Before an optical thin film coating can be fabricated, a coating design is needed that will meet the requirements. Before the design can be finalized, it is necessary to know the n- and k-indices of refraction of the materials to be used. Before the indices can be known, it is necessary to have deposition processes, perform test coatings, and measure and characterize the n- and k-values of the materials to be used in the coating. The deposition processes also require the proper equipment.
The design process is well-defined  and straightforward in modern times, and this is highly facilitated by computer assistance. The fabrication process is usually the area with most of the issues. Almost any coating can be designed, but not every design can be fabricated .
The durability of an optical thin film coating is a major issue. It must stick/adhere to the substrate and have sufficient hardness and durability under any of the environmental conditions in which the coating is expected to operate. The coating may be subject to varying humidity, laser irradiation, temperature extremes, etc. Stress in the coating due to material combinations and processes can cause cracking and delamination, particularly when subjected to temperature extremes. This is mostly due to the difference in the thermal coefficients of expansion of the substrates and the coating materials.
The most common optical thin film coatings are done by physical vapor deposition (PVD) in vacuum coating chambers at pressures from about 10-6 to 10-4 Torr. In such cases, the coating material is heated until it evaporates, and then it condenses on the substrates which are mounted in the chamber to receive the coating. The next most common approach is probably the sputtering process, where the coating material is bombarded with accelerated atoms at chamber pressures that are one or two orders of magnitude higher than the evaporation processes mentioned. Sputtering has the advantage of coating the substrate with more energetic atoms that tend to provide better adhesion and denser coatings. A newer technology which may soon become useful in the optical coating field is atomic layer deposition (ALD). This is typically a chemical vapor deposition (CVD) process similar to one that has been used in the semiconductor industry for many decades, but the chemistry has now been developed to allow processes that work at much lower temperatures where glass and even plastic optics can survive the process without damage.
A major fabrication issue is to maintain a stable process so that the production results will remain consistent. Some of these issues are addressed in current publications . With the typical PVD evaporation process, key parameters that can disrupt stability/reproducibility are the process pressure, temperature, and deposition rate. Contamination of surfaces being coated can also have a major effect on adhesion. This might be from the pumping oils used in many coating systems.
In the 1960’s, it was imagined that optical coatings were glassy/vitreous and dense, but it has since become clear that actual optical coatings are seldom anything like that. Figure 1 from Guenther  shows an approximate two-dimensional computer simulation of the film-growth process as we now understand it.
Figure 1: Computer simulation of film growth
Figure 2 from Liu et al.  shows an electron micrograph of an actual MgF2 film illustrating the porosity and columnar structure of the film.
Figure 2: Magnesium fluoride optical thin film, showing porosity and columnar growth
Figure 3 is a striking micrograph of the structure of TiN from Macleod’s book .
Figure 3: Electron micrograph of unusual titanium nitride growth structure
What is going on here? The following thoughts are an attempt to explain what is observed.
It can be shown that an evaporating atom leaving a heated surface has very little energy, on the order of 0.1 electron Volts (eV). If it travels relatively unimpeded through a near-vacuum and comes in contact with a cold substrate or another cold atom, it looses its energy upon contact and sticks to the surface where it first comes in contact. If, on the other hand, the substrate or other atom were hot enough, then the arriving atom would not give up its energy as quickly, and it might be able to move some distance after contact before all of its energy is gone and it comes to its final resting place. This would allow it to “nestle” in close to other atoms on the surface and form a denser coating. Figure 1 illustrates the “cold” case where the atom stops essentially at its first point of contact. This would build a frost-like or feathery microscopic structure. Models and experience show that more energetic atoms will move somewhat before cooling, and that they will form denser structures in proportion to the energy of the atoms and a higher temperature for the substrate.
The influence of atomic energy and substrate temperature on film properties has become much better understood over the last several decades. In the 1940s, the depositions were mostly done at room temperature, and many materials were investigated to find out which had the best properties under those conditions. Substrate heaters were added to chambers in the 1950s and 1960s to improve the results. Magnesium fluoride became the preferred single-layer antireflection coating (SLAR). When MgF2 is deposited on room-temperature substrates, it is so fragile that it can be wiped from the substrate with a finger. However, if the substrate is at 300°C, the coating will be robust enough to withstand many eraser rub stokes when applied with some significant force.
In the 1970s and 1980s, ion-assisted deposition (IAD) was introduced into optical thin film coating, and depositing surfaces could be bombarded with ion energy to provide energy/heat to the growing film, yielding similar effects to heated substrates (even if the substrates were not otherwise heated). This has resulted, for example, in the ability to produce hard MgF2 coatings on unheated substrates , including plastics (which cannot withstand more than about 100°C). The IAD replaces and/or augments the substrate heating as a means to keep the atoms from cooling too soon and allowing them to migrate into a denser structure.
Figure 4: Typical ion source used for IAD in optical coating chambers
Figure 4 shows a typical ion source  used for IAD in optical coating chambers. An ion source ionizes the gas admitted into it, and it accelerates those ions toward the substrate. The acceleration is controlled to about 60 and 180 eV for best effect. High ion current is usually used to provide the desired effects at as high a material deposition rate as practical. Such sources usually use argon ions because they are inert, but oxygen or nitrogen is often used when oxide or nitride materials are being deposited. With oxides and nitrides, the IAD can densify the films and complete the chemistry of the films to a fully stoichiometric condition. If films like TiO2 and Ta2O5 are not fully oxidized (stoichiometric), they will have some absorption or k-value, which can be an issue.
When films are not fully dense, humidity can fill the voids and raise the effective index of the film. This changes with humidity and will cause the wavelengths of the filter edges to shift by several percent, which can be an issue for the performance. The densification capabilities of IAD can help to avoid this.
As has been mentioned, the energy of the depositing atoms as a function of time is a key issue in determining the properties of the optical thin film produced. This usually is reflected in the microstructure of the layers and its overall physical properties. The intrinsic properties of the coating material chosen and the process used to deposit it are key issues to the success of the coatings. Great progress in the understanding of these issues and the fabrication of optical thin film coatings has been gained in the last half-century.
About The Author
Ron Willey graduated from MIT in optical instrumentation, has an M.S. from the Florida Institute of Technology, and over 40 years of experience in optical system and coating development and production. He is very experienced in practical thin films design, process development, and the application of industrial Design Of Experiments methodology. He is the inventor of a robust plasma/ion source for optical coating applications. He worked in optical instrument development and production at Perkin-Elmer, Block Associates, United Aircraft, Martin Marietta, Opto Mechanik, Hughes, and formed Willey Corporation, which serves a wide variety of clients with consulting, development, prototypes, and production. He has published many papers on optical coating design and production. His recent books are “Practical Design of Optical Thin Films”, 4th Ed. (2014) and “Practical Production of Optical Thin Films,” 2nd Ed. (2012). He is a fellow of the Optical Society of America and SPIE and a past director of the Society of Vacuum Coaters.