PSA Separations

Another important parameter used to gauge the adsorbent s productivity is the product throughput  

Amount (kg) of C3H6 product per hour Amount (kg) of adsorbent 

Run No. Pl (atm) Up (m/s) Up (m/s) CsHg Product Purity (%) CsHg Product Recovery (%) CsHg Product Throughput (kg Product/h/kg of Sorbent) x 10  

Feed pressure Ph = 7.0 atm., step time = 60 s, feed T = 120°C, Pl = des. press., Up = feed velo., Up = purge velo. (Rege and Yang, 2002, with permission). 

The results of the PSA simulations for the four-step cycle described above for the AIPO4-I4 and AgN03/Si02 sorbents with a feed composed of 85% CsHg and 15% CsHg at a temperature of 120 °C are summarized in Table 10.9. The 

---

Another key point of differentiation is the fact that nearly all PSA separations are bulk separations and any investigator interested in a high fidelity description of the problem of adsorption must solve a mass balance equation such as Eq. (9.9), the bulk separation equation, together with the uptake rate model and a set of thermal balance equations of similar form. In addition to the more complicated pde and its attendant boundary and initial conditions the investigator must also solve some approximate form of a momentum balance on the fluid flow as a whole. 

It may not be obvious but driving selectivity to a high value is best done by driving N2 adsorption to some acceptably high value and then driving O2 to a minimum. This dramatically changes the volume of gas that must pass in and out of the macro-pore structure of the adsorbent In aU PSA separations it is the macropore diffusion that is the dominant resistance to mass transfer. 

The mechanics of how adsorption waves and thermal waves move through the adsorption beds is what determines the success of any given separation. Some of the principles underlying design practices have been outlined and unfortunately for current PSA separations only the briefest outline could be given. 

Gupta, R. and Farooq, S. Numerical Simulation of a Kinetically Controlled Bulk PSA Separation Process based on a Bidisperse Pore Diffusion Model. Proceedings of the 8 ArcChE Congress, Vol. 3, p. 1753-1756, August 16-19, 1999, Soul, Korea. 

Farooq et al. [3] introduced a variable diflusivity model to a kinetically controlled PSA separation process. They pointed out that the Dailcen equation with the Langmuir isotherms predicted the experimental data better than the constant diflusivity assumption. Based on those results, following concentration dependent diflusivity model was presented as the adsorption rate models. 

Sircar, S. and Kratz, W.C. Simultaneous production of hydrogen and carbon dioxide from steam reformer off-gas, PSA. Separation Science and Technology, 1988, 23, 2397. 

It is primarily the span of absolnte pressnres, conpled with differences in mole fractions (caused by pressure shifts and/or by admitting different streams to the column), that drive PSA separations. Equilibrium selectivity can canse composition shifts to occur simply by changing the pressure. To illustrate this point, consider an adsorbent bed of zeolite 5A filled with air (assnmed to be only oxygen and nitrogen) at 3 atm. The mole fractions in the gas phase of oxygen and nitrogen are... 

Design pressure swing adsorption (PSA) separation and hydrogen storage systems for the produced hydrogen. 

Figure 4 Hydrogen by steam methane reforming with PSA separation.

Hydrogen by Steam Methane Reforming with PSA Separation... 

The selection of a suitable zeolite adsorbent for CO2 removal from flue gas (mixture of CO2 and N2) has been carried out. The limiting heats of adsorption, Henry s Law constants for CO2 and N2, CO2 pure component adsorption isotherms and expected working capacity curves for Pressure Swing Adsorption (PSA) separation application were determined. The results show that the most promising adsorbent characteristics are a near linear CO2 isotherm and a low Si02/Al203 ratio with a cation in the zeolite structure that has strong electrostatic interaction. 

The prim purpose behind these binary mbcture sorption studies was to develop a theoretical model which was able to predict binary sorption isotherms from the respective single component sorption data with an accuracy which is sufficient for the design of pressure swing adso tion (PSA) separation systems. These studies will, hopefully, also lead to a better understanding of the fundamentals of adsoibent/adsoibate and a oAate/adsoibate interactions. 

The isosteric system described in this paper has been shown to give accurate isosteres when soibates such as hydrocarbons are strongly sorbed by the zeolite phase. These isosteres can be used to calculate accurate isotherms and thermodynamic data. Interesting differences in the sorption behaviour of lower hydrocarbons in silicalite-1 and NaY zeolites have been found which could be used to select the appropriate zeolite for an efficiait PSA separation system for the separation of various binary pairs of suchhydrocarbons. The isosteric system can be employed to study the sorption in silicalite-1 of more weakly sorbed sorbates such as N2 and CH4 and binary mixtures containing one of these gases as a component. However, the strict isosteric principle of constant concentration and composition, in the case of mixtures, of the sorbed phase has been shown not to be adhered to as the temperature is varied across the isostere. Corrections can be made which allow accurate sorption data to be obtained for those systems from which accurate thermodynamic quantities and separation fectois can be obtained. 

The main use for CMS s is nitrogen production from air and CH4/CO2 separation, both by PSA. The latter is applied for (1) landfill gas that contains approximately 50% each of CH4/CO2, and (2) tertiary oil recovery where the effluent gas contains 80% CO2 and 20% of CH4 plus other light hydrocarbons. The PSA separation of CH4/CO2 with Bergbau Forschung CMS has been discussed in detail by Kapoor and Yang (1989) and by Baron (1994), who also discussed several other possible applications. 

Separation and purification can also be accomplished by using differences in diffusivities, that is, kinetic separation (discussed in Chapter 3). 4A zeolite has been used in PSA separation for producing nitrogen for inert purge applications... 

Since 1989, CO separation/recovery by PSA using supported CuCl has been commercialized worldwide. PSA separation results are available in the literature (Kansai Coke Chemicals Co., 1989 Chen et al., 1997 Golden et al., 1998). 

Introducing only 1 Ag per unit cell can significantly improve the PSA separation. Cost estimates have been made that showed that a significant gain can be obtained by the use of Ag-containing Li-LSX zeolite. 

A general comparison can be made of these two types of sorbents. The adsorption of olefins under practical conditions is limited by the pore volume of the sorbent, that is, adsorption at pressures above ambient or adsorption at near room temperature and atmospheric pressure. The pore volumes of zeolites and molecular sieves are substantially lower than that of the jr-complexation sorbents, which are based on silica gel and activated alumina Thus, for propylene, the limiting adsorbed amounts for zeolites and molecular sieves are approximately 2.1-2.4 mmol/g, whereas that for the 7r-complexation sorbents supported on silica gel is weU over 5 mmol/g. As will be shown below, direct comparisons of the PSA separation performances with these two types of sorbents show indeed that the r-complexation sorbents are significantly better than zeolites and molecular sieves. 

This comparison is a direct reflection of the relative pore volumes. The tendency of the N2 isotherm on Sr-ETS-4 to form a plauteau at above 5 atm is also undesirable for PSA separation, particularly when a high feed pressure of natural gas is available (which is usually the case). 

Figure 12. Recovery and purity versus the iso fraction purge to feed ratio for a PSA separation of n/iso-paraffins based on patent data of Minkkinen et al. [3]. Simulations are performed with equilibria, kinetic and a fixed bed model shown in this work.

The chapter concludes with a brief discussion of PSA separation based on the difference of adsorption rates. 

Adsorptive separation is based on the difference of mixture components in the equilibrium adsorption, rate of adsorption or shape and/or size. The size and molecular weight of the two gases are quite close, and their physical or chemical property is also similar, therefore, the difference in the equilibrium adsorption must be somehow enlarged. Enlightened by the monolayer adsorption mechanism, the author s lab successfully enlarged the separation coefficient for several times [33]. As is shown in Fig. 8, the separation coefficient correlates with the specific surface area of adsorbents linearly. Recently, the feasibility of the PSA separation was further proved by a continuous run on a two-column process in the authors laboratory. Its practical application in the future will certainly have an important consequence. 

---