Adsorption Humidity Effects, Microparticle Rate Behavior, and Thermal Swing Adsorption

Abstract

Application of adsorption processes for air purification requires an approach, which accounts for the presence of humidity. Four separate but related studies are conducted to examine the adsorption processes.

A new pure component adsorption isotherm is developed to describe Type 5 adsorption. The results are used to correlate data of water on activated carbon. This model derives from the concept that capillary condensation accounts for Type 5 behavior and is strongly dependent on the pore size distribution. The new model has the advantage over all other prior models of being invertible in terms of loading and partial pressure. The Henry's law limit and heat of adsorption effects are discussed.

A study of coadsorption of water and immiscible organics is also presented. Data for the system chloroethane water on two activated carbons is measured. A new coadsorption model is developed to describe immiscible vapors and water. This model has the advantage of at most one adjustable parameter and can also be solved without iteration. Good agreement is demonstrated between this new model, data measured here and literature data.

The use of thermal swing adsorption for air purification is examined in this work. An experimental system is used to perform cycling experiments under dry and humid conditions. A dynamic simulation model is developed to describe several of cycling runs. Using the coadsorption model developed above the good agreement is found between the data and simulation profiles. Optimization of cycle parameters was investigated to show that some moderation of the feed water content is required to obtain high purification of a light vapor challenge at ambient temperature conditions.

The internal rate effects of commercial adsorbents have been reported in the literature. There is seldom an attempt to review the many approaches. Data was measured using a gravimetric technique for chloroethane and hexane on BPL activated carbon and 13X molecular sieve. A distributed parameter micropore diffusion model was solved to simulate this data. Regression of the adsorption and desorption data was used to determined micropore diffusion coefficients. These values were shown to compare well with literature values.