Pascaline Pré

Quantitative structure–activity relationships for the prediction of VOCs adsorption and desorption energies onto activated carbon

The aim of the study is to investigate quantitative relationships to predict the energetic interactions resulting from either adsorption or desorption of volatile organic compounds (VOCs) onto granular activated carbon. For that purpose, an experimental database was first built. Heats of adsorption and desorption were determined onto one activated carbon material for a 40 VOCs panel. The measurements were performed using differential scanning calorimetry coupled to thermogravimetry analysis. Adsorption energies were found to range between 40 and 80 kJ mol−1, whereas the desorption energies appear to be about 16% higher. Multiple linear regressions were afterwards tested in order to relate energies data with VOCs molecular properties. In a first approach, physical and chemical properties of the organic compounds were selected to investigate the best correlations. From the results obtained, the main influence of the ionization potential and of the polarisability were enlightened. In a similar way, connectivity molecular indexes were used. Some additional information were thus provided, which demonstrated the influence of the molecular shape, its branching and the steric hindrance.

Different families of volatile organic compounds pollution control by microporous carbons in temperature swing adsorption processes

In this research work, the three different VOCs such as acetone, dichloromethane and ethyl formate (with corresponding families like ketone, halogenated-organic, ester) are recovered by using temperature swing adsorption (TSA) process. The vapors of these selected VOCs are adsorbed on a microporous activated carbon. After adsorption step, they are regenerated under the same operating conditions by hot nitrogen regeneration. In each case of regeneration, Factorial Experimental Design (FED) tool had been used to optimize the temperature, and the superficial velocity of the nitrogen for achieving maximum regeneration efficiency (R(E)) at an optimized operating cost (OP(€)). All the experimental results of adsorption step and hot nitrogen regeneration step had been validated by the simulation model PROSIM. The average error percentage between the simulation and experiment based on the mass of adsorption of dichloromethane was 3.1%. The average error percentages between the simulations and experiments based on the mass of dichloromethane regenerated by nitrogen regeneration were 4.5%.

Visualization of the exothermal VOC adsorption in a fixed-bed activated carbon adsorber

Activated carbon fixed beds are classically used to remove volatile organic compounds (VOCs) present in gaseous emissions. In such use, an increase of local temperature due to exothermal adsorption has been reported; some accidental fires in the carbon bed due to the removal of high concentrations of ketones have been published. In this work, removal of VOCs was performed in a laboratory-scale pilot unit. In order to visualize the increase in local temperature, the adsorption front was tracked with a flame ionization detector and the thermal wave was simultaneously visualized with an infrared camera. In extreme conditions, fire in the adsorber and the combustion of activated carbon was achieved during ketone adsorption. Data have been extracted from these experiments, including local temperature, front velocity and carbon bed combustion conditions.

Recovery comparisons—Hot nitrogen Vs steam regeneration of toxic dichloromethane from activated carbon beds in oil sands process

The regeneration experiments of dichloromethane from activated carbon bed had been carried out by both hot nitrogen and steam to evaluate the regeneration performance and the operating cost of the regeneration step. Factorial Experimental Design (FED) tool had been implemented to optimize the temperature of nitrogen and the superficial velocity of the nitrogen to achieve maximum regeneration at an optimized operating cost. All the experimental results of adsorption step, hot nitrogen and steam regeneration step had been validated by the simulation model PROSIM. The average error percentage between the simulation and experiment based on the mass of adsorption of dichloromethane was 2.6%. The average error percentages between the simulations and experiments based on the mass of dichloromethane regenerated by nitrogen regeneration and steam regeneration were 3 and 12%, respectively. From the experiments, it had been shown that both the hot nitrogen and steam regeneration had regenerated 84% of dichloromethane. But the choice of hot nitrogen or steam regeneration depends on the regeneration time, operating costs, and purity of dichloromethane regenerated. A thorough investigation had been made about the advantages and limitations of both the hot nitrogen and steam regeneration of dichloromethane.

Modeling the Temperature Dependence of Adsorption Equilibriums of VOC(s) onto Activated Carbons

In order to optimize the efficiency of the removal of volatile organic compounds (VOCs) by adsorption onto activated carbon beds, process simulations taking into account exothermicity effects are helpful. Significant temperature increases may arise in the bed during the VOC adsorption cycle, especially when high concentrations have to be treated. Consequently, reliable and easy-to-handle isotherms remain a key hurdle to build realistic models. In this study, adsorption models were tested to describe a set of experimental data obtained for three VOCs (acetone, ethyl formate, and dichloromethane) adsorbed onto five commercial activated carbons at four different temperatures (20, 40, 60, and 80°C ). A new expression of the Freundlich equation [ qe = ( a1 T+ a2 T2 ) Ce ( 1/ nf ) ] was shown to be statistically the most efficient to describe the adsorption isotherms of VOCs, single or in mixtures. A second-order polynomial temperature-dependence was introduced in this expression. The so-adapted Freundlich rela...

Characterization and Selection of Materials for Air Biofiltration in Fluidized Beds

A methodology is described for the selection of the best material to employ in a fluidized biofilter applied to volatil organic compounds (VOC) treatment. Two different supports were considered: one natural, scrap wood, and one synthetic, polyurethane foam. In a first part, the main objective was to establish the fluidizability of the selected solids. Particles of increasing moisture content were tested in rigs equipped with different air distribution technologies. Maps of flow regimes were drawn by varying air velocity and bed height. In the second part, the sorption capacities of the selected materials were evaluated at different humidity levels for a soluble organic compound (ethanol) and an insoluble one (toluene). Their ability for supporting viable microorganisms was then measured. Fine particles with 45% moisture content fluidized with the tuyere distributor in channeling and slugging regimes, whereas the conical distributor generated a spouted bed regime. Coarse wood particles with 45% moisture commonly used in fixed bed biofilters were considered as less suitable due to limitations borne by their fluidization regimes. Finally polyurethane foam filled with Agar gel could not endure the mechanical stress to which particles are subjected in fluidized beds. It was shown that a 45% moisture content was optimum to ensure transfer for both soluble and hydrophobic compounds. Considering the biological criterion, it appeared that a 45% humidity level was sufficient to ensure biomass growth and biodegradation of pollutants on fine scrap wood particles.

Hazardous dichloromethane recovery in combined temperature and vacuum pressure swing adsorption process.

Organic vapors emitted from solvents used in chemical and pharmaceutical processes, or from hydrocarbon fuel storage stations at oil terminals, can be efficiently captured by adsorption onto activated carbon beds. To recover vapors after the adsorption step, two modes of regeneration were selected and could be possibly combined: thermal desorption by hot nitrogen flow and vacuum depressurization (VTSA). Because of ignition risks, the conditions in which the beds operate during the adsorption and regeneration steps need to be strictly controlled, as well as optimized to maintain good performances. In this work, the optimal conditions to be applied during the desorption step were determined from factorial experimental design (FED), and validated from the process simulation results. The regeneration performances were compared in terms of bed regeneration rate, concentration of recovered volatile organic compounds (VOC) and operating costs. As an example, this methodology was applied in case of dichloromethane. It has been shown that the combination of thermal and vacuum regeneration allows reaching 82% recovery of dichloromethane. Moreover, the vacuum desorption ended up in cooling the activated carbon bed from 93°C to 63°C and so that it significantly reduces the cooling time before starting a new cycle.

Global statistical predictor model for characteristic adsorption energy of organic vapors–solid interaction: Use in dynamic process simulation

Adsorption of Volatile Organic Compounds (VOCs) is one of the best remediation techniques for controlling industrial air pollution. In this paper, a quantitative predictor model for the characteristic adsorption energy (E) of the Dubinin-Radushkevich (DR) isotherm model has been established with R(2) value of 0.94. A predictor model for characteristic adsorption energy (E) has been established by using Multiple Linear Regression (MLR) analysis in a statistical package MINITAB. The experimental value of characteristic adsorption energy was computed by modeling the isotherm equilibrium data (which contain 120 isotherms involving five VOCs and eight activated carbons at 293, 313, 333, and 353 K) with the Gauss-Newton method in a statistical package R-STAT. The MLR model has been validated with the experimental equilibrium isotherm data points, and it will be implemented in the dynamic adsorption simulation model PROSIM. By implementing this model, it predicts an enormous range of 1200 isotherm equilibrium coefficients of DR model at different temperatures such as 293, 313, 333, and 353K (each isotherm has 10 equilibrium points by changing the concentration) just by a simple MLR characteristic energy model without any experiments.