by David C. Darwin, PhD., and Gregory R. Schoofs
The drive toward lower U-values has created rapid growth in the production of argon-filled insulating glass (IG) units. Desiccant choice and usage can have a substantial impact on the argon concentration and the life of argon-filled IG units. To fabricate high quality argon-filled IG units, there are compelling reasons why we believe a blend of 3A molecular sieve/silica gel should be used in as many sides as practical. Desiccants which contain large pore molecular sieves (commonly referred to as 4A and 13X), either alone or in blends, are detrimental and should be avoided.
The desiccant used in argon-filled IG units should have the following characteristics:
1. The desiccant must adsorb water, thereby protecting the IG unit against moisture fogging;
2. The desiccant must adsorb hydrocarbons (and other volatile species), thereby protecting the IG unit against chemical fogging and against staining coated glass;
3. The desiccant must not adsorb and desorb argon or nitrogen, thereby minimizing glass deflection and seal stress; and
4. The desiccant should not contain pre-sorbed (i.e., previously adsorbed and held) nitrogen, thereby eliminating the outgassing of nitrogen into the sealed IG unit which, in turn, dilutes the argon concentration.
Historically, silica gel, 3A molecular sieve, 4A molecular sieve, 13X molecular sieve, and their blends have been used for desiccation of IG units. Lets take a look at these desiccants with respect to the four characteristics.
It is generally recognized that all of the aforementioned molecular sieves strongly adsorb water. Of these desiccants, 4A molecular sieve is the least expensive to manufacture; silica gel is the most expensive.
Hydrocarbons (and other volatile species) are a common impurity in IG units. Sources of hydrocarbons include:
1. Painted metal muntins or spacers1;
2. Vinyl or plastic internal muntins, spacers, or keys which outgas hydrocarbons when exposed to heat or ultraviolet light (UV);
3. Chemicals used to clean or wipe down muntins, spacers, keys, etc.;
4. Cutting and machine lubricants;
5. "Touch up" paint used to repair marked or scratched muntins or spacers; and
6. Sealant systems containing volatiles.
Types 3A and 4A molecular sieves have no capacity for hydrocarbons because their pore sizes are too small.2 This leaves 13X and silica gel to be considered for hydrocarbon removal.
It has long been recognized that the hydrocarbon capacity of type 13X
molecular sieve is dramatically reduced if water is present. Although the pores or openings of 13X are large enough to admit hydrocarbons, the preferential adsorption of water severely restricts the ability of 13X to adsorb and hold hydrocarbons. As correctly stated by a major desiccant manufacturer, "molecular sieves have an extremely high adsorptive attraction for water. This affinity is so strong that water will normally displace any other material that is already adsorbed on the molecular sieves."3
One might hope that blends of 3A and 13X would reduce the affinity of 13X for water, but this is not the case. Data from molecular sieve suppliers clearly show that the attraction for water on 3A, 4A, and 13X is essentially the same. Thus, if any water is present in an IG unit containing a blend of 3A and 13X, the water will be essentially equally adsorbed and distributed on both molecular sieves. As the 13X molecular sieve adsorbs water, both its attraction and its capacity for hydrocarbons decrease. As water continues to be adsorbed on 13X molecular sieve, previously adsorbed hydrocarbons will be outgassed into the "airspace."
Hydrocarbon and moisture molecules are, literally, like oil and water. Because they are fundamentally different molecules, a blend of fundamentally different adsorbents with totally different properties is best suited for removing the hydrocarbons and moisture from the "airspace." As described in U.S. Patent No. 4,144,196, ("Adsorbent for Use in Double Glazed Windows"), a blend of 3A molecular sieve and silica gel is ideal. This blend is non-separating because the densities of both components are essentially the same. The 3A molecular sieve selectively and preferentially adsorbs water, leaving the silica gel totally active and free to adsorb hydrocarbons without any interference from water.
Day to night temperature changes cause an IG unit to behave like an accordion. In accord with the "Ideal Gas Law," the pressure within a sealed IG unit increases whenever the temperature increases. This is like the pressure increasing inside an automobile tire whenever it heats up. The Ideal Gas Law also mandates that the pressure within an automobile tire or a sealed IG unit decreases whenever the temperature decreases.
Nitrogen and argon adsorption and desorption from 13X and 4A molecular sieves as the temperature changes amplify the accordion effect. As shown in figure one, 13X molecular sieves, and desiccant blends which contain 13X, adsorb nitrogen whenever the temperature decreases and desorb ("outgas") nitrogen whenever the temperature increases. Virtually identical behavior occurs with 4A molecular sieve.4 Similar, though less pronounced, behavior also occurs on 13X and 4A if argon is present instead of nitrogen.5 Desorption, which occurs as the temperature increases, means that nitrogen or argon molecules are outgassed from the desiccant; this increases the number of molecules in the "airspace," which, in turn, causes the pressure to increase. Similarly, adsorption of argon or nitrogen at low temperatures removes molecules from the "airspace," thereby decreasing the pressure.
The pressure changes resulting from the adsorption/desorption of argon or nitrogen amplify the pressure changes above and beyond those mandated by the Ideal Gas Law. Glass deflections which occur unavoidably with any use of 13X or 4A molecular sieves:
1. Distort reflected images;
2. Harm the U-value of the IG unit;
3. Stress or destroy the seal; and
4. Reduce the life of argon-filled IG units.
A blend of 3A molecular sieve and silica gel, which neither adsorbs nor outgasses nitrogen, oxygen, or argon, is the obvious solution to this problem.6
To achieve lower U-values by utilizing argon gas, manufacturers want to replace as much of the ambient air in IG units as possible with argon gas. To attain an argon fill level of 100 percent, all air, which is 79 percent nitrogen, must be eliminated from the "airspace." In the common practice of "lance filling" argon through one or two holes in the conventional IG spacer, the following factors limit the argon concentration to less than 100 percent:
1. Pre-sorbed nitrogen brought into an IG unit on 13X or 4A molecular sieves;
2. Air in the free space between desiccant particles and in the macropores of the particles;
3. Air inside empty spacers; and
4. Air inside hollow muntins.
Large pore molecular sieves, such as 13X, 4A, and desiccant blends containing 13X or 4A, readily pre-sorb nitrogen from ambient air prior to fabricating and sealing an IG unit.4 At sea level and 65F, every ounce of fully-activated 13X has pre-sorbed approximately 12.5 cubic inches of nitrogen gas (see figure one). Type 4A holds a nearly equal amount of nitrogen.4 This translates to 5.5 spacer volumes of nitrogen for each spacer volume filled with 13X or 4A molecular sieve. Additionally, the volume of free air between the particles and in the macropores of the particles corresponds to approximately 0.54 spacer volumes, for a total of 6.04 volumes of nitrogen plus air for each spacer volume filled with 13X or 4A.
A similar, though somewhat less pronounced, effect occurs with any blend containing 4A or 13X molecular sieves. For example, a molecular sieve blend containing 20 percent 13X has pre-sorbed approximately 1.1 spacer volumes of nitrogen for each spacer volume filled with a 20 percent 13X molecular sieve blend. Add to this the volume of free air between the particles, and the total becomes approximately 1.64 spacer volumes of nitrogen plus air for each spacer volume filled with a 20 percent 13X molecular sieve blend.
Type 13X, Type 4A, and molecular sieve blends containing 13X or 4A, will outgas some of the pre-sorbed nitrogen into the "airspace" as they equilibrate with the argon fill.5 Additional pre-sorbed nitrogen will be outgassed as the 13X or 4A preferentially picks up moisture or hydrocarbons from the interior of argon-filled IG units and as the desiccant increases in temperature during normal daytime heat cycles.3,4,5
Nearly all of the pre-sorbed nitrogen will eventually outgas and accumulate in the "airspace." Nitrogen outgassed from 13X and 4A molecular sieves:
1. Dilutes the argon concentration; and
2. Can create an over-pressure condition in IG units, which causes glass deflection and the associated problems.
In contrast, there is virtually no pre-sorbed nitrogen introduced into an IG unit in a 3A sieve due to the pore size which prevents nitrogen adsorption; nor in a silica gel which does not have an affinity for the adsorption of nitrogen or argon.
Argon concentration can be diluted further by air initially present in spacers and hollow muntins. Argon displacement of air from spacers, whether empty or desiccant-filled, and hollow muntins is an inefficient process because it takes a longer time for the air to be displaced and diffuse out of the spacers and muntins than the time typically allocated to argon fill the "airspace."
As an example of the dilution of argon by air from desiccant, spacers, and muntins, we considered a two-foot by three-foot by 1/2-inch IG unit whose "airspace" is argon-filled to an initial level of 100 percent, the best case scenario. Figure two compares the actual argon concentration following nitrogen and air dilution from the four factors outlined above. A minimum initial argon fill (i.e., concentration) of 90 percent has been instituted as a certification standard by at least one North American certification program.7 Figure two shows that it is difficult to meet this certification standard if the IG unit contains 4A or 13X molecular sieves, either alone or in blends.
Filling as many sides as practical with a 3A molecular sieve/silica gel blend provides a simple and inexpensive way to minimize argon dilution in IG units because:
1. There is essentially no pre-sorbed nitrogen to release into the "airspace;" and
2. The blend itself occupies volume in the spacers that would otherwise contain air not readily displaced during argon filling.
Better performance and longer life of argon-filled IG units can easily be obtained with proper desiccant selection and usage. We believe a blend of 3A molecular sieve and silica gel to be unmatched with its combination of high moisture and hydrocarbon capacity, minimum argon adsorption/desorption, negligible amount of pre-sorbed nitrogen, and minimum glass deflection. USG
1. Juergen H. Braun and Daryl P. Cobranchi, Journal of Coatings Technology, December
1995, Volume 67, Number 851, pages 55-62.
2. Donald W. Breck, "Zeolite Molecular Sieves", John Wiley & Sons, 1973.
3. Union Carbide Corporation technical literature, "DRY GAS?".
4. Donald Peterson, Zeolites, July 1981, Volume 1, pages 105-112.
5. George W. Miller, AIChE Symposium Series, 1987, Volume 83, Number 259, pages 28-39.
6. U.S. Patent No. 4,144,196.
7. IGMAC Certification Criteria.
David C. Darwin, Ph.D. is technical service manageradsorbents for Grace Davison in Baltimore, MD. Gregory R. Schoofs is president of Schoofs Inc., in Morage, CA.
© Copyright 1999 Key Communications, Inc. All rights reserved. No reproduction of any type without expressed written permission.