Print version ISSN 0104-6632
Braz. J. Chem. Eng. vol.20 no.1 São Paulo Jan./Mar. 2003
Nonthermal plasma reactors for the production of light hydrocarbon olefins from heavy oil
G.PrietoI; M.OkumotoI; K.TakashimaI; S.KatsuraI; A.MizunoI; O.PrietoII; C.R.GayII
IDepartment of Ecological Engineering, Toyohashi University of Technology Tempaku-cho, Toyohashi, Aichi, 441-8580, Japan
IIDepartment of Chemical Engineering, National University of Tucumán, Argentine Republic, Av. Independencia 1800, (4000), San Miguel de Tucumán, Argentine Republic
During the last decade, nonthermal plasma technology was applied in many different fields, focusing attention on the destruction of harmful compounds in the air. This paper deals with nonthermal plasma reactors for the conversion of heavy oil into light hydrocarbon olefins, to be employed as gasoline components or to be added in small amounts for the catalytic reduction of nitrogen oxide compounds in the treatment of exhaust gas at power plants. For the process, the plate-plate nonthermal plasma reactor driven by AC high voltage was selected. The reactor was modeled as a function of parameter characteristics, using the methodology provided by the statistical experimental design. The parameters studied were gap distance between electrodes, carrier gas flow and applied power. Results indicate that the reactions occurring in the process of heavy oil conversion have an important selective behavior. The products obtained were C1-C4 hydrocarbons with ethylene as the main compound. Operating the parameters of the reactor within the established operative window of the system and close to the optimum conditions, efficiencies as high as 70 (ml/joule) were obtained. These values validate the process as an in-situ method to produce light olefins for the treatment of nitrogen oxides in the exhaust gas from diesel engines.
Keywords: nonthermal plasma reactors, heavy oil conversion, light olefins production.
In the last few years, nonthermal plasma has been applied in many different fields, such as the destruction of harmful compounds in the air and synthesis of products which otherwise require high temperatures and pressures.
This paper deals with the conversion of heavy oil (high molecular weight hydrocarbons) into light hydrocarbons for use, for example, as gasoline components for the new generation of engines in the so-called "hybrid process" or to be added in small amounts to certain catalytic processes for reduction of nitrous oxide compounds in the treatment of exhaust gas from power plants and vehicles. The mathematical models of two nonthermal plasma reactors were developed to evaluate their individual effectiveness in carrying out the conversion of heavy oils into light hydrocarbons. The parameters examined were gap distance between electrodes, carrier gas flow and applied power. Both reactors were modeled empirically applying the techniques provided by the statistical experimental design. The models give us information about the effectiveness of each reactor in accomplishing the proposed goal.
Empirical models are very useful, especially when the phenomenon under study is complicated and not sufficiently understood. Frequently, complete theoretical models are not available and sometimes any attempt to develop one could not be justified from an economic point of view (Box et al., 1978).
The results obtained indicate that the reactions occurring in this process have an important selective behavior. The major byproducts obtained were hydrocarbons with one, two, three and four atoms of carbon, such as CH4, C2H6, C2H4, C3H6 and C4H10, with ethylene as the main compound.
The two nonthermal plasma reactors selected for this study are:
(a) Packed-Bed Plasma Reactor
Figure 1 shows the schematic diagram of the packed-bed reactor.
The packed-bed plasma reactor consists of a tubular reactor made of an acrylic resin tube of 50 mm in inner diameter. Two mesh electrodes pack the bed with almost spherically shaped ferroelectric particles between them. The gas flow enters at the ground-side electrode and exits at the high-voltage-side electrode. An amplified 60 Hz line AC voltage is applied to the reactor to generate partial or spark discharges (Yamamoto et al., 1996).
The selected variables and their corresponding ranges for the experimental runs are gas flow (mL/min): 50 400, gap distance (cm): 0.50 3.00 and input power (W): 5 25.
The reactor is filled with BaTiO3 pellets 3 mm in diameter (dielectric constant eS=4500) up to the corresponding gap distance. Before packing the bed, the amount of pellets already prepared is moistened with 3 to 5 droplets of heavy oil, depending on the bed depth. For every experimental run, pellets are changed to new ones.
Argon is employed as carrier gas. The gas flow is established and the carrier gas is flowed through the reactor during 5 minutes. Then power is applied during 2 minutes without sampling, and after that a gas sample is collected during 3 minutes in the case of the higher flow rates or during 5 minutes in the case of the lower flow rates.
(b) Plate-Plate Plasma Reactor
Figure 2 shows a schematic diagram of the plateplate reactor
The plate-plate plasma reactor consists of an acrylic resin tube with an inner diameter of 50 mm with two stainless steel parallel mesh electrodes with a diameter of about 4 cm. The upper electrode is connected to the high voltage. The high voltage is supplied by a neon transformer of 9 kV (peak to peak), AC current (110V, 50/60 Hz).
The selected variables and their corresponding ranges for the experimental runs are gas flow (mL/min): 50 400, gap distance (cm): 0.30 1.00 and input power (W): 3 12.
For this reactor, a glass fiber sheet imbibed in heavy oil is placed on the lower electrode (covering it wholly), which is connected to the ground. Carrier gas (argon) is flowed during 5 minutes. Voltage is applied during 1 minute, and then a sample is collected during 2 minutes for the higher flow rates or during 4 minutes for the lower flow rates. For every experimental run the glass fiber sheet is replaced.
The performance of each reactor is registered by efficiency (E), expressed in units of mL/J.
Efficiency is defined as the ratio between the total gas flow of carbon compounds and the corresponding input power.
The total gas flow of carbon compounds is calculated after measuring the C compounds sampled at the outlet of the plasma reactor. Gas chromatography is the technique employed to analyze the samples at the inlet and outlet of plasma reactors. The GC is equipped with an FID detector and a G-950 column (length 40 m, film thickness 25 mm and inner diameter 1.2 mm).
A common situation in experimental research is to identify the conditions of experimentation, which are most desirable depending upon some preselected criterion, such as for example, the determination of optimum operating conditions of a reactor. One possible way is to use the experimental information to construct mathematical models capable of representing the actual state of the process. These models must correlate target parameters, such as efficiency or yield, with those variables that may exert influence over them. Then, the study is oriented to acquire experimental data, to select the appropriate model and to estimate the parameters involved in the model.
A methodology that involves mathematical and statistical techniques appropriate to the model to analyze how a target response (selected response of interest) is influenced by many variables is the response surface methodology. When the mathematical form of the function is unknown, the ratio between the variables under study and the selected response can be adjusted to a polynomial function. These polynomials can be best adjusted by the central composite design. This design lends itself to sequential experimentation. The fundamentals and underlying philosophy of these techniques are discussed in many papers and textbooks (Montgomery, 1991; Davies, 1979; Myers, 1971).
ANALYSIS OF DATA AND DISCUSSION
The experimental results were evaluated according to the statistical experimental design methodology.
Packed-Bed Plasma Reactor
The mathematical model for efficiency is given by Eq. (1). To simplify the calculations, the independent variables were coded at a (-1, +1) interval.
G = gap distance,
FR = gas flow rate and
IP = input power.
Figure 3 shows a three-dimensional model for efficiency.
Analysis of the experimental results showed that variable gas flow has no meaningful effect on efficiency when it is increased from 120 to 330 mL/min.
There is an interaction between the variables input power and gap distance. The significance of the interaction is that the effect of these variables cannot be interpreted separately because the effect caused by one of the variables depends on the value set for the other variable. The interaction occurs because at the highest value of input power (21 W) an increase in the gap distance (from 1.0 to 2.5 cm) increases efficiency from 3.2 to 4.3 (mL/J) on average. But at the lower value of input power (9 W) the same increase in gap distance causes a decrease in efficiency, from 3.2 to 2.65 (mL/J) on average. This suggests an interaction because when the gap distance is increased, the variation in efficiency is not the same for the lowest and the highest level of input power.
Plate-Plate Plasma Reactor
The mathematical model for efficiency is given by Eq. (2). To simplify the calculations, the independent variables were coded at a (-1, +1) interval.
G = gap distance,
FR = flow rate and
IP = input power.
Figure 4 shows a three-dimensional model for efficiency.
There is a strong interaction between the variables input power and gap distance, so the effect of these variables cannot be interpreted separately. The interaction occurs because at the lowest value of input power (4.8 W) an increase in gap distance (from 0.56 to 0.89 cm) increases efficiency from 25.40 to 48.75 (mL/J) on average. Nevertheless, at the highest level of input power (10.2 W) the same change in gap distance causes a minor modification in efficiency, which changes from 2.15 to 4.1 (mL/J) on average. Here the interaction between changes in efficiency and increases in gap distance is more evident for the two different values of input power. There is also a strong interaction between the variables gas flow and gap distance, so the effect of these variables cannot be interpreted separately. The interaction occurs because at the lowest value of flow rate (120 mL/min) an increase in gap distance (from 0.56 to 0.89 cm) increases efficiency from 5.15 to 34.55 (mL/J) on average. Nevertheless, the same increment in gap distance at the highest flow rate (330 mL/min) decreases efficiency from 22.4 to 18.3 (mL/J) on average. The interaction is evidenced because the variation in efficiency when the gap distance between electrodes is increased is different depending on whether the flow rate is at the low or the high level.
Selectivity in producing ethylene gave rise to values around 85%.
Nonthermal plasmas efficiently produce highly reactive chemical species. In the reactors, the intense electric field initiates multiple streamers along its length, creating the reactive chemical species.
A clear conclusion about the high concentration of olefin compounds with low molecular weights obtained can be drawn, taking into account that dehydrogenation is a common reaction in plasmas. Compounds such as paraffins normally form a variety of products containing considerable amounts of olefins, in general, by a highly unselective behavior (Hollahan and Bell, 1974). Nevertheless, the reactions in this process of heavy oil conversion showed an important selective behavior. The major byproducts obtained were hydrocarbon compounds with two, three and four atoms of carbon, with ethylene as the main compound (highest concentration).
The set of values for the variables corresponding to the highest part on the surface response plots, shown by each reactor, provides an estimate for the optimum combination of those variables that give rise to the highest efficiency in the production of light olefin compounds. From the analysis of the results obtained with the packed-bed and plate-plate plasma reactors studied here, the plate-plate reactor showed the highest efficiency, about 12 times higher than that of the packed-bed reactor, as well as the highest selectivity in producing ethylene (85%) and achieving notably high efficiencies. The optimal set of values for the variables to get efficiencies on the order of 70 (mL/J) is for the largest gap distance (about 0.9 cm) and the input power at a low value of about 5 W.
It is worth emphasizing the high performance shown by this reactor, making it a very promising reactor for the process studied here.
Box, G. E. P., Hunter, W. G. and Hunter, J. S., Eds., Statistics for Experimenters, John Wiley & Sons, 1978. [ Links ]
Davies, O. L., Ed., The Design and Analysis of Industrial Experiments, Imperial Chemical Industries Limited, Second Edition, New York, 1979. [ Links ]
Hollahan, J. R. and Bell, A.T., Eds., Techniques and Application of Plasma Chemistry, pp. 80, John Wiley & Sons, 1974. [ Links ]
Montgomery, D. C., Ed., Design and Analysis of Experiments, Third Edition, John Wiley & Sons, 1991. [ Links ]
Myers, R. H., Response Surface Methodology, Allyn and Bacon, Boston, 1971. [ Links ]
Yamamoto, T., Chang, J. S., Berezein, A. A., Kohno, H., Honda, S. and Shibuya, A., Decomposition of Toluene, o-Xylene, Trichloroethylene, and their Mixture Using a BaTiO3 Packed-Bed Plasma Reactor, J. Adv. Oxid. Technol., Vol.1, 1996. [ Links ]
Address to correspondence
Received: March 5, 2002
Accepted: September19, 2002