DC Corona Electric Discharges for Air Pollution Control. Part 1

Jul 14, 2007 - For both, the energy efficiency is significantly better with negative than with positive DC corona, especially in humid air. ... Citati...
1 downloads 10 Views 259KB Size
Environ. Sci. Technol. 2007, 41, 5862-5868

DC Corona Electric Discharges for Air Pollution Control. Part 1. Efficiency and Products of Hydrocarbon Processing ESTER MAROTTA,† ALESSANDRO CALLEA,† MASSIMO REA,‡ AND C R I S T I N A P A R A D I S I * ,† Department of Chemical Sciences, Universita` di Padova, 35131 Padova, Italy, and Department of Electrical Engineering, Universita` di Padova, 35131 Padova, Italy

A large (ca 0.7 L) wire-cylinder benchtop reactor was developed and tested for DC corona processing of VOC (volatile organic compound)-contaminated air at room temperature and pressure. The aim of our research is the identification and rationalization of the chemical reactions responsible for VOC removal. Model hydrocarbons, n-hexane and 2,2,4-trimethylpentane (i-octane), were used to characterize the process and compare the effects of DC corona polarity and of humidity on its energy efficiency and products. n-Hexane and i-octane behave very similarly. For both, the energy efficiency is significantly better with negative than with positive DC corona, especially in humid air. The effect of humidity is most interesting. Thus, while with -DC corona the process efficiency is significantly better in humid air, a slight inhibition is observed with +DC corona. Differences between +DC and -DC corona are also found in the amounts of volatile products formed, which include CO2, CO, and minor quantities of organic byproducts (aldehydes, ketones, alcohols, and lower hydrocarbons). A significant fraction of the carbon originally present as VOC is, however, unaccounted for by the analysis of gaseous and volatile organic products and must, therefore, end up as nonvolatile materials and aerosols.

Introduction Among the novel technologies for air-pollution control, nonthermal plasma (NTP)-based processes are gaining an increasingly important role (1-8). In NTPs, which are conveniently produced by corona discharges in a gas at room temperature and pressure, temperature (i.e., kinetic energy) is not in thermal equilibrium and differs substantially between the electrons, typically in the range 1-20 eV, and the more massive species (neutral molecules, radicals, atoms, and ions) which remain at near room temperature. The highly energetic electrons interact with molecules of the bulk gas, i.e., N2 and O2 in the case of air, and induce excitation, dissociation, and ionization leading to excited molecules, atoms, and ions. In * Corresponding author phone: ++39 049 827 5661; fax: ++39 049 827 5239; e-mail: [email protected]. † Department of Chemical Sciences. ‡ Department of Electrical Engineering. 5862

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 16, 2007

addition, thermalized electrons produce negative ions via electron attachment, an important process in air which leads to the formation of O2- and O3- (9). Removal of VOC molecules (typically present in the hundreds to thousands ppm range), is generally attributed to radical reactions initiated by the attack of O or HO• (Scheme 1a): the VOCderived radicals thus produced undergo oxidation according to a scheme similar to that established for their tropospheric degradation (10) (Scheme 1b). Subsequent analogous steps lead eventually to exhaustive oxidation to CO2. The involvement of ionic initiation steps has also been proposed as an alternative to 1a (11-13). It was noted that Scheme 1b could still apply since VOC-derived ions undergo fragmentation to form smaller ions and radicals (13, 14). In any case, selective input of energy into VOC molecules occurs, leading to their oxidation at room temperature and pressure. This remarkable feature of NTP based processes makes them particularly attractive and advantageous for air purification which typically requires the treatment of large volumes contaminated by pollutants in low concentration. Different types of reactors and power supplies have been developed for the production of NTPs by means of electrical discharges for air pollution control (1). Recognized merits of such reactors include mild operation conditions (room temperature and pressure) and easy maintenance. While the early research efforts concentrated on the improvement of the process energy and cost efficiencies, emphasis has recently focused on the products being released (1, 5, 15). It is now well-known that high removal efficiencies are not always accompanied by high CO2 selectivities. Undesirable byproducts include NOx, ozone, volatile and nonvolatile VOCs, and nanosized particles (5, 15). A better characterization of the underlying chemistry (products, kinetics, structure vs reactivity relationships, and mechanisms) is necessary to develop a more efficient and cleaner process. Quantitative experimental studies require stable and reproducible experimental conditions and the possibility to change one process variable at a time without affecting other system variables. To this end we built a simple wire-cylinder benchtop corona reactor and tested it with two model hydrocarbons, n-hexane and 2,2,4-trimethylpentane (i-octane), both important gasoline components. The unusually large size (0.7 L) of the reactor we developed for this project represents a reasonable compromise between two opposite constraints: small enough for quantitative determinations necessary for fundamental studies and large enough for extrapolation of the results to industrial applications. It is well-known that scaling up is a problem for corona processing since the physical phenomena and the ensuing corona regime are not linearly related to the geometrical setup parameters, notably the interelectrode gap. We used DC corona, which, although certainly not the most energy efficient, is well modeled and fit to produce stable and reproducible plasma conditions (16). To avoid interferences by any contaminants of ambient air, we used certified synthetic mixtures both for pure air (80% nitrogen and 20% oxygen) and for VOC-containing air. Because of the high costs of such mixtures, all experiments were run with a fixed VOC concentration of 500 ppm, a typical mixing ratio for NTPs applications.

Experimental Section Reactor, Power Supply, and Gas Line. Figure 1 shows a schematics of the plasma reactor (Figure 1a) and of the gas line (Figure 1b). The reactor is a 3.85 cm × 60 cm (0.698 L) stainless steel cylinder set at ground potential, with a stainless 10.1021/es0707411 CCC: $37.00

 2007 American Chemical Society Published on Web 07/14/2007

FIGURE 1. Schematics of the experimental setup. (a) Wire-cylinder corona reactor. (b) Gas flow line. steel wire (the energized electrode) of 1 mm diameter fixed along its axis. The stainless steel cylinder, which has a window (1 × 10 cm) for the observation of the corona luminescence, is contained into a Pyrex glass cylinder of slightly larger diameter securely held in place by two Teflon caps connected by steel guys and viton o-rings. A description and a schematic (Figure 1S) of the electrical power supply circuit are given as Supporting Information. Pure air or VOC-containing air from appropriate gas cylinders is allowed into the reactor at a controlled flow rate (150-800 mL‚min-1, flowmeter 1) through a 3 mm i.d. Teflon tube. For experiments with humid air, a portion of the air flow, set by flowmeter 4, goes through a bubbler filled with deionized water to an appropriate level, 5. The humidity resulting when the two portions of the flow are newly mixed is measured by an hygrometer (Rotronic A1H), 6. Samples of the treated gas for off-line chemical analyses are collected with a gastight syringe from a glass reservoir, 7, connected to the gas outlet. The air tightness of the system is periodically checked by filling it with air at a pressure slightly higher than 1 bar and by verifying the absence of any pressure drop over several minutes with a manometer, 3. Chemicals. Air was a synthetic mixture (80% nitrogen and 20% oxygen) from Air Liquid with specified impurities of H2O (