Site Search
+1 800.241.2523

Coating Materials News Vol 9 Issue 1


Volume 9, Issue 1    March, 1999

In this issue we examine a common problem found with optical coatings, and discuss the various forms and consequences that this problem exhibits as well as techniques for minimizing the effect.

Water in Coatings
Thin film coatings grow with a density less than that of the bulk form of the material and consequently possess a considerable void volume. Density is dependent on the energetics of the layer growth and the incorporation of gasses during deposition. In review of earlier discussions, increased packing density is produced by high substrate temperatures, ion bombardment of various forms, or high impact energy of the primary species. Figure 1 is a SEM microphoto of a thick CdTe film grown at 60° C substrate temperature. When the temperature was raised to 200° C, the film becomes featureless and apparently amorphous. Amorphous growth, as opposed to crystalline or columnar microstructure, is required to provide a high packing density and therefore lower penetration of water vapor.
High substrate temperature assists in removing surface contamination, thereby conditioning the surface for nucleation, increases surface mobility of the arriving adatoms, and can either decrease or increase the water and other gas content of the atmosphere. Cases of increased water content can result from wall and tooling outgassing at high temperatures. Ion bombardment occurs when a plasma is present in the vicinity of the substrate as in the many forms of sputtering and plasma-assisted processes. High impact energies are produced in sputtering or in ion-assisted deposition (IAD) where the substrate is deliberately showered by high-energy argon ions. The greater arrival energy of e-beam vs resistance-heated evaporation generally results in denser films. These deposition techniques are employed to densify deposited layers, however, other means can be and are used towards this end.
As a first step, effort must be made to remove as much water from the deposition environment as possible. Chamber walls and tooling hold surface water that is continually being released to the deposition atmosphere. High rate exhaustion of this water and other gasses through baking out under vacuum, cold trapping, or pumping is required. Cryopumps, in contrast to diffusion pumps, are better at removing water. Sometimes a cold trap is placed in the coating volume to trap volatile water through condensation.
The starting material and its form and preparation are important in limiting the introduction of water. Absorbed and adsorbed water and other gases are released in going from atmosphere to vacuum. Some materials hold more water than others. Materials that have been conditioned by pre-melting, fusing, or mixing retain less water than materials that are simply sized into smaller pieces of the parent crystallized material, for example. We are all aware of the sometimes explosive showering and spitting that occurs with dust and small particles in the crucible. All pure fluoride compounds hold water, while few oxide compounds do. IR materials such as ZnS, Ge, and other semiconductors generally remain dry, as do metals. Many salts, used in far-IR coatings, are hydrophilic and even hygroscopic and require special care, for example cryopumping, to be successfully deposited with stable film properties.
When a film layer condenses with porosity instead of bulk density, water vapor can diffuse in and occupy the internal voids upon exposure of the layer to the atmosphere. The internal surfaces are highly reactive and bonding forces are high. Several phenomena can then transpire.

Spectral Shifting

The effective refractive index of the layer increases with increased diffusion of water vapor, causing a spectral shift (longward). This change would be acceptable if it was permanent and could be compensated for in anticipation, but some fraction of the water bonds are weak, permitting water to escape upon evacuation of the layer. This is a consistent problem for filters and AR coatings that are intended for vacuum operation as in space missions. Oxide compounds that grow with columnar microstructure are notorious for this problem. Titania layers exhibit the effect to a greater degree than silica layers because the latter tend to grow as glassy films (amorphous). Hafnia films exhibit the water-caused shift , and it is observed with nearly all oxide compounds. The problem is especially visible with alumina films combined with silica or titania layers. Blotchiness is reported that requires the passage of days for absorption to complete and achieve uniform appearance.
An example of the shift between atmospheric humidity (~50%) and dry nitrogen purging illustrates the phenomenon [1]. A thirty-layer stack of titania and silica quarterwaves deposited by e-beam on a 250° C substrate was examined. The startling observation was that an 8 nm shift of the spectral edge of the coating was observed to occur within 1-1.5 min, and was completely reversible. The porosity of the 30-layer stack evidently was quite high.

Water Bands

Another phenomenon that occurs is the appearance of spectral absorption water bands, specifically in the 2.8 to 3.2 µm and 6.0 to 7.4 µm regions of the IR. These absorptions can be troublesome for certain applications and must be minimized. The depths of the absorption bands can be used as a measure of the success of the deposition parameters in compacting the layer. Since IR coatings nearly always require a fluoride compound for the low index component, and fluorides notoriously retain water, efforts must be made to reduce the water content of the layers [2]. These efforts take the forms of deposition by e-beam at high substrate temperatures, IAD, preconditioning the material to drive as much water off as possible, using mixed materials, and water pumping. Generally, slow heating of the starting material has a significant effect in driving the water out. Fluoride sputtering techniques are being developed, but are not as straightforward or hazard-free as sputtering oxide compounds. Therefore, starting material preparation and pre-deposition conditioning are the processes steps that should routinely become part of the deposition parameter set.
If a material is hydrophilic, small particle sizes with their greater surface area must be avoided. It is preferable to start with larger chunks. Under the shutter and in vacuum, the material is slowly heated, and, if e-beam is used, the low power beam must be swept rapidly. When outgassing stops or a fused surface is formed, the material is ready for evaporation. Some materials behave very well in terms of low outgassing and absence of particulate showering (spitting). CERAC IRX® and IRB™ are especially notable for these properties. These materials are solid solutions of mixed fluoride compounds. Silicon monoxide also behaves well because of the way it is formed as a starting material and its condensation as an amorphous microstructure. Materials that melt produce layers that absorb less water than materials that sublime.

Stress Effects

Another effect of water absorption / desorption is changes in the nature and magnitude of layer stress. Fluoride films often develop tensile stress when exposed to atmospheric moisture. This effect is reversible to the degree that the volatile (weakly bound) component is exchanged between vacuum and air. In some cases, the sign and magnitude of the stress can move in directions that improve the integrity of the coating upon exposure to air. The coating is said to have “seasoned”.
In the case of multi-layer coatings, changes in total stress level can result in crazing, cracking, and loss of adhesion to the substrate. The change occur over hours or days depending on the diffusion rate. It has been observed that a coating that appears sound upon venting can spontaneously break up after some time of residence in normal atmosphere. The MIL-spec humidity soak is designed to detect this tendency. The preventative measure is to insure that the materials of the multi-layer grow with amorphous form to prevent moisture penetration, and form strong mutual bonds. This is not always accomplished with high substrate temperature because high temperatures promote large grain growth with accompanying larger interstitial spaces. Slow growth under high vacuum conditions can help some materials, but obviously not oxides that require reactive combination with a substantial background pressure of oxygen. IAD is useful under these deposition conditions, and has been observed to reduce stress mainly because columnar growth microstructure is discouraged in the presence of high impact energies.

Laser Damage Threshold

There is evidence that the absorbed water lowers the LDT of oxide coatings for near-IR laser applications. Water has an absorption bands near 830 nm, 905 nm, 955 nm, 1.34 µm, and 1.37 µm. Heating by high-energy irradiation can cause debonding and mechanical effects that result in coating destruction. Again, densifying methods are required to increase the LDT.


1. Samuel F. Pellicori and Herbert L. Hettich, Appl. Opt. V27, No. 15, 3061 (1988).
2. S. F. Pellicori and E. Colton, Thin Solid Films, 209, 109 (1992).
3. Fink, Yoel, Winn, Fan, Chen, Michel, Joannopolous, Thomas, Science, 282, 1679 (1998).
4. Angus Macleod, Macleod Medium, V7, No. 1 (1999).

Dr. Mitchell C. Colton, Editor
CERAC, inc.
P.O. Box 1178 | Milwaukee, WI 53201
Phone: 414-289-9800 | FAX: 414-289-9805

Samuel Pellicori, Principal Contributor
Pellicori Optical Consulting
P.O. Box 60723 | Santa Barbara, CA 93160
Phone/FAX: 805-682-1922

back to top Copyright 1999, CERAC, inc.

All printed, graphic and pictorial materials made available on this website are owned by CERAC and protected by Federal Copyright laws. None of the materials, in whole or in part, may be reprinted and distributed or otherwise made available to others for any purpose without CERAC's prior written consent.

Phone:  414-289-9800 /  FAX: 414-289-9805  /