Regeneration Of Adsorbents By Using Liquid, Subcritical And Supercritical Carbon Dioxide
REGENERATION OF ADSORBENTS BY USING LIQUID, SUBCRITICAL AND SUPERCRITICAL CARBON DIOXIDE
H. Grajek
Military Technical Academy, Institute of Chemistry,
00-908 Warsaw, Kaliski str.2, Poland
Introduction
As it is commonly known gases fulfil the ideal gas laws only in narrow ranges of pressure, volume and temperature. At proper temperatures and under sufficiently high pressures all gases condense forming liquid phase. It is the excessive aspect of gases imperfection, which can be described by means of any cubic equation of state complying the following stability condition: . For higher temperatures, e.g. 323.15K isotherms (functions p=f(J,T)) have a typical hyperbolic shape (cubic equations of state) expected for an ideal gas. Deviations from hyperbolic shape appear during decreasing temperature. At 304.21K the inflexion point appears on the isotherm, and it is called the critical point: , , and Jc =0.0957 dm3/mol. Differences between gas (vapour) and liquid phases disappear at that point (condition). Subcritical region for CO2 is found in the triangular region formed by the melting curve, the boiling curve and the line that define the critical pressure. Subcritical liquid CO2 exists from 216.58 to 304.21 K and from 0.52 to 7.38 MPa. The critical region has its origin at the critical point. At this point we can define a supercritical fluid (SCF) as any substance (gas or liquid) that is above its critical temperature (Tc) and critical pressure (pc). The region below the critical temperature (T7.38 MPa) and below the melting point (T>216.58 K) is called the region of liquid state of carbon dioxide.
The purpose of writing this review is to obtain a greater understanding of the dynamic extraction behaviour of different adsorbates from adsorbents with supercritical CO2. The understanding of the process requires
characterization of adsorbent-SCF, adsorbent-adsorbate and adsorbate-SCF interactions. Such knowledge is important for the fundamental understanding, and design of environmental applications of SCF extraction such as desorption of analytes from adsorbents, and activated carbons regeneration. Activated carbon has been most often selected as the solid matrix because it is well defined and has uniform properties. The properties such as organic content, particle size, arborescent pore structure, and specific surface area differ significantly between poor-porous adsorbent and the varying influences of these properties can interfere with the interpretation of the achieved results. It will be very difficult to establish whether the observed behaviour will due to the adsorbent or to the desorption conditions.
Adsorption of Supercritical Gases onto Porous Adsorbents
Supercritical fluids adsorb onto carbon and non-carbon adsorbents and even porous polymers, altering both the chemical and physical properties of them.
The supercritical gases having no concept of the saturated vapour pressure cannot be adsorbed on the flat surface, macropores, and even mesopores with physical adsorption. The adsorption of vapour in micropores, which is called micropore filling, is enhanced at a very low pressure region due to overlapping of the molecule and pore-wall interactions. The deep potential well of the micropore gives rise to even adsorption of the supercritical gas to some extent. That is, the quite narrow micropore whose width is commensurated with the size of the adsorbate molecule has a very deep molecular potential well and is effective even for adsorption of supercritical gas. Accordingly, the supercritical gas is not an objective gas for predominant micropore filling. In case of micropore filling of vapour, the potential well of the micropore whose width is even more than trilayer thickness of the molecular size is deep enough, so that the vapour molecules can be sufficiently filled at a low pressure region without any obstacle for the intrapore diffusion. It is inconvenient, because it hinders the micropore penetration by SCF and being in contact with the adsorbate molecules which may be extracted from the micropore interior. It may have a significant influence on the removing efficiency of molecules adsorbed in micropores.
Applications of Liquid, Subcritical,
and Supercritical CO2 as Extraction Solvent
Extraction of solutes from solid matrices takes place through four different mechanisms:
– if there are no interactions between the solute and the solid phase (an individual or a simple mixture) the process is simple dissolution of the solute in a suitable solvent which does not dissolve the solid matrix;
– if there are interactions between the solid and the solute, then the extraction process is desorption in the presence of the solvent and the adsorption isotherm of the solute on the solid in presence of the solvent determines the equilibrium;
– the third mechanism is swelling of the soil phase or the destruction of solid texture by the solvent accompanied by the extraction of the entrapped solute through the first two mechanisms;
– the fourth mechanism is reactive extraction, where the insoluble solute reacts with the solvent and the reaction products are soluble hence extractable.
In supercritical extraction the important aspect relates to solvent-solute interactions. Normally, the interactions between the soil and the solute determine the ease of extraction, i.e. the strength of the adsorption isotherm is determined by interactions between the adsorbent and the adsorbate.
Generally, the regeneration of an activated carbon consists of removing the adsorbed substances from its surface and in restoring, as far as possible, its initial adsorptive properties. The regeneration of activated carbons by using supercritical CO2 has received widespread attention over the past years. DeFilippi et al. investigated the adsorption and desorption of different pesticides (trifluralin, diazinon, alachlor, atrazine, carbaryl, pentachlorophenol, phenol and acetic acid) on different activated carbons, and carbonaceous resinous adsorbent Ambersorb XE-348 in supercritical CO21.
The ability to achieve a constant but lower adsorption capacity probably is an example of such relative forces representing the ability of SC CO2 to remove the less strongly held adsorbate molecules. The first regeneration cycle probably removes only the molecules held by weaker physical adsorption. The subsequent adsorptions and regenerations probably operate only on the available low-energy sites. The chemisorbed molecules are still present on the carbon surface. The process of removing the irreversibly bound adsorbate from surface was designated reactivation.
The key to designing large-scale supercritical desorption processes is understanding how the desorption is influenced by process variables such as pressure, temperature, and extraction solvent flow rate.
The earliest workers concluded that solubility limitations have not a decisive influence on supercritical desorption, and the shape of the adsorption isotherm has the dominant influence on the desorption.
To achieve higher desorption efficiencies, the DDT solubility at the extraction conditions needs to be increased. The increase of temperature and/or pressure was unlikely to be sufficient for achieving proper extraction conditions. An alternative method of raising the solubility would be through the addition of a polar cosolvent or a reactive one to the SC CO2.
It is well known that the adsorption is determined by the properties of the adsorbent and the adsorbate. The adsorbate must often compete for adsorption sites with other components of the solution, predominantly with water. The regeneration efficiency of activated carbon spent with toluene was 96% without water and 85% with water, but the initial rates of the desorption (indicated by the slope of the desorption profile) were similar for both trials. Water has apparently had a shielding effect for low concentrations of toluene. Efficiencies of 85% and 89% were observed for 2-chlorophenol in two separate trials6.
In most cases supercritical regeneration has been carried out by using supercritical CO2. A desorption of solutes from adsorbents can also be successfully carried out with liquid and subcritical CO2. The processes of regeneration of activated carbon and zeolite spent with buthyl acetate and xylenes by using liquid and SC CO2 were investigated by Vallee and Barth7. Majewski et al.8 investigated regeneration of activated carbon bed, activated carbon ceramic cartridge and zeolite spent with different organic solvents by using liquid, subcritical and SC CO2. There were no significant differences in the regeneration behaviour carried out by using the aforementioned extraction solvents.
Influence of Temperature on the Efficiency of Supercritical Fluid Regeneration of Activated Carbons
During the desorption of ethyl acetate from activated carbon with SC CO2 Tan and Liou noticed that regeneration efficiency decreased with temperature at fixed pressure5. The high efficiency was achieved in the liquid state of CO2 (8.8MPa, 300K) rather than in the supercritical region. But when the regeneration pressure was raised to 13.1MPa, more efficient regeneration was achieved in the supercritical region instead of in the liquid state region. Similar results were achieved by Srinivasan et al. at nearing regeneration conditions4. These results have suggested that, at some pressures higher than 13.1MPa5 or 16.3MPa4, the amount of ethyl acetate desorbed will increase with increasing temperature. This unconventional phenomenon was analogous to that of the solubility in a supercritical fluid. In the supercritical region, the solubility decreased with increasing temperature at low supercritical pressures.
Influence of Flow Rate of Supercritical Solvent and Particle Size
of Adsorbent on Regeneration Efficiency
Many experiments were carried out to examine the influence of SC CO2 flow rate on the desorption behavior of fixed SC CO2 density (pressure) and temperature. The achieved data show that when the flow rate fell the desorption became slower but not with respect to dimensionless time. Similar results were reported previously4,5. It attributed to an external film resistance that is significant at low flow rates and effectively slows down the rate at which the adsorbate can enter the bulk flowing fluid and be carried out of the bed.
Macnaughton and Foster plotted the fraction of DDT desorbed corresponding with the flow rate data as a function of the total number of CO2 moles passed through the bed3. At the highest flow rate, the equilibrium was not achieved, i.e. the effluent concentrations at each average loading were below the equilibrium concentration. This behavior indicated the presence of a mass-transfer resistance which was probably an intraparticle diffusion or an external film resistance. The desorption occurring at the lower flow rates were primarily influenced by adsorption equilibrium constraints rather than mass-transfer limitations. The results achieved by Srinivasan et al. exhibited similar trends4.
The Local-Equilibrium Theory of Fixed-Bed
The desorption curve profile is a function of the adsorption equilibrium relationship; in this case it is applied to isothermal, fixed-bed, piston flow desorption. The so-called local equilibrium exists at all points and all times between the particles and the adjacent SCF. Longitudinal diffusion along the axis of the adsorbent bed is neglected and the piston flow of the fluid is assumed.
DeFilippi et al.1 summarized some column profiles as the coverage versus distance down the adsorbent bed in column at various times during the regeneration process. Each curve presented there has been a cross-sectional “snap-shot” of the adsorbate remaining on the adsorbent bed at the designated time.
If the slope of adsorption isotherm is a constant for all solute concentrations (usually at low concentrations), the desorption profile predicted by the local equilibrium theory is perfect. For non-linear adsorption isotherm a more rigorous model involving the effects of mass transfer axial dispersion, and effective diffusivities is needed.
Costs of Supercritical Regeneration of Adsorbents
DeFilippi et al.1 estimated the cost of supercritical regeneration of activated carbon spent with phenol for the pilot plant on a daily basis. The plant was designed on the basis of the steady-state granulated activated carbon capacity. The estimated operating cost was about 0.17$ (1980 dollars) per kilogram of regenerated activated carbon. A similar estimation was carried out by Tomasko et al.6 on the pilot plant consisting of a three element desorber with two stage flash separation at the yield of a 24 ton of adsorbent per day. After the optimization of the process around minimizing the cost of recycling the SC CO2 through an efficient recompression scheme and a cycle configuration in the desorber unit the processing cost was about 0.23$ (1993 dollars) per kilogram of activated carbon.
Summary and Conclusions
The regeneration efficiency of adsorbents is determined by two competing effects, solvation and adsorption. The solvation is dictated by thermodynamic equilibrium and can be estimated by using an equation of state models for adsorbent-SCF phase equilibrium.
The cross-over effect for desorption (i.e. desorption decreases with increasing temperature) is usually observed during the SCF regeneration of activated carbons. The related cross-over effect is well-known for the equilibrium solubility of adsorbate (the solubility decreases with increasing temperature).
Supercritical fluid technologies will be better served if all scientists discuss the results of experiments in terms of density rather than pressure. The extraction solvent density in liquid, sub- and supercritical state has a decisive influence on the regeneration efficiency.
The apparent activation energy for desorption decreases with the increase of SC CO2 density. The higher density may enhance the solubility of an adsorbate in a SCF, but the higher viscosity may have an adverse effect on the diffusion rate.
The magnitude of the adsorbate loading lends credence to the assumption that adsorption-equilibrium limitations are responsible for the shape of the desorption curve and that the adsorbate solubility in SC CO2 does not represent the limiting step of the regeneration process.
A relatively higher desorption extent can be obtained in the liquid phase of CO2 rather than in the supercritical region.
The regeneration processes carried out with liquid, subcritical, and supercritical carbon dioxide behave similarly.
The increase of the flow rate of the regenerating fluid causes the decrease of desorption time required.
The susceptibility of activated carbons loaded with tert-butylbenzene to the SC CO2 regeneration decreases with the increase of micropore structure in the total pore structure of adsorbent, because of decreasing ability to penetrate the micropore structure by SCF.
The extent of SCF desorption increases with the decrease of the adsorbent particle size.
The presence of water on activated carbon surface slightly inhibits the efficiency of regeneration.
The rate of mass transfer, the axial dispersion, and the diffusion into the micropores are the engineering parameters needed for designing the regeneration system. The prediction of the solvent behaviour of compressed gases is difficult because of the limitation (sometimes even the lack !) of suitable scientific knowledge.
The desorption profiles predicted by the local equilibrium theory are perfect only for the adsorption systems characterizing a linear adsorption isotherm. In cases of non-linear adsorption isotherm, a more rigorous extraction model involving the effects of mass transfer, the axial dispersion, and the effective diffusivities have to be taken into account.
The regeneration cost of 1 kilogram of spent activated carbon by using SC CO2 as the extraction solvent is around 0.23$ at conditions of the continuous process. The adsorbate (contaminant) properties influence the SCF regeneration economy more strongly than the adsorbent properties.
Acknowledgments
This work was partly supported by the Polish State Committee for Scientific Research, Grant No. 3 T09B 036 16.
References
1. DeFilippi, R.P., Krukonis, V.J. Robey, R.J. and Modell, M. Raport EPA-600/2-80-054, Washington DC, 1980.
2. DeFilippi, R.P. Chem. Ind., 1982, 12, 390.
3. Macnaughton, S.J. and Foster, N.R., Ind. Eng. Chem. Res., 1995, 34, 275.
4. Srinivasan, M.P., Smith, J.M. and McCoy, B.J., Chem. Eng. Sci., 1990, 45(7), 1885.
5. Tan, C.-S. and Liou, D.-C., Ind. Eng. Chem. Res., 1988, 27, 988.
6. Tomasko, D.L., Hay, K.J., Leman, G.W. and Eckert, C.A., Environ. Prog., 1993, 12(3), 208.
7. Vallee, G. and Barth, D., Proc. of the 4th Meeting on Supercritical Fluids, Institut National Polytechnique DeLorraine, Villeurbanne, 1997, 99.
8. Majewski, W., Perrut, M. and Goupy, M., Proc. of the 4th Meeting on Supercritical Fluids, Institut National Polytechnique DeLorraine, Villeurbanne, 1997, 111.
Related articles::