| The emission of volatile organic compounds (VOCs) into the atmosphere is harmful to human health and environmental protection. Currently, there are two types of technologies for VOCs emission controlling:degradation (combustion, chemical oxidation and biological oxidation) and recovery (condensation, absorption, adsorption, membrane separation, etc.), among which adsorption has been widely used for many merits, such as high efficiency, stable work effect and low energy consumption. The most common adsorbent is activated carbon. However, due to its problems of difficult regeneration, hygroscopic, and short operating life, exploring new alternative adsorbents for VOCs recovery is of great significance. Hypercrosslinked polymeric resin (HPR), a new type of resin with great specific surface area, stable physical-chemical properties, and good regenerability, has been employed in many cases by this study group in recent years. However, before the design of an adsorpton unit, a large deal of experiments need to be conducted to gain the breakthrough curves under corresponding conditions so as to determine the key parameters, including dosage of HPRs, diameter, and height of adsorption columns, etc., consuming much time and energies. In order to provide theoretical basis and scientific guidance in designing VOCs recovery applying with HPRs, the mathematical models were established for adsorption of VOCs on HPRs in fixed bed in this study, and the predictive validity was verified with adorption experiments, based on the adsorption properties of HPRs. The main research and resulrts are as followed:1. Adsorption euilibrium data of 14 VOCs (pentane, hexane, heptane, benzene, dichloromethane, butanone, benzene, chlorobenzene, trichloromethane, dichloroethane, trichloroethene, acetone, butanone, methyl acetate, ethyl acetate, propyl acetate) onto 5 HPRs (HY-1, HPsorbent, ND-100, Resin-1, and Resin-2) under different temperatures were applied to predict the adsorption isotherms. The data were fitted with Dubinin-Radushkevch (D-R) equations, and the fitting values of qo (the volume adsorbed capacity per unit of adsorbent) and E (adsorption characteristic energy) were obtained. The result shows that the fittings are good, R2>0.9388; the fitting go values are widely different from the micropore volumes, with the relative errors range from 0.55% to 30.50%. Therefore, it is unreasonable to assume qo be equivalent to micropore volume of HPRs in the prediciton of equilibrium adsorption capacity using D-R equation.2. Multiple linear regression (MLR) method was applied to correlate qo and E in D-R equation with the properties parameters of adsorbents and adsorbates, in which specific surface area and micropore volume of HPRs and physical-chemical properties (molar polarizability, molar volume, parachor and ionization potential) of VOCs are independent variables, and the fitting qo and E are dependent variables. The result shows that the regression is highly significant, with both P=0.000<0.01: for qo. the significant coefficients in order are micropore volume of HPRs (0.361). ionization potential (0.0306), molar polarizability (0.0277), molar volume (0.0177), and parachor (-0.0103) of adsorbates; and for E they are, molar polarizability of adsorbates (0.275), and specific surface area of HPRs (0.00318); compared to the experimental values, the predicted qo and E have the relative errors range from-8.43% to 10.24% and from-12.75% to 10.78%, respectively, indicating high accuracy for the predicting model. On the basis of the corelation equations obtained for qo and E, the equilibrium adsorption capacities of pentane, hexane, and heptane on HY-1 were predicted employing D-R equations. The result shows that adsorption isotherms predicted by MLR method are very close to the experimental ones,83.3% of the predicted data have the relative errors below 10%, which is significantly superior to the conventional method, in which qo in D-R equation equals to the micropore volume of HPRs, and E is predicted by affinity coefficient method; while only 17%,25%, and 29% for three affinity coefficient methods (molar volume, parachor, and molar polarizability), respectively.3. The mass transfer model for isothermal adsorption in fixed bed was established by coupling the obtained D-R equations with mass balance and Linear Driving Force equation (LDF); then, the breakthrough curves for isothermal adsorption of alkanes on HY-1 were obtained by solving the equation set using Matlab Software. Comparing the predicted breakthrough curves with the experimental ones, we find that the predicted breakthrough curves are consistent with the experimental ones, and the relative errors of breakthrough time between predictions and experiments are from-7.28% to 5.80%. The sensitive parameters of this model include linear gas flowrate U, inlet concentration Co, colunm length L, qo, and E for predicting breakthrough time. Among these parameters, U, Co and L are clearly given by operational condition, thus, qo and E are the most sensitive facotors, which can be accurately predicted by MLR models. Compared with the convential model, in which qo equals to the micropore volume of HPRs, and E is predicted by affinity coefficient in solving the D-R equation, the breakthrough curves predicted by the model in this study are more consistent with experimental results.4. The mass and heat transfer model for quasi-adiabatic adsorption in fixed bed was established by coupling the obtained D-R equation with mass balance, heat balance, and LDF equation. The temperature rises and breakthrough curves for quasi-adiabatic adsorption of alkanes on HY-1 were obtained by solving the equation set using Matlab Software. The maximum temperature rise for pentane, hexane, and heptane during the quasi-adiabatic adsorption were 39.3℃,24.7℃ and 19.9℃, respectively. Compared with isothermal adsorption, the adsorption rate constants for quasi-adiabatic adsorption were smaller, and the breakthrough adsorption capacities were reduced due to the bed temperature rise. Comparing the predicted bed temperature rises and breakthrough curves with the experimental ones, we find that the predicted maximum temperature rises and the breakthrough curves were close to the experimental ones with all the relative errors below 20%. The sensitive parameters of this model include initial temperature To, linear gas flowrate U, inlet concentration Co, colunm length L, qo and E for predicting the temperature rise and breakthrough time. Among these parameters,T0, U, Co, and L are clearly given by operational condition, thus, qo and E are the most sensitive factors, which can be accurately predicted by MLR models. Compared with the conventional approach, in which qo equals to the micropore volume of HPRs, and E is predicted by affinity coefficient method in solving the D-R equations, the bed temperautre rises and breakthrough curves predicted by the proposed model in this study are more consistent with the experimental results.In conclusion, the MLR model suggested in this study is accurate enough to predict the parameters qo and E in D-R equations. By coupling the predicted D-R equations with mass balance, heat balance, and LDF equations, the mass and heat transfer model for fixed-bed adsorption was established, and the breakthrough curves of isothermal and quasi-adiabatic adsorption of alkanes on HPR can be well predicted. |
PDF Full Text Request