COPAC-4 ABATEMENT OF NITROGEN OXIDES EMISSIONS FROM STATIONARY SOURCES National Academy of Engineering National Research Council ------- COPAC-4 ABATEMENT OF NITROGEN OXIDES EMISSIONS FROM STATIONARY SOURCES Prepared by Ad Hoc Panel on Abatement of Nitrogen Oxides Emissions from Stationary Sources Committee on Air Quality Management Committees on Pollution Abatement and Control Division of Engineering National Research Council National Academy of Engineering Washington, D.C. J 972 ------- This is the report of a study undertaken by the Committee on Air Quality Management Ad Hoc Panel on Abatement of Nitrogen Oxides Emissions from Stationary Sources for the National Academy of Engineering in execu- tion of work under Contract No. CPA 70-48 with the Office of Air Programs of the Environmental Protection Agency. As a part of the Division of Engineering of the National Research Council, the Committees on Pollution Abatement and Control perform study, evaluation, or advisory functions through groups composed of individuals se- lected from academic, governmental, and in- dustrial sources for their competence and Interest in the subject under consideration. Members .of these groups serve as individuals contributing their personal knowledge and judgments and not as representatives of any organization by which they are employed or with which they may be associated. 11 ------- PREFACE Americans are justifiably concerned with the deteriorating quality of their physical environment. The determination of government—at the federal, state, and local levels—to reverse this trend is reflected in recently enacted laws and in strengthened enforcement actions. Control of pollution involves technical deci- sions and economic trade-offs of benefits received for costs incurred, and the decisions that are made will be the wisest if a considerable portion of the concerned public is informed on the problems and the facts that underlie them. This report, which is an assessment of the control technology for a significant air pollutant— nitrogen oxides—is addressed primarily to a technical audience. However, it is hoped that it will have value to others who are concerned primarily with political and social aspects of air pollution control. In preparing this report the ad hoc Panel on Abatement of Nitrogen Oxides Emissions from Station- ary Sources met in Washington, D,C. several times over a period of seven or eight months, generally for two- day sessions, during which a large number of experts from government and industry presented Information and their views and insights. This was supplemented by review of many technical documents, reports, and papers. All information received was extensively debated by the members of the Panel, each of whom is directly con- cerned in his professional activities with some aspect of air-pollution control. On this basis, the Panel arrived at a collective judgment, recognizing that knowledge about both the degree of health hazard and the technology of control Is increasing rapidly and that current judgments may undergo-considerable change within a few years. The goal of the Panel has been to provide a balanced viewpoint, verifiable quantita- tively as far as possible, concerning the current status of abatement of nitrogen oxides emissions from stationary sources and what is reasonably achievable in the near future. It is hoped this will be of value to govern- ment and to the public in making the complex decisions that lie ahead. Charles N. Satterfield, Chairman iii ------- NATIONAL ACADEMY OF ENGINEERING NATIONAL RESEARCH COUNCIL DIVISION OF ENGINEERING COMMITTEE ON AIR QUALITY MANAGEMENT Jack E. McKee, California Institute of Technology, Chairman Reid A. Bryson, The University of Wisconsin, Ex Officio* Thomas H. Chilton, Retired, E. I. du Pont de Nemours and Company, Inc. Merrell R. Fenske,** The Pennsylvania State University S. K. Friedlander, California Institute of Technology Robert L. Hershey4 Retired, E. I. du Pont de Nemours and Company, Inc. Chalmer G. Kirkbride, Retired, Sun Oil Company Charles N. Satterfield, Massachusetts Institute of Technology Thomas K. Sherwood, University of California at Berkeley Staff R. W. Crozier, Executive Secretary^ Committees on Pollu- tion Abatement and Control, National Research Council J. M. Marchello, Staff Engineer, Committees on Pollutiot • Abatement and Control, National Research Council Barbara P. Sowers, Administrative Secretarys Committees on Pollution Abatement and Control, National Research Council Liaison Representative - EPA John 0. Smith, Chief, Office of Engineering Analysis, Division of Control Systems, Stationary Sources Pollution Control Programs, Office of Air Programs, Environmental Protection Agency *Liaison—NAS-NAE Environmental Studies Board #*Deceased iv ------- NATIONAL ACADEMY OF ENGINEERING NATIONAL RESEARCH COUNCIL DIVISION OF ENGINEERING COMMITTEE ON AIR QUALITY MANAGEMENT AD HOC PANEL ON ABATEMENT OF NITROGEN OXIDES EMISSIONS FROM STATIONARY SOURCES Charles N. Satterfield, Massachusetts Institute of Technology, Chairman David Archer, Westinghouse Electric Corporation William Bartok, ESSO Research and Engineering Company Thomas H. Chilton, Retired, E. I. du Pont de Nemours and Company, Inc. Donald N. Felgar, Southern California Edison Company R. M. Lundberg, Commonwealth Edison Company William H. Manogue, E. I* du Pont de Nemours and Company, Inc. M. S. Peters, University of Colorado Staff R. W. Crozier, Executive Secretary, Committees on Pollu- tion Abatement and Control, National Research Council J. M. Marchello, Staff Engineer, Committees on Pollution Abatement and Control, National Research Council Barbara P. Sowers, Administrative Secretary, Committees on Pollution Abatement and Control, National Research Council Liaison Representative - EPA Stanley J. Bunas, Division of Control Systems, Office of Air Programs, Environmental Protection Agency v ------- CONTENTS I. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS . . 1 A. Summary . * . . . ....... 1 B. Conclusions 2 C. Recommendations B II. INTRODUCTION 11 A. The Clean Air Amendments of 1970 11 B. Role of the Academies . 13 C. Federal Research, Development, and Demonstration 14 III. SOURCES OF NITROGEN OXIDES 18 IV. FORMATION AND CONTROL OF EMISSIONS FROM COMBUSTION SOURCES. . 27 A. Equilibrium and Kinetics. . 27 B. Factors Affecting Utility Boiler Emissions . 29 C. Fluidized-Bed Combustion. ........ 36 D. Other Combustion Processes 38 1. Gas Turbines, 38 2. Domestic Heaters. . 41 3. Industrial Furnaces . . 42 4* Incinerators 43 V, STACK-GAS CLEANING. 44 VI. FORMATION AND CONTROL OF EMISSIONS FROM CHEMICAL OPERATIONS 47 VII. SAMPLING AND ANALYTICAL METHODS 49 TABLES 1. Summary of estimated nitrogen oxide emissions in the U.S., 1969 . * 19 2. Breakdown of estimated NOg emissions from stationary industrial sources, 1968 .... 20 APPENDIX A. Bibliography. . ...... 50 vi ------- I SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS A. SUMMARY National and regional standards for air-quality management are being defined under the Clean Air Amend- ments cbf 1970 (Public Law 91-604).1 Keeping the costs of meeting these standards within bounds and minimizing the burden on our national economy will call for the best efforts and most careful planning at all levels, from individuals, civic groups, and companies to local, regional, state, and federal agencies. The nitrogen oxides emitted from industrial* sources are essentially nitric oxide (NO) and nitrogen dioxide (NO2). They are generally grouped together and, for convenience, termed N0x» Nitrous oxide, ^0, at the levels emitted by most chemical processes, is believed to be innocuous and is not included in the definition of N0X. About 53 percent of the total man- made N0X emissions in the United States are from sta- tionary sources. (The remainder is emitted by vehicles.) The largest stationary-source contributions are from the fossil-fuel-fired boilers of electric utilities and from industrial furnaces. At the 1970 level of control, NOx emissions from stationary sources would approximately double by the year 2000. The need to reverse this trend is clear. On a world-wide basis, man-made sources of N0X produce but a tenth of that produced naturally. But the distribution of man-aade NOx is closely re- lated to population distribution; over 60 percent of the emissions in the United States occur in urban areas* The relative contributions from stationary and mobile sources vary substantially from city to city and the proportionate contribution at ground level is presum- ably more from mobile sources than from stack emissions. Stack height and meteorological factors also affect the relative contributions. N0X has a residence time in the atmosphere of three to four days. Thus, pollu- tion from N0X Is a regional rather than a global problem. 1 ------- Unlike emissions of sulfur oxides (S0X), whij are directly proportional to the sulfur content of the] fuel, N0X is formed largely by the reaction of nitroge^ and oxygen from the atmosphere at the high temperature! existing during combustion. A. smaller contribution is from organo-nitrogen compounds in the fuel. The most promising prospects for significant early reduction of Wx in fuel-combustion stack gases lie chiefly in appl'> cation of some combination of combustion-modification processes to reduce the N0X formed. The probability that processes can be developed for removal of Wx frorh stack gases is not encouraging. B, CONCLUSIONS With regard to combustion-control modificatio! the following conclusions are drawn: 1. Of the three fuels used in firing, gas, oil, and coal, gas allows the most precise control in the attainment of the lowest levels of N0X. The term "coal" covers a variety of types of solid fuels varying greatly in their combustion characteristil and the nature of the ash formed. A vari-i ety of boilers and burners are required to burn these various types satisfactorily Present emission levels from coal firing vary greatly. Of the three fuels, least is known about coal relative to minimizinf N0X formation from combustion. A realieti objective for new plants using natural gad to be placed in operation by 19803 is a reduction in N0X concentration to about 100 ppm from present-day uncontrolled levels, which average about 350 to 400 ppmj but range as high as ls400 ppm. How- ever, natural gas may not be available as a fuel for utility boilers very far into the future. For oil, the most common range today, when the combustion process has not been modified, is about 180 to 280 ppm for tangentially fired units and 300 to 700 ppm for horizontally fired units. A realistic objective for oil-firei 2 ------- ;plants placed in operation by 1980s achiev- able by flue-gas recirculation and off- stoichiometric combustion, is about ISO to 200 ppm. For smaller furnaces or where low-nitrogen oil is available, there is a possibility of reaching this objective at an earlier date. Control methods are not yet established for coal. The Panel recom- mends a review to establish realistic ob- jectives after more data become available in the next two or three years, 2. Both theory and practice indicate that N0X emissions from combustion sources can be lowered by: (a) reducing the amount of oxygen present in the flame zone, as by use of staged admission of air (or off- stoichiometric combustion), and (b) re- ducing the peak flame temperature, as by use of flue-gas recirculation to the flame zone. The practicality of these abatement techniques has been developed primarily in furnaces burning gas or oil. Little has been done on coal-fired units. 3. The principal problem in reducing N0X emissions by the use of staged combustion is to avoid the significant increase of emissions of CO, hydrocarbons, and smoke. In addition, with coal, it is important to avoid increasing the hazard of flame- outs, the rate of corrosion of boiler components^caused by a reducing atmosphere, and the percentage of unburned carbon in the ash. The applicability of the above techniques to coal firing is not well understood and it will vary considerably with burner design; e.g., flue-gas re- circulation may be less effective with cyclone burners than with other burner designs. 4. The amount of NOx formed per unit quantity of heat released on combustion can vary 3 ------- by a factor of about 10 depending on a number of interrelated factors: (a) the fuel (coal, oil, or gas); (b) the per- centage of excess air used in combustion; and (c) the size of the furnace—as fur- nace sizes increase from domestic heating units to large utility boilers, the amount of NOx formed per Btu released usually in- creases, probably because lower surface- to-volume ratios and increased heat-re- lease rates per cubic foot lead to less rapid thermal quenching of the combustion process. For large utility boilers, in- creased furnace volume (while holding all other variables constant) will act to re- duce the average temperatures and will therefore always act to reduce NOx. However, it appears that the effect of furnace volume on N0X formation is of secondary importance when compared with other combustion-control modifications such as off-stoichiometric combustion* This is not to say that furnace volume is insignificant. In designs in which combustion is spread out in the furnace, e.g., tangentially fired units, it seems likely that increasing furnace volume would cause a reduction in N0X; (d) burner design—designs that, produce more intense combustion and higher temperatures, e.g., cyclone burners for coal, produce consid- erably more NOjj than designs that allow combustion to occur out in the furnace, e.g., tangentially fired boilers. It is impractical, however, in an existing in- stallation to replace cyclone burners with tangential burners located in the furnace corners, for this would require nearly complete rebuilding of the furnace. It is not the fact that the burners are positioned to admit fuel and air in a tangential configuration that brings about a reduction in NO formation, but rather the manner in which fuel and air are 4 ------- admitted and mixed; and (e) load—as load is first reduced in any particular instal- lation, the concentration of N0X formed at first drops. With further reduction in load, the change in N0X concentration is determined primarily by the degree to which increased air-fuel ratio may be re- quired to prevent excessive carbon monoxide or smoke at lower loads. 5. Preliminary data exist regarding the re- lative importance of conversion to N0X of fixed nitrogen in oil fuels. These data suggest that combustion of oil or coal under reducing (sub-stoichiometric air) conditions in a first stage helps reduce the amount of fuel nitrogen con- verted into N0X or increases the con- version of NOx to N2 and the combustion products, but additional research is needed. 6. Presently feasible technology that can be applied to NOj,. control in utility boilers varies with the fuel utilized and the nature of the installation—retrofitting an existing boiler or designing a new boiler. Furnace design (e.g., opposed vs. tangential firing), burner configura- tions (e.g.,. cyclone vs. conventional burners), and provisions for adjusting fuel and air flows and recycling combus- tion gases determine the degree of N0X control achievable in an existing Instal- lation. Research, development, and design studies are needed to determine combustor configurations and designs that will facil- itate complete combustion of coal and oil (avoiding carbon, hydrocarbon, and CO losses and preventing smoke and soot for- mation), and will at the same time mini- mize emission of NO^ (formed either by N2-O2 fixation or oxidation of nitrogen in the fuel). 5 ------- 7. Combined-cycle systems involving the com| bustion of coal or oil in combinations ol gas and steam turbines are of considerabl interest to utilities. Some of these maj involve external coal gasification close! coupled to a turbine or boiler. N0X emi^ sions are expected to be low, comparable to those from natural gas, but little in- formation is available. 8. On the basis of laboratory and pilot-sca] tests, fluidized-bed-combustion boilers show promise of appreciably reduced N0X emissions, compared with conventional coal-fired boilers. Burning of the coal is carried out at relatively low tempera- tures, 1400-1900° F.; tests indicate that the N0X emissions originate almost entire, from nitrogen in the fuel. In fluidized combustors burning coal at atmospheric pressure, two-stage combustion has reduced! N0X emissions to about 70 pp®. In pres- surized combustors, NOx emissions have been reduced substantially in single-stag^ combustion. Fluidized-bed-combustion boilers, therefore, show excellent poten- tial for N0x control, and their develop- ment should be pursued. Removal of N0X from stack gases may offer potential for control in the future. However, no proven process is available for substantial removal of Wx from combustion stack gases. Conclusions based on presently available information are: 9. Any wet scrubber system for NC^ removal will be expensive for two reasons: (a) Most of the N0X is in the form of NO, which is relatively unreactive and in- soluble. The maximum rate of absorption in an aqueous system occurs at a NO/NO2 mole ratio of unity, which requires either (1) recycle of NO2 and a method of gener- ating NO2 from the scrubbing system, in 6 ------- turn requiring very high scrubbing effi- ciencies for substantial overall removal of N0X, or (2) oxidation of about half of the NO to NO2 prior to scrubbing. The rate of oxidation of NO to NO2 is slow and de- creases with increasing temperature; and (b) large vessels are required for scrub- bing because of the large volumes of gas that must be handled and the necessity for low pressure drop, 10. Catalytic reduction of N0X to W2 by a re- ducing agent as a process for treating stack gases from large utility boilers requires a sulfur-resistant catalyst if coal or oil is used as a fuel. Space velocity (i.e., catalyst activity) and catalyst life also limit this approach at present. The catalytic reduction of N0X with ammonia or other reducing agents is being studied by several groups. Informa- tion is insufficient to assess fully the potential of any catalytic reduction methods of control. 11. Decomposition of NOx in the absence of a reducing agent requires such high tempera- tures, on even the best catalysts known, as to be impractical. The following conclusions have been reached for other stationary combustion sources and for chemical manufacture: 12. Somewhere between two percent and 21 per- cent of the N0X emissions from stationary sources are produced by internal-combus- tion engines (burning natural gas or die- sel fuel, used in conjunction with pipe- lines and gas plants)* With diesel engines, techniques such as control of fuel injec- tion, exhaust-gas recirculation, water injection, and alteration of combustlon- chambex design, are available to reduce 7 ------- N0X emissions. With gas turbines, rede- sign of combustors and methods of fuel in- jection accompanied by more fuel-lean con- ditions in the combustion zone should pro- duce significant reduction in NOx emissions. 13. More data are required on emissions from industrial process and commercial furnaces, residential furnaces and heaters, and in- cinerators. 14. Emissions from nitric acid plants and chemical operations may be "decolorized" (conversion of NO2 to NO) by catalytic re- duction with natural gas. A similar method can be used to reduce NO to N2 (typically from 3,000 ppm to 100 to 500 ppm), but requires careful control. Ad- sorption by molecular sieves has been shown in the laboratory to produce even lower emission levels but no demonstrated commercial process is yet available. Al- kaline scrubbing may be used but involves liquid-waste disposal problems. 15. In some chemical processing with nitric acid, a substantial portion of the nitro- gen oxides emitted may be in the form of N2O, which is considered to be harmless. C. RECOMMENDATIONS On the basis of its review and in conjunction with the above conclusions, the Panel recommends that: 1. Combustion-modification studies be given first priority in research and development to control N0X emissions. Studies of coal combustion are especially required. Stud- ie8 of the effect of fuel nitrogen on N0X emissions and the potential of flame-tem- perature -control techniques in oil and coal burning are also needed. A sub- stantial reduction in the amounts of N0X 8 ------- released to the atmosphere (i.e., of the order of 50 to 80 percent) will come least expensively from modifications of the com- bustion process rather than from scrubbing or adsorption systems to remove NOx from stack gases. 2. Experimentation to develop firing methods for minimizing NOx emissions be accompanied by data correlation and theoretical analy- ses of the data obtained in order to de- velop the basic understanding for config- uring and designing new combustors and for choosing operational modes in a vari- ety of applications. 3. Boiler manufacturers and utilities incor- porate as much flexibility as possible in the design of new boilers to permit taking advantage in the future of increasing knowledge of the factors affecting N0X emissions in combustion. 4. Additional work be funded on new energy- conversion concepts—such as fluidized- bed combustion, coal gasification for elec- tric-power production, and combined-cycle gas- and steam-turbine generating plants operating in conjunction with such com- bustors and gasifiers—to develop their potential for reducing NOx and other pollutants. 5. New concepts claiming potential for the economic simultaneous removal of NOx and S0X be evaluated carefully. 6. Evaluation of all new electric-power-gen- eration techniques utilizing fossil fuels- -including magnetohydrodynamics (MHD), fuel cells, and combined-cycle plants with fluidized-bed combustion or coal gasifica- tion—incorporating estimates of K0X for- mation and the economic cost of NO* con- trol, begin as soon as practicable. 9 ------- 7. Careful consideration be given to improv- ing present methods of sampling and analysis of NO and NO2, particularly in the presence of other pollutants from stationary sources. 8. The potential for the generation of N0X by such sources as stationary internal- combustion engines, industrial and commercial furnaces, residential furnaces and heaters, incinerators, electrostatic precipitators, and other high-voltage equipment needs further evaluation. If the level of emissions is significant and the effect on ambient air quality is detri mental, then control techniques should be sought. 10 ------- II INTRODUCTION A. THE CLEAN AIR AMENDMENTS OF 1970 The increasing concern of Americans for the deterioration of the nation's air resources is re- flected in the Air Quality Act of 1967 (Public Law 90-148 as amended)2 and the Clean Air Amendments of 19701 (Public Law 91-604 in which Public Law 90-148 was further amended to provide for a more effective program to improve the quality of the nation's air). These acts include emission standards for existing sources that are to be enforced by the states, and performance standards for new sources that were promul- gated during 1971 and are to be enforced by the Environmental Protection Agency (EPA). Responsibility for carrying out the provisions of these laws is assigned to the EPA Administrator. Under the provisions of these laws, the EPA published in the Federal Registerguidelines to the states for preparation, adoption, and submittal of implementation plans for enforcement of national ambient air quality standards. On April 30, 1971, the EPA published the National Primary and Secon- dary Ambient Air Quality Standards.4 National pri- mary and secondary ambient air quality standards are those judged by the Administrator, based on the air quality criteria documents8*9»10»12*13*lif published by the EPA and allowing an adequate margin of safety, as requisite to protect the public health (primary standards) and the public welfare (secondary standards) from any known or anticipated adverse effects associated with the presence of pollutants in the ambient air. The national primary ambient air quality standards for nitrogen oxides were promulgated as follows: The national primary emd secondary ambimt air quality standard for nitrogen dioxide 3 measured 11 ------- by the reference method described^ (sodium hydroxide absorption and subsequent colorimetrie determination)* or by an equivalent method, is: 100 micrograms -per aubia meter (0*055 parts per million)—annual arith- metic mean. Under Section 110,1 the states have nine months to submit a plan for implementation, maintenance, and enforcement of the primary standard in each air- quality-control region and an additional nine months to submit a similar plan for secondary standards. The plans are to include a timetable for compliance within three years, and state standards may be more stringent than the federal standards. For nitrogen dioxide, the national secondary ambient-air-quality standard is the same as the primary standard. In applying these standards it is assumed that any NO present in the air is converted to the di- oxide, but the extent to whiah NO is actually converted to NO2 can vary greatly. Recommendations by EPA5 to the states for possible inclusion in their implementation plans for existing sources cover emissions of nitrogen oxides from fuel-burning equipment and nitric acid manu- facture. These emission standards together with standards of performance for new stationary sources7 are as follows: Nitric Acid Fuel-Burning Equipment Manufacture (pounds per million Btu) (pounds per ton) Gas Oil Coal Required for New Sources 0.20 0.30 0.70 3.0, 10 per- cent opacity Recommended for Implementation 0.30 0.30 5.5 Plans* (175) (230) (575) (400) *The figures in parentheses are approximate ppm, in dry flue gas containing 3 percent oxygen* 12 ------- The reasons for concern about N0X emissions to the atmosphere are threefold: (1) Both NO and NO2 can of themselves have adverse health effects, NO2 being considerably more toxic than an equal concentration of NO14; (2) NO and NO2 interact with hydrocarbons in the atmosphere under the influence of the ultraviolet energy of sunlight by a highly complex and only partially under- stood series of reactions, to generate eye irritants such as PAN (peroxyacetyl nitrate and related compounds), PBZN (peroxybenzyl nitrate and related compounds), and Oq (ozone) .9 *111 In addition, the reddish color of N0« (NO is colorless) can contribute to haze and decreased visibility; and (3) detrimental effects on vegetation.14 Following combustion, NO reacts relatively slowly with residual oxygen in the combustion equipment and in the flue gases to yield N02. This conversion is usually less than 10 percent because of the short resi- dence time. Thus, NO released to the atmosphere from combustion processes is largely in the form of NO*14*15 In the typical daily urban pattern, the concentration of NO and NO2 in the atmosphere begins to increase rapidly at dawn as human activity, particularly auto- motive traffic, increases, the ratio of NO to NO2 initially being very high. Under the influence of sunlight, much of the NO is converted during the day to NO2 and concentrations of PAN, PBZN, and O3 increase. After sunset, photochemical conversion of NO to NO2 ceases but O3 formed during the day continues to react with NO to form N02 until O3 is depleted. In ordinary circumstances, winds and dispersion during the night reduce the concentration of pollutants to a low level and the cycle recommences the following dawn, What happens to as much as 50 percent of the nitrogen oxides that become incorporated into the photochemical complex is still undetermined, since many of the nitrogen oxide end products remain unidentified.14 B* ROLE OF THE ACADEMIES The National Academy of Sciences and the National Academy of Engineering established the Environ- mental Studies Board in 1967 to coordinate activities of the two Academies in the environmental field. One 13 ------- of the first acts of this Board was to recommend estab- lishment of four committees on air, water, noise, and solid-waste management within the Division of Engineer- ing of the National Research Council. These committees have an engineering orientation and are available for furnishing advice and assistance to the Congress and tat agencies of the executive branch of government having responsibility for pollution abatement and control. In June, 1969, the National Air Pollution Conj trol Administration of the U. S. Department of Health, Education, and Welfare (now the Office of Air Programs of the EPA), requested the National Academy of Engineerj ing to make a comprehensive review of present industry : and government research, development, and demonstration programs, to include technical and economic potentials, adequacy of scope, proper integration with similar ef- forts, and responsiveness to national needs directed to ward control of sulfur oxides effluents from stationary sources of combustion. The report on that study, Abate* ment of Sulfur Oxide Emissions from Stationary Combus- tion Sources^ was published in May, 1970. The newly formed Office of Air Programs (OAP) of the EPA then requested the National Academy of Engineering—National Research Council to follow this study with studies of sulfur oxides emissions from in- dustrial sources, nitrogen oxides emissions from sta- tionary sources, and particulate emissions from station- ary sources. Each study includes an assessment of the adequacy of present technology; the need for additional technology; the scope, integration, and coordination of industry and government effort; and responsiveness to the national need for control and abatement of air pollution from each of the three sources. Other similar studies are expected to follow within the next year* This report deals with the emissions of nitrogen oxides. C, FEDERAL RESEARCH, DEVELOPMENT, AND DEMONSTRATION PLANS The EPA has developed a five-year research and development plan for control of N0X emissions from stationary sources.17*18 The Panel hopes this report 14 ------- will be useful to the EPA and its Office of Air Programs, other government agencies, industry, fuel suppliers, and others in planning and coordinating the research, de- velopment, and demonstration programs required to meet the national need for cleaner air. The implementation of this plan will need to be closely coordinated with national energy policies for optimum use of resources.19 The major emphasis will be placed on developing combustion-modification techniques that will reduce N0X formation in coal-, oil-, and gas- fired industrial and utility boilers. Electric utility sources have received the highest priority for funding in the Development Engineering Branch of the EPA's Office of Air Programs for the following reasons: (1) Electric utilities are the largest single source of N0X emissions; (2) the internal-combustion engines that produce the N0X in conjunction with pipelines and gas plants can be controlled with technology developed by OAPfs motor vehicle research and development program and by the automobile industry; (3) industrial, domestic, and commercial combustion sources may be controlled by selectively adapting the technology developed for elec- tric utility sources. Combustion in furnaces produces about two-thirds of the NOx emissions from stationary sources in the United States and will be responsible for an increasing share of this total. Combustion-mod- ification techniques, such as low-excess-air combustion, staged combustion, flue-gas recirculation, and water or steam Injection, will be investigated up to full-scale application. Basic theoretical and experimental com- bustion studies will be performed through fiscal year 1975 in support of the large-scale work. The complex nature of control of NO^, SO2, and other pollutants from coal combustion, and to a, lesser extent from the combustion of other fuels, fully justifies serious federal attention and support for de- velopment of processes for production of "clean" fuels by such methods as fluidized bed gasification, coal gasification, coal liquefaction, and processing of our vast deposits of oil shale. This requires careful planning and coordination by OAP with other federal agencies especially the Office of Coal Research and 15 ------- the Bureau of Mines. Projections for use of fuel re- sources20 should include considerations such as burning low-sulfur fuel in commercial and industrial boilers and higher sulfur fuels in electric utility boilers. The second area of emphasis for the next five years will involve the development of combustion flue- gas treatment processes to remove nitrogen oxides once they have been formed. The control of N0X emissions by combustion modification may be limited by special pro- cess requirements for very high temperatures, limited space restrictions, or possibly the inherent fixed- nitrogen content of the fuel. Flue-gas treatment tech- niques, including selective catalytic reduction, aqueous alkaline scrubbing, and other selective adsorption and absorption techniques, will be scheduled for research and development. The third area of emphasis through fiscal year 1975 will be the development of needed control technology for processes that manufacture or use nitric acid. Such processes emit small quantities but high concentrations of nitrogen oxides. Limited control technology is already available, but OAP support may be necessary to encourage further improvements in such techniques as catalytic reduction, alkaline scrubbing, and adsorption. This area may profit from developments in combustion flue-gas cleaning. The EPA budget for stationary-source nitrogen oxides control in FY 1971 was just over $1 million. In FiT 1972 it is $1.5 million; it is projected to be $4.2 million in FY 1973, $8.0 million in FY 1974, $12.0 million in FY 1975, $23.0 million in FY 1976, and $17.0 million in FY 1977. Funding is then expected to de- crease sharply. The Panel concludes that the research and development plans and the funding levels proposed by EPA are prudent and are required for achievement of national goals as presently defined. Most of the information on health hazards associated with N0X, on ambient air and emission stan- dards, on combustion-modification methods, and on pro- gress in control of N0X from mobile sources has been 16 ------- developed very recently, much of it during the course of the Panel meetings. The Panel believes it to be de- sirable to issue its findings now even though the rapidly developing understanding of this problem may cause some revisions in judgment in the near future, particularly with respect to those sources for which standards are soon to be established. The Panel's assessment was completed before all the background information was available from EPA on the proposed new source perfor- mance standards.6»7 17 ------- Ill SOURCES OF NITROGEN OXIDES Of the various oxides of nitrogen, the most important as air pollutants are nitric oxide (NO) and nitrogen dioxide (NO2)» generally grouped together and referred to as nitrogen oxides (NQx). Other oxides of nitrogen, N2O3, N2O4, N2O5, and NO3 are present in very low concentrations, and although they may participate in photochemical reactions they can be neglected for present purposes. Nitrous oxide or laughing gas CN2O) is present in the atmosphere in concentrations of about 450 ug/m3 (0.25 ppm)14 but is regarded as being harmless physiologically, and there is no evidence that it par- ticipates in photochemical reactions in the lower at- mosphere, ll+ Most N0X is produced biologically; fixation by lightning seems to be relatively unimportant.15 Natural sources produce on the order of 50 x 107 tons per year (about 90 percent of the total) of NOx, world- wide, while man-made sources emit about 5 x 107 tons per year (about 10 percent of the total). Naturally occurring sources of N0X have been found to cause non- urban concentrations, usually less than 10 ug/m3. How- ever, urban concentrations are generally 10 to 100 times higher, which indicates the importance of the man-made sources even though, overall, they are a small fraction of the natural sources. Nationwide, the man- made sources that are the major cause of N0X emissions are fuel combustion in furnaces and in engines.15'21 Industrial chemical processes are responsible for high, localized emissions, but are not significant on a large scale. A summary of estimated man-made N0X emissions in the United States in 1969 is given in Table 1 and a breakdown of estimated emissions from stationary sources, 1968, is given in Table 2. About 53 percent of the total Wx emissions in the United States are from stationary sources. The largest contributions are from fossil-fuel'fired boilers for electric utilities and from industrial furnaces. At the 19?0 level of control, N0X emissions from 18 ------- TABLE 1 SUMMARY OF ESTIMATED MAN-MADE NITROGEN OXIDES EMISSIONS IN THE UNITED STATES, 1969 (Environmental Protection Agency, May 1971) Source 106 Tons/Year Transportation 11.1 Motor vehicles, gasoline 7.6 Motor vehicles, diesel 1.1 Aircraft 0.3 Railroads 0.1 Vessels 0.2 Nonhighway 1.8 Fuel Combustion in Stationary Sources 10.0 Coal 3.8 Fuel oil 1.3 Natural gas 4.7 Wood 0.2 Industrial Process 0.2 Solid waste 0»4 Miscellaneous 2.0 Forest fires 1.6 Structural fires Coal refuse 0.1 Agricultural 0.3 TOTAL 23.7 19 ------- TABLE 2 BREAKDOWN OF ESTIMATED NOx EMISSIONS FROM STATIONARY SOURCES IN THE UNITED STATES, 1968 (Environmental Protection Agency, May 1971) Source Percentage of Total Electric utility boilers 38 Industrial combustion 29 Pipelines and gas plants 21* Domestic and commercial 10 Noncombustion sources 2 *Estimates range from 2 percent to 21 percent. (See remarks on page 39.) 20 ------- stationary sources would approximately double by the year 2000.17 N0X emissions from stationary sources in the United States result primarily from the combustion of fossil fuels in boilers, furnaces, and stationary in- ternal-combustion engines. The bulk of the N0X emitted by industrial sources comes from steam boilers and pro- cess heaters, with smaller amounts from internal-com- bustion engines, boilers burning waste-fuel gases, catalytic cracking regenerators, metallurgical ovens, furnaces, and kilns. On the other hand, emissions from pipeline and gas plant operations, result primarily from the use of internal-combustion engines to drive pumps and compressors, and»to a lesser extent from pro- cess heaters and boilers. Domestic and commercial sources include incinerators, space heaters, water heaters, ranges, and clothes dryers. Essentially all the N0X emitted by noncombustion sources comes from the manufacture or use of nitric acid. N0% distribution is closely related to popu- lation distribution; over 60 percent of the Wx emie- sions in the United States occur in urban areas. The relative contributions from stationary and mobile sources vary substantially from city to city and the contribution at ground level presumably comes propor- tionately more from mobile sources than from stack emissions. Stack height and meteorological factors also affect the relative contributions» has a residence time in the atmosphere of three to four days. Thus, pollution from a regional rather than a global problem. Almost all N0X emissions arise from combus- tion processes. The nitrogen oxides are formed during combustion by chemical combination of oxygen and nitro- gen in the air (fixation), and by oxidation of some of the nitrogen combined with organic substances in the fuel. Only a small fraction of the nitrogen in the air passing through a furnace or an engine is converted to N0X, Likewise, only a portion of the nitrogen in the fuel is converted to N0X. However, the total tonnage of N0X emitted, if it were converted to nitric 21 ------- acid, is 10 times the U.S. production of this chemical. Coal varies between about 1.0 and 1.5 percent by weight and fuel oil between about 0.02 and 1.2 percent by weigh in the amount of combined nitrogen they contain. Natura gas contains no significant amount. Unlike emissions of S0Xi which are directly proportional to the sulfur content of the fuel> NOx is formed in conventional combustion processes largely by fixation of nitrogen and oxygen from the atmosphere at the high temperatures existing during combustion. Under conventional combustion conditions a smaller contribution is from organo-nitrogen compounds in the fuel. The major factors influencing N0X emissions from combustion processes are: the percent of excess air; mixing; heat release and removal rates; fuel type; and fuel composition.17*48 These factors provide a basis for modifying combustion techniques of N0X con- trol. Modification of operating conditions include: operation at reduced load; low excess-air firing; staged or off-stoichiometric combustion; flue-gas recirculation steam or water injection; reduced air preheat; and com- binations of these methods. The N0X level can also be reduced by changes in design such as: altering burner- design configuration, location, and spacing; use of tangential firing; and, potentially, fluidized-bed combustion. The meteorological aspects of the N0X pollu- tion problem can be considered in three general areas, which are not necessarily mutually exclusive in all re- spects:14 1. The transport and dispersion of nitrogen oxides—of particular importance in urban areas. 2. The relationship between concentrations of oxides of nitrogen and meteorological factors, especially in relation to trans- formation processes in the atmosphere. 22 ------- 3. The effects of NO2 concentrations on visibility and sky coloration* Studies in Los Angeles give the following re- sults for the relationships between oxides of nitrogen and meteorological factors: 1. A correlation was found between the "hourly average concentrations" of hydrocarbons (HC) and N0X 2. The HC/N0X and HC/CO ratios indicate that NO concentration is a complex function of temperature. 3. The maximum daily N0X concentration is inversely related to inversion height and wind speed and is independent of light intensity. 4. The maximum daily oxidant concentration is a function of the square root of light intensity. Similar studies In Cincinnati^ generally con- firm these and other observations. For example: HC, NO, and CO reach their peak concentrations at about 8:00 a.m., NO2 about 9:30 a.ia. » while O3 increases monotonically through the period. The ratios C0/N02, CO/NO, and NO2/NO increase rapidly on warm days, where- as the NO2 peak occurs later on cold days. However, to a considerable extent each city is unique. The rela- tive ratios and quantities of pollutants emitted and their effect on ambient conditions vary greatly because of differences in fuels burned, types of boilers and furnaces, nature of industry and of transportation, and meteorological conditions. Although nationwide emission standards are being established, it is impor- tant that better models be developed for individual cities and local air-quallty regions with respect to the sources and fate of air pollutants, so that local decisions can reflect optimum strategies both for con- tinuous control and for action under emergency conditions. 23 ------- The national ambient air quality standards for NO rest very heavily on a series of studies in Chatta- nooga, Tennessee, on school children. 3»i+* 5»5* 7»22*23 These standards call for a mean concentration of NO2 not exceeding 100 pg/m3 during the year. It is not clear, however, that these studies provide a sufficiently broad data base for firmly establishing ambient air stan- dards for N0X. It involved primarily children, and the majority of the N0X was produced by a single source. The fluctuations of N0X concentrations around mean values (including relatively higher peak concentrations) are wider with a single source than witfi multiple sources. Health and environment effects may be more attributable to peak concentrations than to mean values. The current ambient-air standard for NOx is probably being exceeded in portions of major cities of the United States; there- fore, additional data should be sought to confirm this standard. The current view of the EPA appears justified in that it is reasonable and prudent when promulgating air quality standards, to give consideration to re- quirements for a margin of safety that would take into account possible effects on health, vegetation, and materials that might occur. However, the Panel does not feel competent to evaluate the standard itself, from the viewpoint of hazard to health and welfare. A national strategy may be needed to achieve substantial reduction of N0X levels in urban areas. Ground-level concentrations appear to be related to mobile sources and domestic heating more than to large single sources that discharge through tall stacks. A program of monitoring, modeling, and land-use planning will need to be developed for each urban area in order to obtain acceptable air quality with respect to nitro- gen oxides and other pollutants. In 1970, electric utilities placed new orders for electric generating equipment of about 30 million kilowatts capacity to burn fossil fuels and of about 15 million kilowatts capacity using nuclear fuel. For various reasons it is expected that this ratio of two to one for new equipment using fossil and nuclear fuels will be maintained for the next several years, thus ensuring that the generation of N0X will be a continu- ing significant problem. Electric utilities in the 24 ------- past have been dependent upon equipment manufacturers for research and development. Control of NOg and other pollutants may require them to engage in research and development to solve their individual problems. Encour- agement to do so will depend heavily on the attitudes of the regulatory agencies, which must approve the rates charged by the utilities for electrical energy. Fuel use involves many factors* One has been the increasing emphasis on air-pollution control and the substitution of low-sulfur oils for some of the high-sulfur coals. There has also been a balancing of installed equipment in utilities systems to use low- capital -cost t peaking-type (gas ox oil) equipment to supplement the base-loaded nuclear equipment as it is added to a system, so some of the growth in oil firing during the recent past was of the peaking variety, and some of it was the substitution of oil for coal (espec- ially along the Eastern Seaboard) where oil can often be brought directly into the power plant. Considerations of fuel availability and strategies for N0X control could lead to conversion of smaller sources to gas. other factors are expected to spur research and deve op- ment of coal gasification, fluidized-bed combustion, and changes in combustor design. It is generally agreed that at present the state of the art for gas and o 1 firing permits the concentration of NO in the stack gas to be reduced to somewhere in the range of 150 to 250 ppm* on large units employing one or more of the latest combustion modifications. The importance of air-pollution control is clear. However, an economic price must be paid and the cost of reduction of any pollutant, per unit quantity of *The concentration of NO* is obviously affected by the amount of excess air used. To have a common basis or comparison it is customary to calculate the NO* concen- tration that would exist on a water-free basis if the oxygen concentration in the exit gases were 3 percent. For regulatory and other purposes, it may be desirable to express N0X emissions in terms of quantity produced per Btu released or kilowatt-hour generated. 25 ------- that pollutant, increases the higher the percentage re- moval specified. For lower concentration streams the cost per unit removal also increases with the total volume processed. These economic costs can be major and will ultimately be borne by the economy as a whole. They should not be glossed over in the nation's drive toward cleaner air. For those pollutants that already are or may become global problems, it is important that global standards be adopted. 26 ------- IV FORMATION AND CONTROL OF EMISSIONS FROM COMBUSTION SOURCES A. EQUILIBRIUM AND KINETICS It appears that most of the NO is formed "by the high-temperature reaction of molecular nitrogen and oxygen present in the combustion air via the following chain-reaction mechanism:2I+ O2 $ 0* + 0* 0- + N2 $ NO + N« N- + 02 $ NO + 0- Some of the NOx doubtless also results from the oxida- tion of organically bound nitrogen in the fuel. The fixation of nitrogen atid oxygen is thus represented by the overall reaction: N2 + 02 » 2N0 The rate and extent of this reaction increases very rapidly with increased temperature. The theraodynaiaic equilibrium constant is given by:^7 K ¦ 21.9 exp (-43.40G/RT) The formation of NO is favored at high temperatures. Equilibrium concentrations of NO at flame temperatures are of the order of 3,000 ppm, while the NO2 levels are negligible. The concentrations of NO actually found in the stack gases are usually much less than the equilib- rium concentration that -could ©x:U£ at the £lame Tem- perature in the furnace but much greater than that corresponding to the temperature of the gases leaving the furnace. It seems likely, then, that the NO con- centration is determined by the time-temperature-com- position history of the gases as they move through a 27 ------- furnace or other combustion system. This theory is supported by theoretical chemical kinetic calculations and experimental studies on premixed flames that show that the NO is formed after most of the fuel combustion has been completed. These experiments also indicate that the NO-formation process is strongly temperature- dependent (the most rapid rates occurring at peak com- bustion temperatures) and that little, if any, NO de- composition occurs. It is therefore concluded that the amount of NO emitted in gas- and oil-fired systems is essentially determined by the characteristics of the forward reaction forming NO. For coal firing, however, the amount of NO in the gases may go through a maximum and decrease as the gases cool. It is possible that the ash may have a catalytic effect in decomposing some of the NO initially formed at the higher temperatures. Al- though NO2 is favored thermodynamically over NO at lower temperatures, relatively little is formed during the rapid cooling of combustion gases in passing through a furnace and in dispersion into the atmosphere. About 90 to 95 percent of the N0X dispersed into the atmosphere is in the form of NO. Fundamental differences remain in understand- ing just how NO is formed in a furnace. One viewpoint applicable to one-dimensional premixed air-fuel systems, which lends itself readily to quantitative modeling, re- gards most of the NO as being formed after much of the fuel combustion has been completed. The NO concentra- tion is thus regarded as being determined by the time- temper ature-compos it ion history of the gases as they flow through a furnace or other combustion system* An- other viewpoint applicable particularly to gas fuel, but also to some extent to oil and coal, considers most of the NO in a furnace (which is not operated to minimize NO formation) as being formed in the primary combustion zone within a few feet of the burner. NO is visualized as being formed as pockets of fuel react in the primary zone near the burners at near-stoichiometric conditions- -conditions essentially unrelated to the burner air-fuel ratio. The burner air-fuel ratio, rather, determines the number of pockets or eddies that react at near- stoichiometric conditions, each reaching peak tempera- ture and forming high NO. The amount of NO formed in 28 ------- each pocket depends upon the temperature-concentration history of the products in each pocket as it passes out into the bulk-gas region, mixing with air and products of combustion as it goes. Firing fuel-rich may be con- sidered as reducing the number of pockets reacting in the primary zone, thus reducing NO formation. The amount of NOjj formed in boilers can, in principle, be calculated by solving the coupled set of three-dimensional differential equations describing the transfer of mass, momentum, energy, and species. Such a solution would include the effects of flow pat- tern, interaction between flow patterns, molecular and turbulent diffusion, chemical kinetics, two-phase flow, and radiation-heat transfer. However, due to basic uncertainties concerning how to describe such phenomena as turbulence and to the complexity of solving these equations, the mathematical models are necessarily a simplification of the real situation. They have made valuable contributions to understanding the basic pro- cesses relating to NOx formation and thus to suggesting ways of modifying boilers for lower NOx production. They cannot, at present, accurately predict the change in NOx formation that will be caused by a given modification. These mathematical models are undergoing constant refine- ment as their predictions are tested against data and as the effects of various simplifying assumptions be- come clearer. They may be used as aids to the planning of laboratory and field testing of combustion-modifica- tion techniques, in the evaluation of data from such testing, and in the design of new boilers and modifica- tion of old ones to achieve minimum NOx. Experimentation to develop new firing methods for minimizing NOx emissions should be aooompccnied by theoretical analyses -of the data, obtained in orderto develop mathematical models of general usefulness for prediotion, B. FACTORS AFFECTING UTILITY BOILER EMISSIONS The major factors affecting the formation of nitrogen oxides in combustion processes, based on both theory and practice, are: 29 ------- 1. Combustion temperature, NO-formation equilibrium and kinetics are extremely dependent upon peak combustion temperatures, with higher peak tempera- tures favoring higher emissions. 2. Availability of combustion air. NO for- mation is dependent upon the availability of air for the "fixation" reaction shown. 3. Mixing of fuel3 airA and combustion pro- ducts. Internal recirculation or "backmixing" of com- bustion products into the combustion zone dilutes the fuel and air, lowers the flame temperature, and thereby reduces N0X emissions. Distribution of the fuel and air so as to achieve most of the combustion under fuel- rich conditions and the rest under fuel-lean (excess air) conditions also reduces N0X emissions. Slow diffusion of the fuel and air streams can also accom- plish this objective. 4. Heat release and removal. Low heat-re- lease rates and high heat-removal rates reduce NO for- mation, because lower peak temperatures and shorter residence times at high temperatures are achieved. 5. Fuel type. On an equivalent heat-input basis, using modified combustion techniques, coal firing usually emits the most N0X; oil emits less; and gas the least, N0X emissions from combustion sources can be reduced by two major techniques—combustion modification and flue-gas treatment. Combustion-modification tech- niques presently appear to be the quickest and most economical methods of accomplishing major reduction of N0X emissions. Those techniques¦that have been con- ceived (some have been tried) are low-excess-air opera- tion, staged or off-stoichiometric combustion, flue-gas recirculation, reduced air preheat, steam or water in- jection, or a combination of these techniques. Low-excess-air combustion reduces NOjc emis- sions because the availability of oxygen is reduced. Staged or off-stoichiometric combustion is the technique 30 ------- of introducing combustion air in stages so that the par- tially burned fuel and combustion products are permitted to cool before the completion of combustion; thus, the simultaneous exposure of the nitrogen to oxygen and to high temperatures is avoided. Much of the combustion occurs under fuel—rich conditions* External recirculation of flue gas into the combustion zone, like internal "backmixing," dilutes the fuel and air, lowers the flame temperature, and thereby reduces NQX emissions. Injection of steam or water into the combustion zone or the reduction of air preheat also has the same effect. Flue-gas recirculation has been practiced for many years as a method of adjusting tem- perature distribution in a boiler, but this has usually not involved recirculation back to the burner region, which is required for optimum NOx control. Low-excess-air staged-combustion and gas- recirculation techniques have been used successfully by some California utility companies firing gas25 and applicability to oil firing is being investigated by these same companies.26 However, very little work has been done to apply these and other combustion-modifica- tion techniques to coal-fired units. The alteration of firing procedures may also affect the heat-transfer characteristics of a boiler and hence the thermal efficiency, consequences that must be kept in mind in searching for suitable modifications. Flue-gas treatment, the second technique, is the removal or decomposition of the NOx after it has been formed and is described in more detail in Chapter V of this report. The most promising candidates for further development are aqueous scrubbing systems using alkaline solutions or concentrated sulfuric acid. To be effective, however, the N0X must be scrubbed out in equimolar concentrations of NO and NO2* Since less than 10 percent of the NOx present in stack gases is NO2 and the remainder is HO, some method of equalizing the concentrations is needed. The best method appears to be the recycle of N02» since catalytic oxidation of "NO to N02 in the main flue-gas stream is both too expensive and too slow. Other flue-gas treatment 31 ------- techniques that are candidates for development are cata- lytic reduction, adsorption by solids, and catalytic de- composition, the last being probably the least promising. The amount of N0X formed per unit quantity of heat released an uncontrolled combustion can vary by a factor of about 10 depending on a number of inter- related considerations: 1. The fuel (coal3 oita or gas), 2. The percent excess air used in combustion. Z. The size of the furnace. As furnace sizes increase from domestic heating units to large utility boilers, the amount of N0X formed per Btu released usu- ally increases, probably because of lower surface-to- volume ratios, which lead to less rapid thermal quench- ing of the combustion process. For large utility boilers, increased furnace volume (while holding all other variables constant) will act to reduce the average temperatures and will therefore always act to reduce N0X. However, it appears that the effect of furnace volume on NQX formation is of secondary importance when compared to other combustion-control modifications such as off-stoichiometric combustion. This is not to say that furnace volume is insignificant. In designs in which combustion is spread out in the furnace (e.g., tangentially fired units) it seems likely that in- creasing the furnace volume would cause a reduction in H0X. 4. Burner design. Designs that produce more intense combustion and higher temperatures, (e.g., cyclone burners for coal) produce considerably more N0X than designs that allow combustion to occur out in the furnace, (e.g., tangentially fired boilers). It is impractical, however, in an existing installation to replace cyclone burners with tangential burners lo- cated in the furnace corners, for this would require nearly complete rebuilding of the furnace. It is not the fact that the burners are positioned to admit fuel and air in a tangential configuration that brings about a reduction in NO formation, but rather the manner in 32 ------- which fuel and air are admitted and mixed. It may be possible in some instances to incorporate the mixing features of tangential burners into wall-fired burner designs as more is learned of these features. In other instances where cyclone burners are required because of the specific type of coal used, this may not be possible. 5. Load. As load is first reduced in any particular installation, the concentration of N0X formed at first drops. With further reduction in load, the change in N0X concentration is determined primarily by the degree to which increased air-fuel ratio may be required to prevent excessive carbon monoxide or smoke at lower loads. Certain residual fuel oils, e.g., those from California and Venezuela, have a relatively high content of organo-sulfur and organo-nitrogen compounds. Fuel oils from California are especially high in organo- nitrogen compounds. The nitrogen content is usually lower than the sulfur content in oils that have not been desulfurized. The range of the nitrogen in Califor- nia residual fuel oil is about 0.3 to 1.0 percent, and for a Venezuelan fuel oil, from 0.1 to about 0.5 percent. Desulfurization of fuel oil also removes some of the nitrogen and the fraction of nitrogen removed increases as operating conditions are altered to increase sulfur removal. With firing of residual fuel oil in conven- tional equipment, as a rough guide, perhaps about a third of the fixed nitrogen is emitted as N0X. Many coals contain substantial quantities of fixed nitrogen. One percent nitrogen in coal would re- sult in about 1,500 ppra of N0? in the stack gases (stoichiometric conditions) if it were all coiivferted to N0X. Even though much is in fact released in the form of S2, the remainder can be a significant contri- bution to total HQX emissions. Pvetirninapy d/oLtaexist regard'ingthe relative importance of conversion of fixed nt-tvogen vn &&lvfuet$ to N0X. These data suggest that combustion of oil ox* coal under reducing (sub-stoiohiometria air) con- ditions in a first stage helps reduce the amount of 33 ------- fuel nitrogen converted into NOx or increases the con- version of N0X to N2 and combustion products, but addi- tional work is needed. Boiler manufacturers are being asked to pro- vide guarantees on the maximum amount of NO^ that will be emitted. Of the three fuels used in firvng3 gass oilj and coaly gas allows the most precise control in the attainment of the lowest levels of N0X, The term "coal" covers a variety of types of solid fuels varying greatly in their combustion characteristics, and there- fore a substantial variety of boiler and burner designs are required to burn these various types satisfactorily. Present emission levels from coal firing vary greatly. Of the three fuels3 least is known about coal relative to minimizing N0X formation from combustion. The Panel is aware of only three instances in which coal-fired utility boilers have been operated with staged admission of air or two-stage combustion. The results were en- couraging and need to be extended. A long lead time is required for the sequence of research, development, design, construction, and operation. A realistic goal for 1980 for firing of natural gas in new plants is a reduction in concen- tration to about 100 ppm from present-day values, which are about 350 to 400 ppm uncontrolled,, but range as high as 1J400 ppm. However, natural gas may not be available as a fuel for utility boilers very far into the future. For oils the most common range today when the combustion process has not been modified is about 180 to 280 ppm for tangentially fired units and ZOO to 700 ppm for horizontally fired units. A realistic goal for 1980 in new plants, achievable by flue-gas recirculation and/or staged combustion, is 150 to 200 ppm. Both theory and practice indicate that N0X emissions from combustion sources are lowered by: (a) reducing the amount of oxygen present in the flame zone, as by use of staged or off-stoichiometric combus- tion admission of air, and (b) reducing the peak flame temperature, as by use of flue-gas recirculation. The practicality of these techniques has been tested pri- marily in furnaces burning gas or oil. Limited data 34 ------- make it appear that flame-temperature-con caseous may not be as beneficial with oil a® w fuel. Little has been done on coal-fired un The principal problem in reducing N0X e*^sions by the use of staged and off~stoichl^eJl!fJi^vUi,l_ is achieving the reduction without sign or creasing emissions of CO, hydrocarbons, an » «ith coal, increasing the hazard from creasing unbumed carbon in the ash. e ^nns4(jerably of these techniques to coal firing ^V^^^nd with with the slagging characteristics of the coal and with burner designj e.g., gas recirculation may hurner effective with cyclone burners than with other designs, if gas is recirculated into cause most of the combustion occurs within burners. However, if a modified operation ( fltion burner redesign) could be established for rec e_ directly into the cyclone, this approach mig ful. Costs for boiler modifications are , varied. For example, as of March 1971, J-os®* "J Power plant costs range from $100 to $3 P _ of capacity. The boiler alone costs $35 to . kilowatt. 3?or large new gas-fired units the of flue-gas recirculation (to the windbox) adds about $2.65 per kilowatt. The cost of staging would add about $0.15 to $0.2 m _ * ar*. of the fired units some recirculation is generally P basic designj to modify it to recirculate to box adds about $1,50; over-fire air $0.50. $3.50; staging costs an additions • $ . ^ an additional Oversizing a boiler by 50 percent wi „ used. $3.75 to 57.50 per kilowatt depending ' The potential benef its of ln oomparison with reduction of NO appear to be small in P ^lstinjj other methods of NO control• units for gas recirculationor redis ^ clted above. vould be much more expensive than The definition of what cons« "V Plant is difficult but iaportant, since the^firsts^ of emission standards are specified PP y 35 ------- plants." For nitrogen oxides particularly, most of the control methods now available depend on boiler modifica- tions, so it would appear that once the purchaser has specified the boiler, the plant can no longer be termed new. There is about a seven-year pferiod between initial commitment by the utility and the operation of a new power plant. The first set of ambient air standards that will become applicable in 1975 is based, first, on what is desirable, and second, on what is achievable. With respect to utility boilers, presently feasible technology for control of NO varies greatly with the fuel used, the nature of the Durners (e.g., cyclone versus tangential burners), and the age of the unit. Boiler manufacturers and utilities should in- corporate as much flexibility as possible in the design of new boilers to be able to take advantage in the future of the growing knowledge of the factors affecting N0X emissions in combustion. A long time scale is involved between design and actual operation of a boiler which tests out-the effect of substantial mod- ifications. Because boilers are highly complex, only a two-to-one to four-to-one extrapolation on size in construction can be justified and much testing must be done on full-size units. Thus the testing of new con- cepts is a lengthy and expensive procedure, C. FLUIDIZED-BED COMBUSTION Coal (or oil or gas)*can be completely burned at l,400p F. to 1,8Q0° F. within a bed of limestone, dolomite, ash, or inert particles suspended by the flow- ing combustion air and products* Such a system promotes high volumetric heat release, and high heat-transfer rates to steam tubes submerged in the particle bed. Boilers can thus be made smaller and cheaper. Such boilers may also reduce atmospheric pollution. The limestone or dolomite absorbs most of the sulfur con- tent of the fuel and the production of N0X is reduced due to the low combustion temperature in the bed. The quantity of fine particulates produced is also less be- cause typically the coal is in the size of 100 mesh to 1/4-inch material in contrast to 200 mesh in pulverized coal burners. 36 ------- Research and development in fluidized-bed com- bustion was begun by the British Central Electricity Generating Board and has been continued by the National Coal Board (NCB) in Great Britain and by the U.S. Depart- ment of the Interior's Office of Coal Research (OCR), and more recently by contractors of the OAP of the EPA in the United States. Steam-generating pilot plants have been built and operated at both atmospheric and superatmospheric pressures. Fluidized-bed boilers can be operated at at- mospheric pressure as a replacement for conventional pulverized fuel boilers or at pressure (8 to 30 atmos- pheres) with air compressors and gas turbines in a com- bined-cycle power plant. The British NCB and Westing- house, under contract to OAP, have produced boiler de- signs and have estimated the performance of fluidized- bed boilers in utility power plants. Estimates indicate that the cost for power plants based on pressurized fluidized-bed boilers will be 20 to 30 percent less than for conventional plants with stack-gas scrubbing systems (for removal of SO2)• Six contractors to OAP have also demonstrated the air-pollution-control poten- tial of such devices. Fluidized-bed combustion in the past has been used for coal, oil, gas, or waste material. Currently, paper-mill wastes are burned in fluidized-bed combustors, as are oily wastes from refineries, sewage sludge, and trash. N0X formation in fluidized-bed combustion can be reduced to below that ih conventional furnaces. The equilibrium concentration of NO with 10 percent excess air at temperatures between 1,400° F. and 1,900° F. is about 30 to 70 ppm. It has been generally observed that burning coarse coal containing approximately 1.0 to 1,4 percent fixed nitrogen in fluidized beds results in 250 to 500 ppm of N0X in the off-gases. However, the Argonne Laboratory has observed that, with use of 80 percent of stoichiometric air, the N0X concentration is only 50 to 70 ppm. The NCB, under contract to OAP, has carried out fluidized-bed combustion in tests at five atmospheres in which N0X emissions have ranged from 50 to 150 ppm. It appears that the lower the SO2 content, the higher the N0X content of the flue gases, 37 ------- i.e., that some reaction occurs between limestone, S02> and N0X in a fluidized bed to help remove N0X. On the basis of taboratory and small-scale workt pressurized fluid-bed boilers show -promise of having advantages over conventional coal-fired boilers* Combustion occurs at substantially lower temperatures than in ordinary combustion processes and the N0X formed comes largely from fixed nitrogen in the fuel. N0X emissions may be somewhat reduced by additives to the bed. "Two-stage" fluid-bed combustion (reducing con- ditions in the bed, followed by secondary air injection) can reduce N0X to low levels, according to laboratory data. Pressurization also appears to reduce N0X emis- sions from fluidized-bed combustion of coal to a low level. D. OTHER COMBUSTION PROCESSES 1. Gas Turbines Stationary gas turbines and diesel engines are sources of significant NOx emissions. Some of the simplest approaches to control of emissions from diesel engines are: retard the timing to the range in which both the smoke and N0X emissions decrease; and use a larger period of injection for the same amount of fuel so that the rate of heat release in the cylinder is de- creased. Rate of injection and retarded timing tend to be more of a problem on high-swirl chambers and less of a problem on quiescent chambers. Another method is staged injection—pilot injection followed by main in- jection. These techniques have been shown to produce a fairly high reduction in N0X from existing engines, although there is a penalty in some loss in thermal efficiency. Water injected into the intake manifold or emulsified in the fuel for diesel engines has also proven successful, as has exhaust-gas recirculation and lowering of the compression ratio. On turbines the techniques of exhaust re- circulation, water injection, and combustor redesign are also applicable. A combination of fuel-injection techniques and lean operation appears to offer the most 38 ------- promise for reducing N0X from gas turbines.11!12 This involves changing the mixing patterns, adding excess air at various points along the length of the combustor, and recirculation of a portion of the exhaust gases. Various fuel additives may possibly reduce emissions, but no demonstrated technology or promising research data exist. Approximately 20 percent of the capacity of the electrical generating additions in 1970 were in the form of gas turbines for peaking power. In 1985, perhaps as much as 30 percent of the capacity of the additions may be in the form of gas turbines, both for supplying peaking power and for use in combined cycles. The com- bined cycles would be for intermediate generation, 40- to-60-percent load factor. A variety of combined-cycle plants can be visualized but typically a portion of the power is generated by a gas turbine and the remainder by a steam turbine. The outlet temperature from the combustor of a gas turbine is typically 1,600° F. and will probably be 1,800° F. in a few years. A rise to 2,000° F. or 2,200° F. can be projected as the years pass and higher efficiencies are sought and technology becomes available. An 1,800° F. outlet temperature gives approximately 170*27 ppm of N0X with 250-to-300- percent excess air, which is equivalent to 570 ppm (dry, 3 percent O2). These gas-turbine N0x emissions were measured while burning jet fuel; the amount of N0X pro- duced while burning low-Btu gas produced by coal gas- ification is unknown and needs to be determined. With gas turbines, more of the NOx produced is in the form of NO2 because of the high excess air at which they operate, and they can show a noticeable plume with about 9 or 10 ppm NO2. Estimates of N0X emissions produced by in- ternal-combustion enginesj either reciprocating or turbines9 used with pipelines and gas plants vary widely. Estimates range from 21 percent to 2 percent of the total N0X emissions from stationary sources. *Heavy-duty gas turbine at full load (19 megawatts) using jet fuel*27 39 ------- The resolution of this discrepancy requires further in- vestigation. With diesel engines> a variety of tech- niques, such as control of fuel injection, exhaust-gas recirculation, water injection, and alteration of com- bustion-chamber design, are available to reduce NOx emissions to substantially below the uncontrolled level. Exhaust-gas recirculation and water injection are also potentially applicable to stationary engines other than diesels. Catalytic mufflers developed for control of emissions from vehicles may also be applicable to sta- tionary engines. Indeed, control of emissions from stationary engines should be easier than with mobile engines. They usually run continuously at constant load, and space and weight limitations are less. With gas turbines, redesign of combustors and methods of fuel injection accompanied by more fuel-lean conditions in the combustion zone should produce significant re- duction in NCL emissions. Burning coal in a reducing atmosphere (e.g., 60 percent theoretical air) with added water or steam generates a low-Btu gas, which can be burned in a boiler or gas-turbine combustor. This process, called gasification, reduces the quantity of NOx formed in steam or power generation much as does two-stage com- bustion of natural gas. Coal and/or oil gasification provides interesting possibilities not only for re- ducing N0X but also for removing particulates and sulfur in the form of hydrogen sulfide. The volume of fuel gas, usually produced at 10-30 atmospheres pressure is much less than that of the stack gases, and thus particulate and sulfur removal are simplified. Many different gasification processes and power-gen- eration cycles have been suggested. The supercharged combined gas turbine-steam turbine and magnetohydro- dynamic (MHD) cycles have the potential of further improving the heat efficiency of the entire cycle and reducing power-plant capital costs, whereas other sys- tems for removing NQX and S0X may offer only air-pollu- tion control. Combined-cycle systems involving the combus- tion of coal or oil in a combination of steam turbines and gas turbines are of considerable interest to util- ities. Some of these may involve external coal-gasifica- 40 ------- tion processes close-coupled to a turbine or boiler. N0X emissions are expected to be low, but little infor- mation is available. Reliable commercial methods for gasification of coal to produce both a synthetic natural gas to sup- plement existing natural-gas supplies, and a gas of less than pipeline quality to be used as an electric-utility fuel and probably having other commercial, industrial, and residential uses, are of considerable interest. Fuel costs will be higher than projected gas prices or the cost of the same amount of energy from coal. How- ever, the advantages of an environmentally acceptable "clean" fuel and the reduced capital cost and increased efficiency of combined-cycle plants may outweigh the higher fuel costs for power generation. Magnetohydrodynamic generation of electricity uses high pressure and temperature combustion with seed- ing of potassium or cesium to raise the gas-ionization level. The gas is then passed through a magnetic field to generate direct-current electricity. The hot exhaust gas is sent to a boiler or turbine for additional elec- tricity generation so that the complete system is very efficient. The combustion occurs at such high tempera- tures that significant quantities of N0X are produced. Any economic evaluation of MED as a future power source should incorporate as a part of the ana- lysis the economic costs of control of the large amounts of N0X that will undoubtedly be formed. If flue-gas recovery methods for N0X were developed in conjunction with MHD, they could have a major Impact on nitric acid (HNO3) manufacture. Pollution-control costs have a significant influence on the economics of this method of energy generation. 2. Domestic Heaters Since domestic heaters are small and involve a great diversity of equipment, no single solution is universally applicable.28 Rather, control of pollution emissions will require a series of solutions tailored to the various classes of furnaces or boilers. Two 41 ------- courses of action are available for development of the required control techniques: (1) improve current com- bustion equipment for short-term abatement, and (2) de- velop new burner systems for the future. Domestic heaters are generally used in single- family homes and typically have an output of about 150,000 to 200,000 Btu/hr. These units are numerous and emit pollutants at ground levels in highly populated areas. However, the NOx-concentration levels are lower than those of large boilers. The same is true of domestic incinerators. Control of nitrogen oxides and combustible particulates from these units will require a better understanding of the critical features of the burner and combustion chamber designs and the flame structure. Changes in design of nozzles, ignition units, air-introduction devices, pressure controllers, and fuel pumps offer ways of reducing emissions. Better control of heating cycles, improved combustor design, and fuel composition may also be important. 3. Industrial Furnaces Large commercial heating units and industrial furnaces generally fall in the size range between do- mestic furnaces and utility boilers.17* Methods of control developed for these two extremes will be gen- erally applicable to industrial and commercial units. Flue-gas recirculation and "staged" combustion offer significant potential in such units. Industrial furnaces in refineries, cement and lime kilns, glass manufacturing, and metallurgical op- erations represent a special class since their primary function is to provide heating at high temperatures. It is not generally possible to apply such methods as flue-gas recirculation and staged admission of air. Therefore, combustion-modification techniques using peak temperature reduction are not possible. con- trol from these isolated single sources will require special consideration in keeping with the unique appli- cations of the units. In petroleum refineries or petro-chemical 42 ------- processing, feed or process streams are heated in a con- trolled manner and they generally cannot be subjected to conditions that would result in thermal cracking or product degradation, as might occur on dilution or tem- perature reduction. 4. Incinerators Refuse disposal by incineration results in N0X emissions that may be significant locally.29 For large units, emissions average about 25 ppm or approx- imately two to three pounds of NOx per ton of refuse burned. It appears unlikely that reduction of N0X emissions to lower levels can be achieved by process changes. A3 ------- V STACK-GAS CLEANING The observations and conclusions of the Panel presented in this section are in general agreement with a recent study by ESSO Research and Engineering that included an assessment of existing and potential N0X- control technology for stack gases.17 (a) If a flue-gas treatment process is re- quired, aqueous scrubbing systems using aqueous alka- line solutions or concentrated sulfuric acid appear to offer potential for the control of both sulfur oxides and nitrogen oxides emissions. (b) Selective catalytic reduction of N0X with ammonia, hydrogen sulfide, hydrogen, or carbon monoxide is also a possibility. (c) Silica gel, alumina, molecular sieves, chars, and ion-exchange resins all show some catalytic activity for oxidizing NO to N0£, a step that is re- quired in conjunction with any scrubber system to raise the NO2/NO ratio to at least unity. However, the capac- ities are quite low at typical NO flue-gas concentrations. (d) No flue-gas treatment process has so far been directly applied to NOx-emission control in power plants. (e) No catalyst has been found that is effec- tive for a single low-investment NO-decomposition pro- cess, and the probability of discovering such a catalyst is small. In some reports of catalytic decomposition it appears that the N0X was actually reacting with the "catalyst" or with a reducing agent rather than decom- posing as such. (f) Success in the removal of NOx from com- bustion flue gases based on differences in physical characteristics such as molecular size, condensation temperature, or magnetic susceptibility appears to be highly remote. 44 ------- Two systems that appear to offer the best possibility for removal of N0X in the exit gases are magnesium oxide scrubbing and the Tyco modified lead- chamber process.12,20»31 Catalyst systems being de- veloped for automobile exhaust control may also be applicable in specific cases. The low solubility of NO requires that NO2 be added to the flue gas to form soluble N2O3 or that about half of the NO be oxidized to NO2, in either magnesia scrubbing or the Tyco process. The use of two scrubbing systems to remove SOx and N0X in sequence is a major expense. However, the Tyco process has the potential for removing SOx and NOx simultaneously. Since NO2 must be recycled to achieve the least equal parts of SO2 and N02, the scrubber that removes N0X miist be highly efficient* For typical conditions (3,000 ppm S0X) about 90-percent scrubber efficiency in NOx removal is required just to keep the process going for S0X removal even with no net NOx removal. Ninety-nine percent efficiency is required to achieve a level of 50 ppm N0X in the final stack gas, starting with an initial concentration of about 800 ppm. Esti- mates of the cost of the process are highly sensitive to the degree of removal of N0X in the flue gas, assumed scrubber efficiencies, and the value of the sulfuric acid and nitric acid produced. No proven process is available for substantial removal of N0X from combustion stack gases. The Panel's definition of a proven process is one year of satis- factory operation on an industrial scale. Scrubber or adsorption systems proposed primarily for S0X re- moval should also be evaluated for their potential in removing Wx simultaneously. Any wet scrubber system for N0y removal will be expensive because: (a) Most of the N0X is in the form of NO, which is relatively unreactive and relatively insoluble. The maximum rate of absorption in an aqueous system occurs at a NO/NO2 molar ratio of unity, which requires 1) recycle of NO2 and a method of generating NO2 from 45 ------- the scrubbing system or 2) the oxidation of half of the NO to NO2. The rate of oxidation of NO to NO2 is slow and decreases with increasing temperature, (b) Large vessels are required for scrubbing because of the large volume of gas that must be handled and the necessity for low-pressure drop. Aqueous scrubbing systems, such as those using sulfuric acid, are In an early stage of development. It is necessary to generate NO2 and recycle it to the scrubber in order to achieve practical rates of NQx absorption, which in turn requires very high scrubbing efficiencies for substantial overall removal of N0X. Any new concepts showing potential for simul- taneous removal of NOx and S0X should be encouraged, but their economic practicality should be carefully scrutinized. It is deemed unlikely that a good absorbent can be found for removal of NO as such. However, any new ideas in this respect should be examined. Catalytic reduction of N0X to N2 by a reducing agent such as ammonia as a process for treating stack gases from large utility boilers requires a sulfur- resistant catalyst If coal or oil is used as a fuel. Space velocity (i.e., catalyst activity) and catalyst life also limit this approach at present. Available information is insufficient £0 assess the potential of this method of control. Decomposition of N0X In the absence of a re- ducing agent requires such high temperatures, on even the best catalysts known, that It Is impractical. A substantial reduction In t.he amounts of NO* released to the atmosphere (e.g.s of the order of 50 to SO percent) will come most economically from modifica- tions of the combustion process rather than from scrub- bing or adsorption systems to remove N0X from stack gases. 46 ------- VI FORMATION AND CONTROL OF EMISSIONS FROM CHEMICAL OPERATIONS Relative to emissions considered nationally, NOx from chemical operations is quite small; but locally these emissions can be significant.12*32»33 Essentially all these emissions are associated with the manufacture or use of nitric acid. About 75 percent of the nitric acid produced in the United States is consumed in ammo- nium nitrate production. The remainder is used in a variety of processes, with manufacture of adipic acid consuming 9 percent. Because of the high ratio of NO£ to NO in the stack gases from nitric acid plants, colored plumes are visible at relatively low NC^ levels, on the order of a few hundred ppm or much less, depending upon the stack diameter. The present method of control of emissions from nitric acid plants is primarily by catalytic re- duction, using natural gas. It appears that, in much of the current practice, NO2 is reduced only to NO. This is only decolorization, not emission control. Complete reduction to N2 using methane requires complete burnout of the O2 present, more consumption of natural gas, and closer control of the equipment. In addition, CO and hydrocarbon emissions increase with increasing NO reduction. Selective reduction of NO to N2, using ammonia as a fuel, has been described, but there Is insufficient information to determine whether this process is a practical alternative. Scrubbing with caustic soda has long been practiced with N0X emissions from nitration reactions, but it presents a disposal problem. Recent developments with molecular sieves3J+ indicate that adsorption pro- cesses based on their use may be capable of reducing emissions to the 10 to 50 ppm level. In a nitric acid plant, the desorbed N0„ can then be recycled to the absorption tower. Evaluation of a molecular-sieve adsorption process an a large demonstration scale is timely. 47 ------- The quantities of N0X emitted from nitric acid plants and chemical operations are but a small fraction of the total man-made emissions on a nation- wide basisj but they may comprise a significant local source of pollution. Technology for 'decolorization" (conversion of N0£ to NO) by catalytic reduction with natural gas is well established. Present methods of abatement (reduction to N2) are available to reduce NOq from typically 3,000 ppm to 100-500 ppm3 but re- quire careful control*17 Adsorption by molecular sieves or other adsorbents has been shown on a bench scale to result in even lower emission levels but no commercially proven process is yet available. Alka- line scrubbing is a proven process3 but would involve liquid waste disposal problems. In some chemical processing with nitric acid, a substantial portion of the nitrogen oxides emitted may be in the form of N2O. N2O is believed to be innocuous and should be distinguished from NO and NC^. N2O is not included with NOx-emission data. 48 ------- VII SAMPLING AND ANALYTICAL METHODS Enforcing nitrogen oxide emission standards will require monitoring, as will enforcing particulate emission standards.35 The circumstances are different from those associated with sulfur oxide emissions, in which fuel sulfur specification frequently has been satisfactory. The method recommended by EPA for 24- hour sampling of ambient NO2 is the Jacobs-Hochheiser method. 3»l+»5»5»7 Nitrogen dioxide is collected by bubbling air through a sodium hydroxide solution to form sodium nitrite. The nitrite ion is then detected colorimetrically by reacting the exposed reagent with phosphoric acid, sulfanilamide, and N-(l-Napththyl)- ethylenediamine dihydrochloride. Corrections must be made for SC^ interference. Relative standard devi- ations are reported to be between 14 and 21 percent and no accuracy data are available. Alternate tech- niques such as the Saltzraan method may differ by as much as a factor of 3. The Jacobs-Hochheiser method is for ambient-sample analysis and does not bear on the validity of source-sampling data. Nondispersive infrared and electrochemical methods under development offer the potential of rapid analysis and improved accuracy. Photo-ionization or chemiluminescent procedures may also provide general- purpose portable detectors in the future. Maintenance of emission and ambient air quality standards depends on accurate methods of sampling and analysis. Careful consideration must be given to the reliability of present methods of sampling and analysis for NO and NO2, particularly in the presence of other pollutants. Sufficient attention should be paid to those problems to ensure that decisions are being made on the basis of well-established facts. 49 ------- APPENDIX A BIBLIOGRAPHY AND REFERENCES U.S. Congress. Public Law 91-604, H.R. 17225. Clean Air Amendments of 1970. 91st Congress. 2nd Session. December 1970. . Public Law 90-148. The Clean Air Aat of 1967. 90th Congress. 1st Session. January 1969. General Services Administration. National Archives and Records Service. Office of the Federal Register. National Ambient Air Quality Stan- dards. Notice of Proposed Regulations for Preparationj Adoption9 and Submittal of Implementation Plans. Vol. 36. No. 67. Washington: Government Printing Office. April 7, 1971. . National Primary and Secondary Ambient Air Quality Standards. Vol. 36. No. 84 (Part II). Washington: Government Printing Office. April 30, 1971. . Requirements for Preparation, Adoptionj and Submittal of Implementation Plans. Vol, 36. No. 158. Washington: Government Printing Office, August 14, 1971. . Standards of Performanoe for New Stationary Sources. Vol. 36. No. 159. Washington: Government Printing Office. August 17, 1971. . Standards of Performance for New Stationary Sources. Vol. 36. No. 247. Washington: Government Printing Office. December 23, 1971. U.S. Department of Health, Education, and Welfare. Public Health Service Consumer Protection and Environmental Health Service. National Air Pollution Control Administration. Air Quality Criteria for Hydrocarbons. Washing- ton: National Air Pollution Control Adminis- tration. (NAPCA Publication No. AP-64) 1970. 50 ------- 9. . Air Quality Criteria for Photochemical Oxidants. Washington: National Air Pollution Control Administration. (NAPCA Publication No. AP-63) 1970. 10. . Control Techniques for Carbon Monoxide Emis- sions from Stationary Sources. Washington: National Air Pollution Control Administra- tion. (NAPCA Publication No. AP-65) 1970. 11. . Control Techniques for Carbon Monoxidej Nitrogen Oxides9 and Hydrocarbon Emissions from Mobile Sources. Washington: National Air Pollution Control Administration. (NAPCA Publication No. AP-66) 1970. 12. . Control Techniques for Nitrogen Oxides Emissions from Stationary Sources. Washing- ton: National Air Pollution Control Admin- istration. (NAPCA Publication No. AP-67) 1970. 13. . Nationwide Inventory of Air Pollutant Emis- sions 1968. Washington: National Air Pollu- tion Control Administration. (NAPCA Publica- tion No. AP-84) 1971. 14. Environmental Protection Agency. Air Pollution Con- trol Office. Air Quality Criteria for Ni- trogen Oxides. Washington: Air Pollution Control Office. (APCO Publication No. AP-84) 1971. 15. Robinson, E., and R. C, Robbins. Sourcest Abun- dance j and Fate of Gaseous Atmosperic Pol- lutants. SRI Project PR-6755. New York: American Petroleum Institute. February 1968, 16. National Academy of Engineering-National Academy of Sciences-National Research Council. Division of Engineering. Committee on Air Quality Management. Ad Hoc Panel on Control of Sul- fur Oxide Emissions from Stationary Sources. Abatement of Sulfur Oxide Emissions from 51 ------- Stationary Combustion Sources. Springfield, Virginia: National Technical Information Service. (PB 192887) May 1970. 17. Bartok, W., et al. Systems Study for N0X Control Methods for Stationary Sources. Final Re- port. Volumes I-III, Linden, New Jersey: ESSO Research & Engineering Company* Govern- ment Research Laboratory. Prepared under Con- tract PH-22-68-55 (GR-2-NOS-69). November 20, 1969. 18. Hall, R. E., et al. NAPCA Combustion Research Pro- grams to Control Pollutant Emissions from Domestic and Commercial Heating Systems. Cincinnati, Ohio: Environmental Protection Agency. Office of Air Programs. Division of Control Engineering. 19. Risser, E. Power and the Environment—A Potential Crises in Energy Supply. Urbana, Illinois: Illinois State Geological Survey. Environ- mental Geology Notes No. 20. December 1970. 20. Executive Office of the President. Office of Science and Technology. Energy Policy Staff. Elec- tric Power and the Environment, Washington: Government Printing Office. August 1970. 21. Goody, R. M., and C. D. Walshaw. Journal of the Royal Meteorological Society. No. 342. pp 469-500. October 1953. 22. Shy, C. M., et al. The Chattanooga School Children Study: Effects of Community Exposure of Ni- trogen Dioxide—2. MethodsA Description of Pollutant Exposurej and Results of Ventila- tory Function Testing—2. Incidence of Acute Respiratory Illness. Journal of the Air Pollution Control Association* Vol. 20. No. 8. pp 539-588. August 1970. 23. Mencher, S. K,, and H. M. Ellis. Statement Regarding the Proposed National Ambient Air Quality Standards. New York: Gordian Associates. 1971. 52 ------- 24. Zeldovich, Y. B. The Oxidation of Nitrogen in Com- bustion Explosions. Acta Physlcochimica U.S.S.R. Vol. 21. pp 577-628. 1946. 25. Bagwell,.F. A., et al. Oxides of Nitrogen Emission Reduction Program for Oil and Gas Fired Utility Boilers. Paper presented at the 62nd Annual Meeting of the American Power Con- ference. April 1971. 26. Bagwell, F. A., et al. Utility Boiler Operating Modes for Reduced Nitric Oxide Emissions. Paper presented at the 64th Annual Meeting of the Air Pollution Control Association. June 1971. 27. Fenlmore, C. D., et al. Formation and Measurements of Nitrogen Oxides in Gas Turbines. New York: ASME Winter Annual Meeting. November 29- December 3, 1970. 28. Pollution Responsibility—Domestic and Commercial Combustion. (excerpts from Walden Research Corporation. Contract CPA 22-69.85.) A Systematic Study of Air Pollution From Fossil-Fuel Combustion Equipment. 29. Bagg, J. Formation and Control of Oxides of Nitrogen. W. Strauss, ed. "Air Pollution Control, Part X." Wiley Interscience. 1971. 30. First, Melvin W., and Frederick J. Viles, Jr. Clean- ing up Stack Gases Containing High Concentra- tions of Nitrogen Oxides. Paper no. 70-70 presented at the 63rd Annual Meeting of the Air Pollution Control Association. 1970. 31. Bartok, William, and Homer J. Hall. NOx Control from Stationary Sources. Environmental Science and Technology. Vol. 5 No. 4. April 1971. 32. Haridson, L. C. Techniques for Controlling the Oxides of Nitrogen. Journal of the Air Pol- lution Control Association. Vol. 20. No. 20. pp 377-382. 1970. 53 ------- 33. U. S. Department of Health, Education, and Welfare, and Manufacturing Chemists1 Association. Atmospheric Emissions from Nitric Acid Man- ufacturing Process. Springfield, Virginia: National Technical Information Service. (PB 190239) 1966. 34. Union Carbide Corporation* Molecular Sieve Depart- ment. Union Carbide Molecular Sieve Process for N0X Removal and Recovery from Nitric Acid Plant Tail Gas. Tarrytown, New York: Union Carbide Corporation. Molecular Sieve Department. 1970. 35. Mills, John L. Continuous Monitoring. Chemical Engineering Deskbook Issue, pp 217-220. April 27, 1970. 36. Air Pollution Control and Industry's Expanding Role* Filtration Engineering, pp 7-9, November- December 1970* 37. Wilson, William E., Jr., et al. A Study of Sulfur Dioxide in Photochemical Smog. II-Effect of Sulfur Dioxide on Oxidant Formation in Photochemical Smog, Paper presented at the 63rd Annual Meeting of the Air Pollution Control Association. 1970. 38. Fabuss, Bela M., and Ilan Katz. Reduction of Nitro- gen Oxides. Paper presented at the 63rd Annual Meeting of the Air Pollution Associa- tion. 1970. 39. Wilson, William E., Jr., et al. Aerosol Formation of Photochemical Smog. I-Effect3 of Stirring. Paper presented at the 63rd Annual Meeting of the Air Pollution Control Association. 1970. 40. Roberts, J. J., and E. J. Croke. Land Use as an Organization Basis for Urban and Regional Air Resources. Paper presented at the 63rd Annual Meeting of the Air Pollution Control Association. 1970. 54 ------- 41. Shepard, Donald S, A Load Shifting Model for Air "Pollution Control in the Electric Power Industry, Journal of the Air Pollution Control Association. Vol. 20. No, 11. pp 765-761. November 1970. 42. Croke, K. G., and Edward J. Croke. The Influence of Natural Gas Availability on Air Pollu- tion Episode Control feasibility. Journal of the Air Pollution Control Association. Vol. 20. No. 10. 00 649-652. October 1970. 43. Parsons, Donald 0., and Edward J. Croke. An Eco- nomic Evaluation of Sulfur Dioxide Air Pol- lution Incident Control. Paper presented at the 63rd Annual Meeting of the Air Pollution Control Association. 1970. 44. Roberts J. J., and E. J. Croke, Argonne National Laboratory. The Chicago Air Pollution Sys- tem Model: Systems Analysis Applied to Air Resources Management. Chicago: Argonne National Laboratory. October 1969. 45. Roberts J. J., et al. Argonne National Laboratory. Chicago Air Pollution Systems Analysis Pro- gram—A Multiple-Source Urban Atmospheric Dispersion Model. Chicago: Argonne National Laboratory. 1970. 46. Cleaning up the Atmosphere. Battelle Research Out- look. Vol. 2. No, 3. Columbus, Ohio: Battelle Memorial Institute. Columbus Laboratories. 1970. 47. Howekamp, D. P. U.S. Department of Health, Educa- tion, and Welfare. Environmental Health Service. National Air Pollution Control Administration. Flame Retention—Effects of Air Pollution. Paper presented at the 9th Annual Convention, National Oil Fuel In- stitute. June 9-11, 1970. 55 ------- 46. Cleaning up the Atmosphere. Battelle Research Outlook. Vol. 2. No. 3. Columbus, Ohio: Battelle Memorial Institute. Columbus Laboratories. 1970. 47. Flame Retention—Effects of Air Pollution. Prepared by D. P. Howekamp. U.S. Department of Health, Education, and Welfare. Environmental Health Service. National Air Pollution Control Administration. Presented at the 9th Annual Convention, National Oil Fuel Institute. June 9-11, 1970. 48. Battelle Memorial Institute. Federal Research and Development Plan for Air Pollution Control by Combustion Process Modification. Prepared under Contract No. CPA 22-69-147 for the Environmental Protection Agency, Springfield, Virginia: National Technical Information Service. (PB 198066) January 11, 1971, 56 ------- |