Académique Documents
Professionnel Documents
Culture Documents
To cite this article: John M. Kasper , Christian A. Clausen III & C. David Cooper (1996) Control
of Nitrogen Oxide Emissions by Hydrogen Peroxide-Enhanced Gas-Phase Oxidation Of
Nitric Oxide, Journal of the Air & Waste Management Association, 46:2, 127-133, DOI:
10.1080/10473289.1996.10467444
Volume 46 February 199§ Journal of the Air & Waste Management Association 127
Kasper, Clausen III, and Cooper
O 3 -+NO->NO 2 + O 2 (1) Of the first two reactions, reaction (7) is less likely to occur
OH + NO -> NO 2 + H (2) because the energy required to break the HOO-H bond is
OH + NO + M -> HNO 2 + M (3) much greater than the energy required to break the HO-OH
OH + NO 2 + M -»• HNO 3 + M (4) bond.9 Heterogeneous decomposition (reaction 8) occurs
HO 2 + NO -> NO 2 + OH (5) readily at surfaces, and is thought to predominate at tem-
In examining reactions (1) through (5), it is apparent that peratures below about 400 °C; reaction (8) should be mini-
ozone oxidizes NO only to NO2, while OH radicals oxidize mized in order to maximize OH radical formation. 10
NO to HNO2 and HNO3 as well as NO2. Both HNO2 and Hydroxyl attack of hydrogen peroxide can generate HO2 radi-
HNO3 are much more soluble than NO or even NO2. The cals, but is only important at high hydrogen peroxide con-
promise of NOX removal in ordinary scrubbers makes the centrations.11
use of OH radicals more attractive than ozone. Surface reactions affect the decomposition of hydrogen
Absorption of NO and NO2 in aqueous solutions can be peroxide, as noted by a number of authors. 1214 Generally,
complex because these gases may react with water and form peroxides are preserved at acidic surfaces and destroyed at
HNO2 and HNO3.3 NO oxidation to NO2 using ozone, and basic surfaces.15 In one study, boric acid treatment of the
subsequently removing the NO2 in a scrubber, has not been interior walls of a reactor tube significantly reduced the de-
economically attractive for large industrial and power plant composition of hydrogen peroxide.16
applications.4 However, recent studies have shown that us-
ing sulfite scrubber solutions to capture NO2 has promise.5 EXPERIMENTAL APPARATUS AND PROCEDURES
Other researchers have investigated the benefits of liq- The experimental apparatus used here was a modification
uid phase oxidation on the absorption of NOX.68 The pro- of the system used by Cooper and Clausen, who studied the
posed process involved two vessels: in the first, various enhanced oxidation of organic vapors using hydrogen per-
peroxides or other oxidizing compounds in an acidic solu- oxide and ozone. 1718 The experimental system consisted of
tion provided hydroxyl radicals to oxidize NOX to HNO3; in four sub-systems: (1) the gas and hydrogen peroxide supply
the second, the HNO3 was removed in an alkaline solution. sub-systems; (2) the quartz tube reactor and tube furnace;
Oxidizing the NOX in the gas phase would eliminate the (3) the scrubbers; and (4) the gas analyzers. A schematic of
complexity and expense of the first scrubber. the experimental system is shown in Figure 1.
One method of generating OH radicals is the thermal The test gas was 5.0% (by volume) NO in nitrogen, drawn
decomposition of hydrogen peroxide. Hydrogen peroxide from a high pressure bottle. It was diluted with air to achieve
injected into a stream of hot gases can decompose accord- the desired inlet concentration prior to entering the
ing to one or more of the following reactions. reactor. The reactor was a quartz tube capable of operating at
H2O2 -• 2OH (6) 800 °C, heated externally by a Lindberg tube furnace with
H2O2 -> HO2 + H (7) three independently controlled heating zones. The quartz
H2O2->H2O + V2O2 (8) tube was 2.5 m long with a 5.9 mm inside diameter and a
OH + H2O2 -> H2O + HO2 (9) heated volume of 70.5 cc. The inside surface of the first third
PRESSURE
GAGE
INJECTOR
HEATING TAPE
HEATING
TAPE ROTAMETERS
TUBE FURNACE
WITH QUART? TUBE
REACTOR I METERING VALVES
SYRINGE PUMP
FOR INJECTING
H202
(FILTER PACKAGE
«b cb
128 Journal of the Air & Waste Management Association Volume 46 February 1996
Kasper, Clausen III, and Cooper
HEATING TAPE
HEATING TAPE THERMOCOUPLE —
TUBE FURNACE EXTERIOR WALL
NOx/AIR INLET i WICK
•BEGINNING OF HEATED ZONE
INSULATED PLUG
INJECTOR TUBE
•3.8
-8.8-
of the tube was treated daily by wiping with a saturated boric liquid for these tests was a deionized water/NaOH solution
acid solution. Two thermocouples, each inserted down the with initial pH of 9. This scrubber liquid was chosen to
bore of the reactor from opposite ends, were used to monitor separate the HNO2 and HNO3 from the gases for analytical
gas temperatures. To avoid condensation in the exhaust tub- reasons. It was not intended to simulate commercial SO2
ing, a heating tape was used to maintain the exhaust tubing scrubbers which are predominantly lime or limestone
at approximately 50 "C. A similar heating tape preheated based. The NO and NO2 remaining in the gas exiting the
the inlet gas just upstream of the H2O2 injection point. scrubbers were measured by a chemiluminescent type NO*
An aqueous solution of hydrogen peroxide was injected gas analyzer (in this case a Thermo-Environmental, model
into the reactor inlet through a 30 cm long, 1.6 mm OD teflon 42H). Nitrate and nitrite ions in the scrubber liquid were
tube (see Figure 2). The discharge end of the teflon tube was analyzed colorimetrically.19 Although there was probably
wrapped in a wick of Pyrex glass wool. The wick was approxi- some reaction in the liquid of HNO2 to NO 2; 3 the overall
mately 2.5 cm long and loosely filled the inside diameter of nitrogen mass balance was reasonable, as shown in the
the reactor. The wick was found to reduce oscillations in reac- Results section.
tor exit NOX levels during peroxide injection. It is believed The first two scrubbers were each made from a 1-liter
that the wick kept liquid droplets from forming at the end of graduated cylinder filled with Raschig rings (made of 0.635
the tube, and so evenly distributed the liquid peroxide and cm OD glass tubing cut into 0.5 to 1 cm lengths). The first
provided for smooth, continuous evaporation. However, the two scrubbers were filled with 600 mL of scrubbing solution
wick surface may have contributed to significant heteroge- (the net volume of the Raschig ring packing was approxi-
neous decomposition of the H2O2, thus producing high ap- mately 400 mL). The third scrubber was made from a 500
parent usage rates of the peroxide. Hydrogen peroxide flow mL Erlenmeyer flask, containing 350 mL of scrubbing solu-
was delivered by a syringe pump, which accurately controlled tion, with no packing. A 15 cm length of perforated tubing
the very low liquid flow rates required. The hydrogen perox- was attached to the discharge of the inlet tube in the third
ide solution was prepared from a stock solution of approxi- scrubber, through which the gas was bubbled.
mately 30% (by weight) hydrogen peroxide. The hydrogen Pressure drops during scrubbing operation were in the
peroxide concentration was determined precisely by titration range of 30 to 32 inches (76 to 81 cm) of water. A metering
with potassium permanganate.19 valve was placed in the scrubber bypass loop to adjust reac-
Laboratory scrubbers were used to remove and separate tor pressure during scrubber bypass operation to be equal to
HNO2 and HNO3 from the exhaust gas stream. The scrubber the reactor pressure observed during scrubbing.
Volume 46 February 1996 Journal of the Air & Waste Management Association 129
Kasper, Clausen III, and Cooper
RESULTS AND DISCUSSION within about 0.3 seconds. Initial (injected) gas-phase hy-
The first series of tests determined the operating tempera- drogen peroxide concentrations varied up to about 1250
ture range for NO conversion (defined as the oxidation of ppm, and a blank was run by injecting deionized water.
NO to any other species, such as NO2, HNO2, or HNO3). As mentioned, scrubbers were used to remove soluble
Hydrogen peroxide was injected at an inlet concentration nitrogen species from the reactor exhaust stream. The reac-
of 1200 ppm. Nitric oxide conversion varied with reactor tor was started up and allowed to come to steady state with
temperature, as shown in Figure 3, revealing an optimum the gases being exhausted to the atmosphere. The gases were
temperature of about 500 °C. At temperatures below about switched into the scrubbing systems for periods of 15 min-
400 °C, it is thought that radical production is small; at utes at a time to collect data. During scrubbing, a large drop
temperatures above 600 °C, the rate of NO2 thermal reduc- in pH usually was observed.
tion to NO becomes appreciable. After determining the op- Figure 4 shows the decrease in outlet NO concentration
timum temperature, all further tests were run at a reactor as hydrogen peroxide initial concentration increased. At
residence time of 0.7 sec, a reactor temperature of 500 °C, 1222 ppm hydrogen peroxide, 97% conversion of NO was
and an inlet NOX concentration of 500 ppm. Later tests achieved at the reactor exit. At this point, the initial mole
revealed that the conversion of NO takes place quickly, ratio of hydrogen peroxide to NO was 1222/472 or 2.6. At
726 ppm hydrogen peroxide, 75% con-
version of NO occurred. Here, the hy-
drogen peroxide/NO mole ratio was
1.6. According to equations presented
earlier, oxidation of the NO to NO2,
HNO2, and HNO3 requires a hydrogen
peroxide/NO mole ratio of 0.5 to 1.0.
Thus, one should expect a hydrogen
peroxide/NO mole ratio between 0.5
and 1.0 based on stoichiometry, if all
of the hydrogen peroxide were con-
sumed in the oxidation of NO.
Non-useful destruction of OH radi-
cals (e.g., at reactor surfaces) would in-
crease the observed hydrogen
peroxide/NO mole ratio for a given NO
conversion. Because the observed hy-
drogen peroxide/NO mole ratio was
Figure 3. Variation of NO conversion with reactor temperature (at an initial concentration of higher than stoichiometric, it appears
1200 ppm H O in the gases). that significant destruction of OH radi-
cals still occurred at the wick and at
the reactor walls, despite the boric acid
treatment. It is expected that in a full-
scale application, these surface effects
would not occur, and a much lower
mole ratio would be observed.
Deionized water, with no peroxide,
was injected as a blank; the results show
that deionized water had no effect on
NO conversion relative to no injection
at all. The observed small drop in
NO concentration through the reactor
(16 ppm) was due primarily to the dilu-
INJECTED HYDROGEN PEROXIDE CONCENTRATION, ppm tion effects of adding the water.
Nitrogen species distribution varied
with injected peroxide concentration
as shown in Figure 5. This figure shows
Figure 4. Effect of hydrogen peroxide on NO conversion (all tests at 500 °C and 0.7 seconds that more HNO2 and HNO3 are pro-
residence time in reactor). duced at higher H2O2 injection rates.
130 Journal of the Air & Waste Management Association Volume 46 February 1996
Kasper, Clausen III, and Cooper
Reactor Reactor Reactor Scrubber Scrubber Scrubber Scrubber Nitrogen Nitrogen Mass
Reactor Inlet Inlet Exit Exit Exit Liq Liq Into Out of Balance
Test [H202] [NOJ [NO] [NO] [NO] [NO2] [NO2—N]* [NO3~N1* System System Error
Point ppmv ppmv ppmv ppmv ppmv ppmv mg mg mg mg mg%
Average absolute error 4.5%. *NO2-N = nitrite nitrogen; NO3-N = nitrate nitrogen.
Volume 46 February 1996 Journal of the Air & Waste Management Association i 31
Kasper, Clausen III, and Cooper
Table 3. Simplified cost comparison of H2O2 injection (with existing wet scrubber) vs new SCR have wide application to power
for NOx control at a base case power plant. plants and other combustion
sources of NOX because the acids
SCR H?O? with Existing FGD
High Ratio Low Ratio
HNO2 and HNO3 will be removed in
existing wet or dry scrubbers used
Capital Cost ($) for the removal of SO2 and/or HC1.
$/kw $150/kw NA NA High conversions of NO were
for Base Plant $75 million $2 million demonstrated in the laboratory. A
97% conversion of NO was observed
Levelized cap. cost ($/yr)
at an initial mole ratio of peroxide/
(20 yrs, 6% interest) $ 6,600,000 $170,000
NO of 2.6; a 75% conversion of NO
(20 yrs, 4% interest) $ 5,500,000 $ 150,000
was seen at a peroxide/NO mole ra-
Chemicals Cost ($/yr) tio of 1.6. It is likely that commer-
Ammonia ($315/ton of 100% NH3) $ 840,000 cial applications of this process will
Catalyst (replace every 4 yrs) $ 300,000 use lower hydrogen peroxide/NO
mole ratios for a given NO conver-
H2O2 ($1360/ton of 100% H2O2) $18 million $3.6 million sion, due to reduced OH radical de-
Extra limestone $ 150,000 $150,000
struction at reactor surfaces. The
* Notes: optimum temperature is between
1. Base case is assumed to be a 500 MW plant with a wet limestone FGD system, operating at a capacity factor 400 °C and 650 °C.
of 0.65, and requiring a NOX emissions reduction of 7200 tons per year.
2. High ratio is 2.5 mol H2O2/mol NOX; low ratio is 0.5 mol/mol.
Daily treatment of the reactor
3. Hydrogen peroxide is available at 50% or 70% solutions, but cost is the same on a "pure" basis. surface with a boric acid solution
4. Estimate for capital cost of H2O2 system includes rail car unloading, pumps, pipes, tankage, and injectors. significantly increased NO conver-
sion in the laboratory tube reactor.
This treatment was believed to pre-
or MgO supported on a fabric. Thus, it is anticipated that vent destruction of hydrogen peroxide and OH radi-
commercial removal of these acid gases could be done in cals on the surfaces of the reactor. However, the higher
an existing baghouse with dry sorbent injection. than theoretical mole ratios of H2O2 to NOX indicated
The economic feasibility of this proposed method for that there was still significant non-productive destruc-
NOX control was not studied in detail, but a "back of tion of H2O2.
the envelope" estimation was made. Assume as a base
case a 500 MW power plant with an existing wet flue RECOMMENDATIONS
gas desulfurization system, that must now remove 80 to Laboratory scale demonstration of the enhanced oxida-
90% of its NOX emissions, or about 7200 tons per year. As tion and removal of NOX has been accomplished. More
shown in Table 3, an SCR system would cost about $75 data is required to accurately determine the mechanism
million to install, and would incur about $1.1 million per and the kinetic constants. Pilot scale studies and larger
year in ammonia and catalyst costs. The H2O2 injection scale demonstrations using gases with other constitu-
system could probably be installed (either just upstream ents found in power plant flue gas should be started to
or downstream of the economizer) for about $2 million, gather economic and technical feasibility data for full
and would incur chemical costs of about $3.6 to $18 mil- scale industrial application of this technology. The eco-
lion per year, depending on the actual usage ratio of hy- nomic feasibility depends critically on the usage ratio of
drogen peroxide to NOX. Depending on this ratio, the H2O2 to NOX. Also, studies of the effects of nitrites and
process might be economically viable, with its savings in nitrates on scrubber chemistry and sludge disposal re-
capital costs offsetting its higher chemical costs. quirements are needed.
CONCLUSIONS ACKNOWLEDGMENTS
It has been experimentally demonstrated that NO can This project was funded in part with Federal funds as part
be oxidized to NO2, HNO2, and HNO3 in the gas phase of the program of the Gulf Coast Hazardous Substance
by OH radicals generated by the thermal decomposi- Research Center, which is supported under cooperative
tion of hydrogen peroxide. The process temperatures agreement R815197 with the United States Environmen-
and residence times used were representative of down- tal Protection Agency. The contents do not necessarily
stream zones of many types of combustion devices. This reflect the views and policies of the U.S. EPA nor does the
process may be directly applicable to the treatment of mention of trade names or commercial products consti-
flue gases for the removal of NOX. This process may tute endorsement or recommendation for use.
132 Journal of the Air & Waste Management Association Volume 46 February 1996
Kasper, Clausen III, and Cooper
REFERENCES 15. Drysdale, D.D.; Lloyd, A.C. "Gas phase reactions of the hydroxyl
1. Anderson, J.G.; Margitan JJ; Kaufman, F. "Gas phase recombina- radical," Oxidation and Combustion Reviews 1970, 4.
tion of OH with NO and NO 2 " /. Chemical Physics 1974, 60, No. 18. 16. McLane, C.K. "Hydrogen peroxide in the thermal hydrogen oxygen
2. Chao, D.C. Kinetic Modeling of the Hydrogen Peroxide or Ozone En- reaction: Thermal decomposition of hydrogen peroxide" Journal of
hanced Incineration of Nitric Oxide and/or Carbon Monoxide, Master's Chemical Physics 1949, 17, 379-385.
Thesis, University of Central Florida: Orlando, FL, 1994. 17. Cooper, CD.; Clausen, C.A.; Tomlin, D.; Hewett, M.; Martinez, A.
3. Counce, R.M.; Perona, JJ. "Scrubbing of gaseous nitrogen oxides in "Enhancement of organic vapor incineration using hydrogen per-
packed towers," AIChE Journal 1983, 29, 26. oxide," /. Haz. Materials 1991, 27.
4. Harrison, J. W. Technology and Economics ofFlue Gas NOx Oxidation by 18. Clausen, C.A.; Cooper, CD.; Hewett, M.; Martinez, A. "Enhance-
Ozone U.S. Environmental Protection Agency. U.S. Government ment of organic vapor incineration using ozone," /. Haz. Materials
Printing Office: Washington, DC, 1976; EPA-600/7-76-033. 1992, 31.
5. Shen, C.H.; Rochelle, G.T. NO2 Absorption into Sulfite Solutions: 9/92- 19. Standard Methods for the Examination of Water and Wastewater;
9/93 Progress Report Department of Chemical Engineering, Univer- (Greenberg, A.E.; Trussell, R.R.; Clesceri, L.S., Editors); Port City Press:
sity of Texas at Austin, 1993. Baltimore, MD, 1985.
6. Perlmutter, H.D.; Ao, H.; Shaw, H. "Absorption of NO promoted by
strong oxidizing agents: Organic tertiary hydroperoxides in n-
hexadecane," Environmental Science and Technology 1993, 27, No. 1.
7. Littlejohn, D.; Chang, S.G. "Removal of NOx and SO2 from flue gas
by peracid solutions,", Ind. Eng. Chem. Res. 1990, 29, No. 7.
8. Cooper, H.B.H. "Removal and recovery of nitrogen oxides and sul-
fur dioxide from gaseous mixtures containing them," U.S. Patent About the Authors
4426364, 1984.
9. Sanderson, R.T. Simple Inorganic Substances. Robert E. Krieger Pub- Mr. John Kasper, a recent Master's graduate from the
lishing Company: Malabar, FL, 1989. University of Central Florida, is an environmental engi-
10. Satterfield, C.N.; Stein, T.W. "Homogeneous decomposition of hy-
neer with Grove Scientific Company in Orlando,
drogen peroxide vapor," Journal ofPhysical Chemistry 1957,61, 537.
11. Martinez, A.; Geiger, C; Hewett, M.; Clausen, C.A.; Cooper, CD. Florida. Dr. Christian A. Clausen III is a Professor in
"Using hydrogen peroxide or ozone to enhance the incineration of the Chemistry Department at UCF. Dr. C. David Coo-
volatile organic vapors," Waste Management 1993, 13, 261-270.
per (corresponding author) is Professor of Engineer-
12. Wilson, W.E., Jr. "A critical review of the gas-phase reaction kinetics
of the hydroxyl radical," /. Phys. Chem. Ref. Data 1972,1, No. 2. ing in the Civil and Environmental Engineering
13. Hoare, D.E. "Studies of the reactions of the hydroxyl radicals," Pro- Department at UCF, Orlando, Florida, 32816-2450. All
ceedings of the Royal Society, London, 1966, A291, 76. work was conducted at the University of Central Florida.
14. Schumb, W.B.; Satterfield, C.N.; Wentworth, R.L. Hydrogen Peroxide
A.C.S. Monograph No. 128, Reinhold Publishing: New York, 1958.
• »• *
« • _ *•
SCHOLARSHIP PROGRAM
est 1983
Jtiff
FUNDING THE FUTURE
•4 WANTED: Scholarship Contributors and Applicants
Volume 46 February 1i96 Journal of the Air & Waste Management Association 133