An Osmotic Membrane Dehumidifier

  • Type:
    Conference Presentation
  • Conference Type:
    AIChE Annual Meeting
  • Presentation Date:
    October 17, 2011
  • Skill Level:
  • PDHs:

Share This Post:

You will be able to download and print a certificate for PDH credits once the content has been viewed. If you have already viewed this content, please click here to login.


It has long been known that relying on a vapor compression system to cool air down to the dew point in order to dehumidify is expensive to build and uses far too much energy to operate. The ability to dehumidify independently of cooling provides substantial advantage, provided that it can be done using inexpensive materials, uses little energy to operate, and is easily packaged with a separate cooling system.

An osmotic membrane dehumidifier has been developed that uses a flexible, semi-permeable membrane to facilitate capillary condensation of water vapor and the transport of condensed water through the membrane into a salt solution by osmosis. Here a humid gas stream is brought into contact with a semi-permeable membrane, which separates the gas stream from an osmotic (e.g., salt) solution. Some of the pores of the membrane are small enough to permit capillary condensation. Liquid formed within these pores connects with liquid formed in adjacent pores, collectively forming continuous paths of liquid. These ‘liquid bridges’ extend across the thickness of the semi-permeable membrane and provide paths by which water can travel across the membrane. Because the membrane is so thin, water concentration gradients across the membrane can be large. This can provide a large driving force for water transport between the humid air and the osmotic fluid. The flexibility of the polymeric membrane allows for considerable design flexibility that enhances the potential for retrofit with any cooling system.


Conditioning of a gas stream, such as air, in warm weather generally involves removal of moisture and decrease of temperature to make the gas stream suitable for its intended environment. Current dehumidification technology is commonly based on the conventional, refrigerant vapor compression technology (DX technology) or on desiccant substrate capture (DS technology). DX technology requires cooling humid supply air, such as the air within a room and/or outside air, to the water vapor condensation point, with external heat rejection on the compression side. This usually requires the supply air to be cooled below comfortable temperatures, and, thereafter, either reheated or mixed with warmer dry air to raise its temperature to an acceptable level before directing it into the space being dehumidified. Twenty to thirty-five percent of the energy expended in cooling the high humidity air is utilized to remove the latent heat of condensation from the air. Cooling and dehumidification of the air are thus coupled. That makes it impossible to control comfort parameters independently, making the DX cycle less efficient, from an overall system perspective, than a technology that would allow independent control of sensible and latent heat.

In applications where the outside air has both high humidity and temperature and the functional use of the interior space generates high water vapor levels (e.g., populated convention halls, exercise rooms, school buildings, etc.), it may not be possible for the DX technology to maintain the air introduced into the interior space at the correct humidity and temperature for maintaining comfort. The air delivered is cool but ‘muggy’, since further cooling to remove additional water would result in the air being uncomfortably cool.

In stand-alone dehumidification using a conventional compression cycle, heat reject is in direct contact with room air. As a consequence, the room air becomes more comfortable from a humidity side, but may be too warm. Again, the comfort parameters are coupled.

DS systems are generally applied in central air, ducted systems. Water vapor capture on a solid phase substrate is efficient and rapid. However, removal of the water vapor from the pores requires energy input. It also requires removing the substrate from the high humidity air stream and placing it in an exhaust, water-reject stream, before adding the re-evaporation energy. Alternatively, the substrate remains fixed and the treated air and exhaust streams flow directions interchanged as is done in a parallel bed, desiccant dryer system. DS technology requires, in steady state operation, the addition of this energy at a rate equal to or greater than the latent heat of condensation of water in the desiccant substrate.

The advantage of DS technology is that humidity levels in the outside air and/or re-circulated air can be adjusted independently of the cooling step. The disadvantage is the requirement to move the substrate and treated air stream relative to each other for capture and rejection of the water vapor.

In a liquid desiccant system, direct contact equipment, such as packed-bed towers, allows for simultaneous heat and mass transfer between the air and the liquid. The essential issue here is the requirement for intimate contact of desiccant and air. The conventional technique of using a packed-bed adsorption tower is not energy efficient due to its low contact surface area per unit volume. Moreover, evaporation of some liquid desiccant in a packed-bed tower may cause contamination of the conditioned air stream.

Osmotic dehumidification provides a more compact and efficient option. The process is described in Refs. 1 and 2. In the earlier work, a capillary condenser with pore sizes around 2 to 3 nanometers was used to condense water vapor in air at moderately high relative humidity. The condenser (originally designed and fabricated by Mobil Oil) had been fabricated from a ceramic substrate using a molecular template which was subsequently calcined at high temperature to leave behind a fairly uniform porous structure. A semi-permeable membrane was used to separate the capillary condenser from an osmotic fluid. An osmotic driving force, resulting from a water concentration gradient, transports the condensed water through the condenser and into the osmotic fluid. This type of device is effective in promoting water transport at fluxes in excess of l liter/square meter-hour. Under some circumstances, however, the pore structure of the capillary condenser can become unstable and degrade over a relatively short period of time in a humid environment. In addition, the capillary condenser is typically made from a rigid material and is therefore limited to those applications where a rigid body is acceptable.


The feasibility of using commercially available membranes for capillary condensation and subsequent transport of condensate into a salt solution was demonstrated first in a laboratory scale experiment. A glass container, filled to the top with a salt solution, was covered with a membrane. A narrow tube attached to the side of the container extends up above the container. Humid air was blown across the top of the membrane and the rate of removal of water from the air was measured by the change in liquid level in the tube. Membranes tested were those used commercially for water purification. Salt solutions that were very effective were lithium chloride and magnesium chloride. The latter salt is inexpensive, benign and readily available. Results suggest that 1 square meter of the membranes that worked most effectively removes about 0.7 liters/hr of water vapor at relative humidity levels between 80 and 90%. The rate of condensation per unit of surface area is considerably smaller than for a uniform pore size capillary condenser. However, the polymer membranes are stable, inexpensive, flexible, and offer simple methods for providing large surface area interfaces between humid air and films of liquid salt solution.


Subsequent experiments were performed using a larger apparatus in which shallow trays were filled with a salt solution and each covered with a membrane. Trays were arranged in a cafeteria style with narrow spaces between the trays for air flow over the membranes. Air was blown over the membranes and then through an exhaust manifold. An anemometer was used to measure air flow and humidistats were placed upstream and downstream of the trays to measure temperature and humidity. A circulating pump was used to move the salt solution through the trays from inlet to outlet manifolds.

Tests were performed on arrays of trays and on single trays with various spacing and air flow. A single tray, for example, covered with a membrane of 0.13 square meters reduced relative humidity of an air stream flowing at 15 cfm (0.4 cubic meters per minute) by 7%. At 20 cfm, the relative humidity was reduced by 5%. An array of 12 of these trays (occupying a volume of less than 0.02 cubic meters) would remove about 0.5 liters/hr of water from a humid air stream. Reducing volumetric flow rates results in increased relative humidity reduction. However, capillary condensation with current membranes is limited by pore size and relative humidity is generally not reduced below 65-70%.

A mathematical model of the transport of air and removal of water vapor at membrane surfaces has been developed to help to optimize the design of the system. The tray system is but one of many potential configurations.

Ref. 1: Kesten, A.S. and Blechner, J.N., Versatile Dehumidification Process and Apparatus, U.S. Patent 7,758,671 B2, July 20, 2010

Ref. 2: Kesten, et al, Dehumidification Process and Apparatus, U.S. Patent 6,539,731 B2, April 1, 2003




Do you already own this?

Log In for instructions on accessing this content.


AIChE Member Credits 0.5
AIChE Members $15.00
AIChE Undergraduate Student Members Free
AIChE Graduate Student Members Free
Non-Members $25.00