Ind. Eng. Chem. Process Des. Dev., Vol. 18, No. 1, 1979
143
Decomposition of Nitric Oxide in a Silent Discharge Larry A. Haas, Carl F. Anderson, and Sanaa E. Khalafalla' Twin Cities Metallurgy Research Center, Bureau of Mines, U S . Department of the Interior, Twin Cities, Minnesota 55 I 1 1
The effect of some physical and chemical parameters on NO decomposition was determined at 25 O C in a laboratory silent discharge reactor on a gas containing 1% NO in helium. Using a 3-mm electrode-gap reactor, the NO decomposition increased with increasing applied voltages in the range of 9 to 24 kV. The threshold value was estimated at about 3.3 kV. At 24 kV, over 98% of the NO was principally decomposed to its elements with gas residence times greater than 1.2 min. At shorter residence times, neither the N2 nor 0, formation yields were equivalent, the O2 being persistently lower than the N,. The disparity between N2 and O2 formation and NO decomposition became progressively less as the gas residence time increased. The formation of an intermediate dinitrogen trioxide, N,03, can account for the observed results. The addition of H,O, CO,,and CO decreased the NO removal. When CO was added to NO or N,O,COPgas formed indicating that both nitrogen oxides were reduced.
Introduction The decomposition of nitric oxide (NO) in a silent discharge was investigated by the Bureau of Mines as a possible means of converting it to less toxic products by nonthermal means. The oxides of nitrogen are the second most abundant urban atmospheric pollutant ranking next to SOz (Environmental Quality Laboratory, 1972). The principal oxide of nitrogen is NO, but it usually coexists with small quantities of NO2,and this mixture is referred to as NO,. When these nitrogen oxides react with the moisture in the air, chemical smogs and acidic mists are produced. The latter causes undue corrosion to structural materials. Plant life also becomes damaged when the NO, concentration reaches 25 ppm while the safe limit for humans is about 5 ppm (Lyman, 1964). A conventional coal and MHD power plant emits up to 2000 and 5000 ppm, respectively (LaMantia and Field, 1969). A 2-ton electric steel furnace emits up to 96 lb of nitrogen oxides per day (Schueneman et al., 1963). Nitric oxide can also be chemically emitted from low-temperature processes such as that of stainless steel pickling (McCabe, 1952). However, the major source of NO pollution originates from gasoline-powered motor vehicles, which daily emit about 1000 tons in the Metropolitan areas of Los Angeles (Environmental Quality Laboratory, 1972). Nitric oxide is one of the rare endothermic compounds, and its formation is favored with increasing temperature. In order to isolate the small amounts formed in air a t high temperatures (about 3% a t 3000 " C ) , the equilibrium mixture must be rapidly cooled. As NO is cooled in air, only a small quantity of NOz is formed. This fact is related to the peculiar temperature and concentration dependence of the rate of this oxidation reaction, one of the few that slows down as the temperature increases. Therefore, at high combustion temperatures the rate of reaction is very low. At the lower temperatures encountered outside the combustion zone, the rate increases. This rate increase is counterbalanced, however, by the dilution that takes place as the exhaust gases leave the combustion zone and become diluted with air. The dilution lowers the reactant concentrations and consequently the reaction rate. Continuous monitoring of the atmosphere shows that peak NO concentrations of 1 to 2 ppm are common, but NOz concentrations are usually about 0.5 ppm. On the basis of estimated global background levels and annual emission rates for NO,, the average lifetime of NO2 and NO in the atmosphere is estimated (Environmental Protection Agency, 1971) to be about 3 days and 4 days, This ariicle not subject to
U S Copyright.
respectively. These lifetimes indicate that natural processes are active in removing NO, from the atmosphere. Many of the serious effects of NO, pollution are caused not by the oxides themselves but by products resulting from the involvement of NO, in photochemical reactions. These products, called photochemical oxidants, are the more harmful components of smog and are produced when other pollutants take part in a group of naturally occurring atmospheric reactions involving NO and NOz. These reactions, collectively known as the NOz photolytic cycle, are a direct consequence of an interaction between sunlight and NOz. The steps in this cycle are illustrated in Figure 1 and can be described in the following sequence: (1) NO2 absorbs energy from the sun in the form of ultraviolet light. ( 2 ) The absorbed energy causes the NO2 molecule to break into NO and oxygen atoms. The resulting atomic oxygen is extremely reactive. (3) The atomic oxygen reacts with atmospheric oxygen to produce ozone, a secondary pollutant. (4)The resulting Ozone reacts with NO to give NO2 and Oz, and the cycle is completed. The net effect of this cycle is the rapid cycling of NO,; if it were not for competing reactions in the atmosphere, the cycle would produce no net overall effect. The atmospheric NO and NOz concentrations would not change, because O2 and NO would be formed and destroyed in equal quantities. The competing reactions involve hydrocarbons, which are often emitted from the same source as NO,. The hydrocarbons interact in a way that causes the cycle to become unbalanced, so that NO is converted into NO2 faster than NOz is dissociated into NO and 0 (Environmental Protection Agency, 1971). The control of NO, pollution has proved to be quite difficult, and no widely applicable methods have been developed. Sorbents such as molecular sieves (Kranich et al., 1973) and iron oxides (Otto and Shelef, 1970) have been suggested to remove NO from the waste gases of stationary sources, but this method only concentrates the pollutant, and no decontamination is accomplished. Catalytic reactors have been considered to be possible answers to treat automobile exhaust gases, and over 600 catalysts have already been evaluated (Hydrocarbon Processing, 1971). However, with this method, toxic products (e.g., NH, and cyanides) have been produced when NO is reduced with CO, Hz, and hydrocarbons at 200 and 400 " C (Anderson and Green, 1961). Attempts have also been made to thermally decompose NO to its elements in the absence of reductants. The equilibrium data, calculated from JANAF data (Stull and
Published 1978 by the American Chemical Society
144
Ind. Eng. Chem. Process Des. Dev., Vol. 18, No. 1, 1979 2
t a:
I
I
I I
CI
Figure 1. The photolytic NO2 cycle.
i c,+cz+-c,-+------
CI *cz-
nb
CO I
\f
\\\
\\, \\
I 0
5
10 15 TIME, m i n
20
25
Figure 3. Composite representation of typical gas chromatograms obtained with one or two columns.
10' I 0 ' O l
0
201,
400
677
600
1C M P E R A l LIRE
I000
I200
I400
K
Figure 2. Effect of temperature on the equilibrium constant of reactions involving nitrogen oxides.
Prophet, 19701, indicate that all nitrogen oxides should dissociate to N2 and O2 at temperatures below 1400 "C (Figure 2). Well-documented evidence exists that metastable nitrogen oxides such as N20, NO, N203,N204,NO2, and N205can be experimentally isolated in the gas phase a t room temperature. At low temperature and high pressure, NO decomposes to N 2 0 and NO2 (Bamford, 1972). Nitrous oxide (N20) can be dissociated at 600 "C with catalysts such as chromia (Keenan and Iyengar, 19661, manganese supported on MgO (Cimino and Indovina, 19701, spinels [ (Ni,Mg)Al,O,] (Cimino and Schiavello, 1971), or rare-earth oxides (Winter, 1969). The objective of this investigation is to study the decomposition of NO by nonthermal methods a t ambient temperature and pressure. Extent of NO decomposition was determined as a function of several physical and chemical variables in a silent electric discharge reactor yielding data not heretofore available on this subject. The electrically induced reactions leading to the dissociation of NO in low-pressure glow-discharge reactors are reported in a parallel investigation (Kuehn et al., 1975). Experimental Method A concentric tube silent electric discharge reactor was used in this investigation. Details of this apparatus were described (Haas et al., 1972) in a previous publication. The reactant gases moved up between the annular gap formed by two Vycor (Reference to specific trade names does not imply endorsement by the Bureau of Mines.) glass tubes
(or one Vycor and one Monel tube). Electrical contact to the tubes was made by the use of a 7.7 M zinc chloride electrolyte. The inner tube was removable in order to permit a study of the effect of gap-width by inserting different diameter inner tubes. A high-voltage transformer supplied the power to the reactor. The primary input power corresponded to about 10 W a t an applied voltage of 24 k V using a reactor with a 3-mm electrode gap and a 80 cm3 volume. The applied rms (root-mean-square) voltage was measured with a high-voltage probe and a high-input impedance voltmeter. Chemically pure gases from compressed-gas cylinders were metered into a mixing chamber by means of low-flow needle valves constructed from stainless steel. Rotameters and an oil-filled wet-test meter were used to measure the individual and total gas flows. The inlet gases were a t a pressure of 1 atm and a temperature of 25 "C. The gas composition was determined with a chromatograph calibrated with commercial gas mixtures. The chromatograph was operated at 30 O C with thermoconductivity detectors, helium carrier gas, and two partitioning columns. Each column was 120 in. long and 0.125 in. in diameter. Column 1 (C,) was filled with -50 +80 mesh Porapak QS, and column 2 (C21 was filled with -40 +60 mesh Linde 5A molecular sieve. Column 1 separated C 0 2 and N 2 0 whereas column 2 separated 02,N2,NO, and CO (Figure 3, A and B, respectively). Columns 1 and 2 were operated in series the first 3.5 min (Figure 3C) after which time O2 was eluted from columns 1 and 2, and N P ,NO, and CO were pushed into column 2. Column 2 was then bypassed in order to elute the C 0 2 and N 2 0 from column 1. As soon as the N 2 0 peak was eluted from column 1, the columns were reconnected to the original position, and N2, NO, and CO eluted from column 2. Other oxides of nitrogen were not detected and perhaps were trapped in the columns. Semiquantitative analyses of NO2 were conducted using the Saltzman method (Davis and O'Neill, 1966). Basically, the technique utilizes a colorless NO2 absorbent that becomes purple in contact with NO2. The color intensity as determined spectrophotometrically is proportional to
Ind. Eng. Chem. Process Des. Dev., Vol. 18, No. 1, 1979
Table I. Experimental Results Obtained with a Gas Mixture Containing 1%NO in Helium total gas
resi-
test flow, time, no. L/min min
v
N2
decomposed %
Electrode Gap = 3.0 mm, 0.049 1.63 98 0.065 1.23 98 0.107 0.75 9 5 0.232 0.34 80
Discharge Volume = 8 0 22 9 28 10 11 45 83 14
2
cm3 6 5 3 1
,
,
,
10
IS
20
~,Of,~
20
0
5
CeI
where f is the total gas flow in liters per minute. The formation yields were calculated from the above equation by putting c1 = 0. The residence time of the gas was defined as the ratio of the electric discharge volume (VJ to the total volumetric gas flow (0. Results and Discussion Effects of Voltage and Residence Time on Reaction Products. A gas mixture containing 1% NO in helium was passed a t a flow of 0.24 L/min through the 3-mm gap reactor at various applied voltages. The NO decomposition increased with the applied voltage in the range of 9 to 24
kV (Figure 4). The extrapolated threshold voltage was estimated to be 3.3 kV. The dependence of the degree of NO decomposition on gas-residence time a t 24 kV was investigated by varying the total gas flow rate. The results in Table I, columns 2 and 3, indicate that 98% of the 1%NO in the inlet gas was decomposed at gas-residence times longer than 1.2 min (with either the 3.0 or the 5.5-mm electrode gap reactor as shown by tests 1, 2, 5 , and 6). However, a t a short gas residence time of 0.6 min, only 20% of the NO was decomposed. The dependence of NO decomposition on residence time is shown in Figure 5. At long residence times (1.6 rnin), the NO decomposition approached approximately the sum of the formation of N2 and 02,as expected from the dissociation reaction 9 (Figure 2). However, as the gasresidence time decreased, the discrepancy between the sum of the N2and O2 products and the decomposition of the reactant, NO, became more apparent. The O2 formation yield was also considerably lower than the N2yield at short residence times. This disparity indicates that other products besides N2and O2 may be formed. Some N 2 0 (Table 11,test 15) was observed, but this could not account for the discrepancy between the O2 and N2formation and the NO decomposition. The disparity could be explained by assuming that some of the unreacted NO would be oxidized by its plasma dissociation product O2 to form NO2 or some higher oxide. Nitrogen dioxide was not detected, and methods for analysis of the other higher oxides of nitrogen were unavailable.
Table 11. Experimental Results Obtained with Various Gas Additives in NO and N,O Gas Mixtures'
test no.
gas additive
total flow. L/min
14 15 16 17 18 19 20 21 22 23
none none 3.0% CO 3.0% CO 11.0%co 0.2% co2 0.2% co, 0.8% CO, 3.0% H,O 3.0% H,O
0 24 0 24 24 0 24 24 0 24
0.240 0.240 0.246 0.241 0.271 0.234 0.234 0.239 0.069 0.069
24 25 26 27
none none 13.0% CO 13.0% CO
0 24 0 24
0.610 0.610 0.24 1 0.241
25
Figure 4. Effect of applied voltage on the NO decomposition yield.
2.2414 X
voltaee. kv- '
,
APPLIED VOLTAGE, k V
the NOz concentration. The absorbent solution consists of acetic acid, sulfanilic acid, and a coupling agent N(1-naphthy1)ethylenediaminedihydrochloride. After the desired reactant gas concentrations and flows were obtained, the electric discharge was initiated. The yield ( Y , expressed in micromoles per minute) was calculated from the material balance of the individual reactants and products. If the volume percentage of NO (or N 2 0 ) in the inlet gas is c1 and in the exit products is ce, then the balancing of this material across the reactor terminals leads to flcl -
60
i 4 0 1
Electrode Gap = 5.5 mm, Discharge Volume = 1 1 6 cm3 5 0.059 1.36 98 25 11 10 6 0.098 0.82 98 43 14 8 7 0.146 0.55 97 63 14 5 8 0.205 0.39 92 84 16 3 9 0.240 0.33 8 9 95 16 3 10 0.455 0.18 61 124 16 2 11 0.809 0.10 36 130 18 4 12 1.085 0.07 24 116 19 5 1 3 1.314 0.06 20 117 19 6
Y=
I
0,
formed, formed, pmol/min pmol/min pmol/min
1 2 3 4
I
BO
XT/-.
I\
I
145
N, 0 2 co * formed. formed. formed. pmol/min pmol/mih pmol/mih pmol/miA
No decomposed %
1%NO in Helium
