that was originally white, which comprises all colors, loses certain color components and appears in a residual color. As a result, objects with coatings acquire an interference color under illumination. Each layer thickness corresponds to a different color, with colors repeating periodically with growing layer thickness. This is because light is not only annihilated at a path difference of λ/2, but also at 3 λ/2, 5 λ/2 and other odd multiples of λ.

Inhomogeneous colors due to field effects

When PECVD processes are used to deposit transparent interference layers on metals, the result is almost always inhomogeneous coloring [2], however. This is because interference layers grow to a certain extent inhomogeneously on 3D parts. Even differences of a very few nanometers in layer thickness create the impression of different colors. The reason for these inhomogeneous layer thicknesses is that electrical fields are distributed differently on substrates. In particular, different thicknesses appear along edges or in recesses.

The inhomogeneous field distribution can be explained as follows. Metallic surfaces are electrically conductive and therefore electrically equipotential areas – i.e. all points on the surface have the same electrical potential. In the simple case of an electrically conductive and charged sphere, the surface charge is evenly distributed. Excess charges immediately flow to are electrically conductive point on the surface of the sphere, resulting in potential equalization (like a Faraday cage). The smaller the radius of a sphere, or a bulge or peak on a metallic surface that is approximated by a sphere, the higher the surface charge density and the field strength on this area of a component to be coated.

This simplified model, does not take into account the shielding effects of plasma or dark space. It does, however, offer a plausible explanation for increased deposition levels on corners and edges. In extreme cases, the field elevation lead to peak discharges: Plasma engineers are familiar with the dreaded phenomenon of electrical flashover, which can be caused by sharp edges and surface peaks. In case the surface geometries are not pointed, it is easier to increase the deposition rate of in layers by increasing the magnitude of the electric field. Consequently, field shielding (the Faraday cage effect) results in reduced deposits in recesses. Depending on the required application, this inhomogeneous rate of deposition can be used to create effects, for example in the form of rainbow-color bicycle chains or sprockets [6].

One possibility to produce such color gradients is the deposition method by means of PECVD, which creates iridescent interference colors. If an increased deposition is not wanted, one can attempt to homogenize the electric field around components being coated by using auxiliary electrodes. However, this approach to create layers in uniformly distributed thicknesses along edges entails a huge amount of equipment and control technology, , and processes quickly reach their limitations with increasing levels of component complexity. Furthermore, with PECVD processes, layer deposition in tiny recesses in the sub-millimeter range becomes extremely difficult, if not impossible. This means a different process is required for coating components such as watch parts. One viable solution for this is atomic layer deposition, or ALD.

Atomic layer deposition

Unlike CVD or PECVD processes, atomic layer deposition (ALD) is a two-stage process [3, 4]. A surface is exposed to two different reactants in succession: The first reactant forms a monolayer through adsorption on the surface, i.e. a layer is created only one molecule in thickness. It is crucial that no further molecules of the first reactant bind to the monolayer once it is complete. This process is therefore self-limiting. After a step of purging and pumping, a second reactant is introduced, which reacts . with the monolayer to a solid product in form of a thin, monomolecular or monoatomic layer.

With thermal ALD processes, specific limits of applicable temperatures must be observed for successful deposition. If the temperature is too low, the reaction speed drops significantly and it takes a long time to complete a single cycle. In addition, condensation may occur, resulting in a higher deposition rate than expected. On the other hand, if the temperature is too high, this either leads to desorption of the precursor from the surface (deposition rate lower than expected) or, if this does not appear, to thermal decomposition on the surface (deposition rate higher than expected). In some cases, it is not possible to find a processing window that works in practical terms.

To circumvent this limitation or make it possible to use an ALD process in general, the reactivity of a precursor can be increased, for example by activation or ionizion in a plasma [5]. This technology is known as plasma-enhanced atomic layer deposition (PE-ALD). It not only makes deposition at significantly lower temperatures possible, but also allows temperature-sensitive materials (such as many types of plastics) to be coated. Layers produced using PE-ALD have zero defects and are extremely homogeneous. It is important to note that the plasma is only used to activate the co-reactant, so the process remains a two-stage procedure. Obviously therefore, PE-ALD processes offer clear advantages over straightforward PECVD processes if components with more complex geometries need to be coated.

Structure of an ALD coating system

The core component of an ALD system is a process chamber, in which the substrate is maintained at processing temperature and process gases are introduced. A pumping unit comprising two different vacuum pumps is used to evacuate the chamber. The aim here is to avoid reactions with molecules in the residual atmoshere. The optimum base pressure of the vacuum usually lies in the range of around 0.002 mbar. To monitor the pressure, suitable pressure sensors are used. High-speed ALD valves are used to introduce the starting chemical (precursors) and the purge gas. Liquid and solid precursors are heated, gaseous precursors and purge gas are regulated using a mass flow controller (MFC). To generate the plasma, a high-frequency plasma generator is used with an electrode located above the substrate holder.

Producing decorative and durable ALD interference coatings

Deposition of colored decorative layers is only possible on shiny metallic base materials. Matt substrates offer insufficient reflection properties along lower interfaces. As a result, plastics first have to be coated in metal to create an interference layer. The glossier the surface on which the coating is deposited, the more brilliant the interference color appears.

The best effects are created on chrome-plated surfaces, although stainless steel and titanium also produce excellent results. On aluminum, colors usually look slightly duller.

ALD layers deposited at low temperature (T < 100 °C) usually consist of metal oxides such as titanium dioxide or aluminum oxide. Layer thicknesses range from around 20 to 100 nm. These oxide layers are thermodynamically stable and offer excellent resistance to acids, alkalis, and organic solvents. As a result, they also offer optimum corrosion protection. In addition to this, with the right coating they can also be guaranteed to be skin-friendly or biocompatible. With sufficiently hard substrates, ALD oxide layers can even be produced to be highly scratch-resistant. Watch parts, such as springs or screws, can be effectively colored with it.

Theory and practice always go hand in hand

The research conducted on ALD coating shows once again how to succeed with knowledge-sharing and the transferal of know-how into practice. While the experts at Furtwangen University are laying a theoretical foundation, and the team from the Steinbeis Transfer Center for Surface and Coating Technology are conducting experiments to test the suitability of the concept for use on small, complex-shaped components.

This is a revised version of an article first published in German in issue 10/2023 [7] of WOMag magazine.