Preparation of a carbon molecular sieve and application to separation of N2, O2 and CO2 in a fixed bed

Preparation of a carbon molecular sieve and application to separation of N₂, O₂ and CO₂ in a fixed bed

J.L.Soares; H.J.José; R.F.P.M.Moreira

Department of Chemical and Food Engineering, Federal University of Santa Catarina, Campus Universitário, Trindade 88040-670, Florianópolis, SC, Brazil

^{Address to correspondence}

ABSTRACT

The emission of CO₂ from power plants that burn fossil fuels is the major cause of the accumulation of CO₂ in the atmosphere. The separation of CO₂ from CO₂/air mixtures can play a key role in alleviating this problem. This separation can be carried out by using suitable adsorbents, such as carbon molecular sieves. In this work, a CMS was prepared by deposition of polyfurfuryl alcohol polymer on activated carbon. After deposition of the polymer, the material was carbonized at 800ºC for 2 hours. This material was used to separate O₂/N₂ mixtures and CO₂ in a fixed bed at room temperature. Experimental breakthrough curves obtained were fitted to theoretical models in order to establish the main mechanisms of mass transfer. The breakthrough curves showed that it is possible to separate O₂, N₂ and CO₂. The shape of the breakthrough curves was not influenced by the total flow, indicating that the gas contact for the gas mixture was good. The experimental data were fitted to theoretical models and it was established that the main mechanism of mass transfer was intraparticle diffusion.

Keywords: gas separation, carbon molecular sieve, activated carbon.

INTRODUCTION

The emission of CO₂ from power plants that burn fossil fuels is the major cause of the accumulation of CO₂ in the atmosphere, and this causes long-range environmental problems. Separation can play a key role in alleviating this problem. Typical flue gases contain around 17% CO₂, the balance being N₂ (79%) and O₂ (4%). Trace amounts of SO₂ and NO_x can also be found, but they are usually less than 1% in total (Yang et al., 1993).

Carbon molecular sieves (CMS's) are valuable materials for the separation and purification of gas mixtures. CMS's consist of activated carbon with a uniform pore size distribution and a pore size of several ångströns (Casa-Lillo et al., 1998). In this material, gas separation is based on the different adsorption kinetics of the gases, though the amounts adsorbed at equilibrium may be similar (Casa-Lillo et al., 1998). The different methods used for the preparation of CMS's have as a final objective either the synthesis of a material with a homogeneous pore size distribution (i.e. mainly with narrow microporosity) or the modification of the existing porosity of a suitable activated carbon.

In this work, a CMS was prepared by deposition of polyfurfuryl alcohol polymer on activated carbon, and this material was used to separate N₂, O₂ and CO₂ in a fixed bed. Experimental breakthrough curves obtained were fitted to theoretical models in order to establish the main mechanisms of mass transfer.

MATHEMATICAL MODEL

The flow pattern in a fixed bed can be represented by the axial dispersed plug-flow model, according to a mass balance for an element in the column, for the basic differential equation governing the dynamic behavior. The rate of mass transfer to the solid was described using a linear driving force expression (Ruthven, 1984).

Overall mass balance (Ruthven, 1984):

Rate of mass transfer to the solid (Glueckauf, 1955):

Intraparticle mass transfer (Ruthven, 1984):

Boundary and initial conditions (Ruthven, 1984):

The evaluation of the auxiliary parameters for the simulation of the proposed model was made using an empiric correlation proposed in the literature. Molecular diffusivity (D_m) was calculated by the Fuller, Schettler and Giddings equation (Reid et al., 1977); Knudsen diffusivity (D_K) was calculated by the Knudsen equation (Ruthven, 1984). To estimate the axial mass dispersion coefficient (D_ax), the correlation proposed by Leitão and Rodrigues (1995) was used. The isotherms of adsorption for CO₂, O₂ and N₂ were obtained by Moreira et al. (2001).

Equations 1-4 were solved using the PDECOL (Madsen and Sincovec, 1979) package in the FORTRAN language, which is based on the method of orthogonal collocation of finite elements for partial differential equations. Normally, 100 elements are used in the calculations, and 60 sec are needed for the numeric simulation of the breakthrough curves.

EXPERIMENTAL DETAILS

Materials

Activated carbon (CA1, Norit, The Netherlands) was used as the precursor of a CMS that was prepared by deposition of polyfurfuryl alcohol polymer. After deposition of the polymer, the material was carbonized at 800ºC for 2 hours (Moreira et al., 2001).

The CO₂ was 99.99% and He 99.99% pure and they were supplied by White Martins Ltda. The physical properties of the activated carbon used as precursor and the synthesized CMS are shown in Table 1 (Moreira et al., 2001).

System of Adsorption

The adsorption system consists of two lines, one for circulation of the inert gas (He) as the carrier gas or desorption gas and the other line for circulation of the adsorption gas, CO₂ or compressed air (N₂ and O₂). The experimental apparatus is shown in Figure 1. Each line has two flowmeters to fit the flow, and the two lines of gas join and enter into the adsorption column with the CMS. The flow was also measured at the end of the column, using a flowmeter of bubbles in order to check the value of the flow.

At certain times, a 1.0 mL sample of the eluted gas from the column was analyzed by gas chromatography. To avoid the effect of variation in pressure in the gas line during the sampling, an Erlenmeyer with water was placed at the end of the system. The samples of gases eluted from the fixed bed were analyzed in a CG-35 chromatograph using a thermal conductivity detector (TCD) and helium as the carrier gas. The column temperature: Porapak-Q column (for N₂ and O₂ analysis) and Molecular Sieve 5A column (for CO₂ samples) was 120^oC.

Two runs for pure CO₂ were carried out under different operating conditions: 1.5 atm total pressure and 16 cm³/min flow at 15^oC; and 3.0 atm total pressure and 12 cm³/min flow at 19^oC. The breakthrough experiment for air separation (bicomponent system) was carried out only at a total pressure of 2.0 atm, and the flow was 41.4 cm³/min at 16^oC.

RESULTS AND DISCUSSION

Figure 2 shows the experimental results obtained from the adsorption bed with CMS's at ambient temperature (approximately 16^oC) for different flow and pressure systems of CO₂ and air (N₂ and O₂).

We may observe that there is a prolonged time of retention for CO₂ due to the high adsorption capacity of these gases on the CMS. Desorption is very slow, as shown in Figure 2, because when desorption with the same flow is attempted a considerable delay occurs. However, in Figure 2B a short retention time is observed for the adsorption of N₂ and O₂. It is notable that there is no chromatographic separation of the two gases, and the same breakthrough is seen. Table 2 shows the parameters used for the simulation. The simulation for air samples should be for a bicomponent system, but owing to the fact that the N₂ and O₂ isotherms are approximately the same and were also linear, the simulation actually used was that for a monocomponent system; in other words, all the simulations in this paper assume a single adsorbed component.

In Figure 3, we have the experimental data predicted by the models proposed in Eqs. (1-4), while in Table 3 we have the values of the parameters obtained through the empiric correlations that were applied in the numeric simulation of the proposed model in Eqs. (1-4). From the results, it can be affirmed that the models predict the experimental data, confirming the validity of the proposed models.

In relation to the results shown in Table 3, it was observed that the mass transfer in the film and the superficial diffusivity were negligible. A strong influence of the molecular diffusion was observed in the resistance to the diffusion inside the particles. The values for the Peclet number were very close, with an average of 86.5 for all the experiments, and thus they should be considered to influence column dispersion, which confirms the expected results due to the low flow values.

CONCLUSIONS

The large capacity of adsorption for CO₂, as seen by the retention time in the bed and low values of adsorption for N₂ and O₂, were achieved. The CMS is good for gaseous separation with CO₂, but is insufficient for separation of O₂ and N₂. The models represent the experimental data well. They show the strong influence of intraparticular diffusion, which is specifically controlled by molecular diffusion. The significant dispersion values are also proven by the Peclet number obtained for each breakthrough curve.

NOMENCLATURE

Greek Letters

ACKNOWLEDGMENT

José Luciano Soares is a Ph.D. student supported by the Capes Foundation (Portugal University work) and by CNPq (Brazil).

Received: March 5, 2002

Accepted: August 21, 2002

• Casa-Lillo, M. A., Monge, J. A., Pinero, E. R., Amorós, D. C., Linares-Solano, A. (1998). Molecular Sieve Properties of General Purpose Carbon Fibers. Carbon., 36, 1353.
• Glueckauf, E. (1955). Theory of Chromatography. Part 10. Formulae for Diffusion Into Spheres, Theory of Chromatography. Trans. Faraday Soc., 51, 1540.
• Grande, C. A., Rodrigues, A. E. (2001). Adsorption Equilibria and Kinetics of Propane and Propylene in Silica Gel. Ind. Eng. Chem. Res., 40, 1686.
• Leitão, A., Rodrigues, A. E. (1995). The Simulation of Solid-Liquid Adsorption in Activated Carbon Columns Using Estimates of Intraparticle Kinetic Parameters Obtained from Continuous Stirred Tank Reactor Experiments. Chem. Eng. Journal, 58, 239.
• Madsen, N. K., Sincovec, R. F. (1979). PDECOL: General Collocation Software for Partial Differential Equations. ACM Trans. Math. Software, 5, 326.
• Maleq, A., Farooq, S. (1995). Determination of Equilibrium Isotherms Using Dynamic Column Breakthrough and Constant Flow Equilibrium Desorption. J. Chem. Eng. Data, 40, 25.
• Maleq, A., Farooq, S. (1996). Comparison of Isotherm Models for Hydrocarbon Adsorption on Activated Carbon. AIChE J., 42, 3191.
• Moreira, R. F. P. M., Rodrigues, A. E., José, H. J. (2001) Modification of Pore Size in Activated Carbon by Polymer Deposition and Its Effects on Molecular Sieve Selectivity. Carbon, 39, 2269.
• Reid, C. R., Prausnitz, J. M., Sherwood, T. K. (1977). The Properties of Gases and Liquids. McGraw-Hill Book Company. Third Edition.
• Ruthven, D. M. (1984). Principles of Adsorption & Adsorption Processes. John Wiley & Sons.
• Silva, V. M. T. M., Rodrigues, A. E. (2001). Diethylacetal Synthesis with Acid Resin Catalyst: Dynamics of a Fixed Bed Adsorptive Reactor. Nascre 1 Conference. Houston, Texas, USA.
• Xiu, G., Li, P. (2000). Prediction of Breakthrough Curves for Adsorption of Lead (II) on Activated Carbon Fibers in a Fixed Bed. Carbon, 38, 975.
• Yang, R. T., Kikkinides, E. S., Cho, S. H. (1993). Concentration and Recovery of CO₂ from Fuel Gas by Pressure Swing Adsorption. Ind. Chem. Eng., 32, 2714.

Publication Dates

• Publication in this collection
  19 Mar 2003
• Date of issue
  Mar 2003

History

• Accepted
  21 Aug 2002
• Received
  05 Mar 2002