Topic: Characterization of Adsorbents

Nitrogen production from air by PSA (pressure swing adsorption) using CMS (Carbon Molecular Sieve) adsorbents was first proposed by Juntgen et al. in 1981 [1]. Juntgen and co-authors disclosed a process for producing nitrogen at 99.9 % purity with 40 % recovery (though the nitrogen stream also contains about 1 % of argon). Presently a PSA unit can produce nitrogen at purity of 99.9995 % and up to 10000 m3N/h, at 97 % purity. This technology is finding a very significant place in the market for nitrogen production, mainly because the lower costs and higher flexibility compared to cryogenic production [2].

In a PSA adsorption column, nitrogen is separated from air based on the kinetic selectivity of oxygen over nitrogen in absorbents such as CMS. Oxygen is much faster entering the adsorbent micropores while nitrogen, which is much slower is limited to the intraparticle voids and is drawn from the adsorbent column at high pressure. The basic PSA cycle for nitrogen production consists of two adsorbent columns working out of phase and has four basic steps: pressurization, production, blowdown and desorption [3].

The kinetic selectivity of a CMS adsorbent is based on a narrow pore size distribution that can sterically discriminate between oxygen and nitrogen species. Since the expiration of the original patent, several companies began developing and producing new CMS adsorbents that show improved performance. There is then an emerging interest in characterizing these adsorbents to aid in developing optimized PSA cycles for nitrogen production.

This paper reports the characterization of three CMS adsorbent samples from two different suppliers: a CMS D from Carbotech and a KP325 and MSC3K-162 from Japan EnviroChemicals, Ltd. These adsorbents were characterized by measuring the adsorption isotherm and uptake curves of oxygen, nitrogen and argon, from 0 bar to 7 bar and at various temperatures. Afterwards, the effective diffusivity of oxygen, nitrogen and argon was obtained as function of the pressure at the selected temperatures. A model was fitted to the effective diffusivities as function of the pressure and this inserted in a mass balance equation modeling breakthrough experiments. Finally, some breakthrough experiments were performed using a packed bed column and the results used to assess the model proposed.

It was concluded that for all three CMS samples, the effective diffusivity increases with sorbate concentration. Preliminary results show that the oxygen/nitrogen selectivities are similar for all adsorbents, but CMS D has the highest diffusivities followed by MSC3K-162. KP325 shows the highest adsorption capacity. The breakthrough experiments will allow validating the proposed model that will be used to select the best adsorbent and to help the development of the new generations of CMS adsorbents.

References

[1] – Juntgen, et al.; Process for the recovery of nitrogen-rich gases from gases containing at least oxygen as other component. 1981. US 4,264,339

[2] – Agrawal, 2001 R. Agrawal, Separations: perspective of a process developer/ designer, A.I.Ch.E. Journal 47 (2001), p. 967

[3] – Ruthven, D. M., Farooq, S., Knaebel, K. S. (1994). Pressure Swing Adsorption, VCH Publishers, New York.