United States Environmental Protection Agency Air and Energy Engineering Research Laboratory Research Triangle Park, NC 27711 Research and Development EPA/600/S7-86/005 May 1986 Project Summary Environmental Assessment of NH3 Injection for an Industrial Package Boiler C. Castaldini, R. DeRosier, and L R. Waterland This report discusses emission results from comprehensive flue gas sampling of a gas- and oil-fired industrial boiler equipped with Exxon's Thermal DeNO, Ammonia Injection Process for NOX reduc- tion. The objective of the tests was to evaluate criteria and noncriteria pollutant emissions in the flue gas during a baseline (uncontrolled) and a low-NO, condition with ammonia injection. The test boiler was a 7.57 kg/s (60,000 Ib/hr) watertube unit equipped with sidefire air and Ther- mal DeNOx. Comprehensive emission measurements included continuous monitoring of flue gas emissions; source assessment sampling system (SASS) testing with subsequent laboratory analysis of samples to give total flue gas organics in two boiling point ranges, specific quantitation of semivolatile organic priority pollutant species, and flue gas concentrations of 73 trace elements; EPA Method 5/17 for solid and conden- sible particulate emissions and ammonia emissions; controlled condensation system for SO2 and SO3; and N2O emis- sion sampling. Comparison of the baseline and con- trolled emission results showed that am- monia injection at a NH3/NO molar ratio of 2.52 gave a NO, reduction of 41 percent from an uncontrolled level of 234 ppm to a controlled level of 137 ppm (corrected to 3 percent 02). NH3 emissions increased from 11 ppm for the baseline to an average of 430 ppm for ammonia injection. Nitrous oxides, N2O, was reduced 68 percent from a 50 ppm baseline level to a 17 ppm con- trolled level. Total particulate emissions in- creased by an order of magnitude from a baseline of 17.7 ng/J (0.042 lb/10* Btu) to a controlled level of 182 ng/J (0.43 lb/106 Btu). This increase is in part attri- buted to formation of ammonia surfate and bisulfate from residual ammonia and SO,. Total organic emissions were at a mod- erate level and showed a relative concen- tration in the nonvolatile category (boiling point greater than 300 °C). Organic emis- sions were lower by a factor of five with ammonia injection. Emissions of carbon monoxide and trace inorganic elements were not significantly affected by am- monia injection. This Project Summary was developed by EPA's Air and Energy Engineering Research Laboratory, Research Triangle Park. NC, to announce key findings of the research project that is fully docu- mented in two separate volumes (see Project Report ordering information at back). Introduction Increasing stringency in stationary source NO, emissions regulations has created applications for advanced control technologies. Over the past 5 years, the Exxon Thermal DeNO, selective non- catalytic reduction process has been in- stalled in a number of process heaters and boilers to augment NO, reduction from combustion modification at a lower cost than selective catalytic reduction. This process reduces NO, through the gas- phase reaction with ammonia in the temperature range of 870 to 1,200 °C (1,600 to 2,200 °F). NO, reduction as high as 60 to 70 percent has been achieved in some industrial applications. The reduc- tion efficiency is affected by the NH3 feed rate relative to NO, concentrations, by the degree of flue gas thermal stratification in ------- the ammonia injection section, and by the flue gas residence time within the ap- propriate temperature window. Potential emission side-effects of the process in- clude the presence of unreacted ammonia in the flue gas and the formation of am- monium sulfates for sources fired with sulfur bearing fuels. Few data have been reported on the effects of full-scale am- monia injection on emissions of other inorganic and organic species. This test was undertaken to perform a comprehen- sive emission characterization of the boiler with and without ammonia injection to supply data on overall emission impacts of the process. This test was conducted as part of the Combustion Modification Environmental Assessment Program under which similar tests have been performed on 17 boilers, process heaters, and engines. Summary and Conclusions Source Description The boiler tested was a packaged two-drum Zurn Keystone watertube unit equipped with an economizer and a rated capacity of 7.57 kg/s (60,000 Ib/hr) of superheated steam at 2.51 MPa (350 psi) and 260°C (500 °F). The boiler was equipped with a single ammonia injection grid to mix ammonia with the combustion gases in the appropriate temperature win- dow. Steam was used as the ammonia car- rier. Hydrogen injection with the ammonia was also available to lower the effective temperature range for use at low load operation. Test Program Two tests were performed: a baseline uncontrolled run, and an ammonia injec- tion controlled run. Boiler steam flowrate was about 4.0 kg/s (32,000 Ib/hr) for both tests, corresponding to a total steam in- put of about 13 MW (45 million Btu/hr). The boiler was fired with refinery gas and residual oil. Refinery gas supplied 44 per- cent of the total heat input. Ammonia in- jection rates for the controlled test were 1.53 g/s (12.1 Ib/hr) corresponding to a NH3/NO molar ratio of 2.5. The program for emission measure- ments at the two test conditions con- formed to a modified EPA level 1 protocol. In addition, NH3 flue gas emissions were measured to calculate the amount of unreacted NH3 being emitted under the boiler and control system parameters investigated. Flue gas was measured at the stack downstream of the boiler econo- mizer where the gas temperature was about 188 °C (370 °F). Flue gas measure- ments included: • Continuous monitors for NO* CO, CO2, and O2. • Source assessment sampling system (SASS) train sampling for organic and inorganic pollutant species. • EPA Method 5 with water impingers and an EPA Method 17 backup for solid and condensible particulate mass emissions. • Controlled condensation system (CCS) for S02 and S03 • Grab sample for onsite analysis of gaseous C, to Ce hydrocarbons by gas chromatography. • EPA Method 17 with HCI impinger solutions for ammonia sampling. In addition to this detailed test program, short-term tests (varying NH3/NO molar ratio, hydrogren injection, and oil/gas fuel ratio) were performed. The objective of these short-term tests was to map the per- formance of the Thermal DeNO, Process over a wide range of system and boiler operating parameters. Measurements for these short-term tests were confined to continuous monitoring of O2 and NO«. Emission Measurements and Results Results of the short-term performance tests are summarized in Figure 1. Baseline NOX emissions averaged 235 ppm at 3 per- cent O2 for two oil/gas fuel mixtures in- vestigated. The results show that NOX reduction for a given NH3/NO ratio de- pends on the fuel mixture. The system was less effective at the oil/gas ratio of 56/44 percent than at the lower oil/gas ratio of 37/63 percent (closer to the typical opera- tion of the unit and the design basis for the NH3 injection grid installed). The ad- dition of hydrogen did not improve system performance at the lower oil/gas ratio, but resulted in significant, further NO. reduc- tion to levels below 100 ppm at the higher oil/gas ratio. With the higher oil/gas ratio, boiler convective section gas temperatures are lower at the grid location, thereby decreasing the effectiveness of NH3 alone. For both fuel mixtures, the NO, reduc- tion performance appears to peak at a NH3/NO ratio of about 2.5 with little or no additional reduction gained with further in- crease in NH3 injection rate. Table 1 summarizes criteria and other gas species emissions measured during the two comprehensive tests. NH3 injec- tion, at a rate of 2.52, resulted in a 41 per- cent N0» reduction. CO emissions showed no significant change. Indications of higher Legend Boiler heat input =~13 MW (45 million Btu/hr) Stack 02 = 2.5 to 2.8 percent O 44 percent refinery gas 56 percent residual oil 63 percent refinery gas 37 percent residual oil Hydrogen injection; numbers specify Hi/NHy molar ratio if known Test point for detailed emission measurements 1.0 1.5 2.0 2.5 3.0 NHs/NO (molar ratio, based in inlet NO) Figure 1. Thermal De /V0« performance on the packaged industrial boiler. ------- Table 1. Criteria and Other Gas Species Emissions Pollutant As measured by: Continuous gas analyses O2, percent CO 2, percent NOX, ppm f\f\ _ _ __ CO, ppm Wet chemical methods SO2, ppm SO3, ppm NH-i, ppm Offsite gas chroma tography N2O Onsite gas chroma tography Cj, ppm Co, ppm C3, ppm €4, ppm Cg, ppm C6, ppm Corrected emissions ppm0 /VOX« 254 CO 30 SO2 NA SO3f NA NHji 11 N20 52 c, C2 C3 2.5 C4 cs C6 Total C1 to Ce 2.5 Paniculate mass Emissions: Method 5/17 solid - Method 5/17 - con dGnsible inorgdnic Method 5/17 - condensible organic Method 5/17 total — SASS solid - aNH3 emissions ranged from 3 to bNH3 emissions ranged from 280 °Dry ppm at 3 percent O2 dOn heat input basis eAs NO2 'As H2SO4 sArithmetic average NA — Sample lost in transit ND — Not detected Test 1 (baseline! 2.6 11.7 239 Of J/ NA NA 11a 53 ND ND 2.6 ND ND ND ng/jo 115 9.0 NA NA 2.0 25 1.3 — _ 1.3 4.6 10.0 2.9 17.7 2.2 25 ppm from Test 2 (NH3 injection) 2.5 11.6 141 Oyf -i4 82 13 44Ob 77 0.8 0.6 ND 6.2 5.1 5.0 lb/106 Btuf ppm ng/J lb/106 Btu 0.268 137 67 0.16 0.02 23 6.9 0.02 NA 80 55 0.13 NA 13 13 0.03 0.005 430 78 0.18 0.056 17 8.0 0.019 0.8 0.15 0.0003 0.6 0.21 0.0005 0.003 - 6.0 3.2 0.007 5.0 4.3 0.010 4.9 5.0 0.012 0.003 17.0 13.0 0.030 0.01 - 1.8 0.004 0.023 - 180 0.42 0.007 - 0.2 0.0004 0.042 - 182 0.43 0.005 - 2.6 0.006 three separate flue gas measurements to 600 ppm from two separate flue gas measurements gaseous hydrocarbons, especially in the C4 to C6 range, were recorded with ammonia injection. This may have been due to burner tip coking which required frequent cleaning. Baseline NH3 emissions ranged from 3 to 25 ppm, averaging 11 ppm (0.23 Ib/hr). During NH3 injection, unreacted NH3 emis- sions from two consecutive measure- ments ranged between 200 and 600 ppm. averaging 430 ppm (8.4 Ib/hr). A third measurement resulted in NH3 concentra- tions of 840 ppm (16.0 Ib/hr). This measurement was considered erroneous because it resulted in more NH3 emitted than actually injected through the grid. Analyses of the EPA Method 5 and 17 im- pinger solutions indicated concentrations of NH3 corresponding to a stack concen- tration of about 6 ppm (0.12 Ib/hr) for baseline and 360 ppm (7 Ib/hr) for the NH3 injection test. These results are in general agreement with the ammonia emission sampling system. Nitrous oxide (N2O) averaged about 50 ppm during baseline, and dropped to 17 ppm during the second test. This 68 per- cent reduction in N2O exceeds the 41 percent reduction in NOX. Total paniculate matter during the NH3 injection test increased by more than one order of magnitude. The largest contribu- tion to this increase was from the in- organic condensate matter collected in the impinger section. This can be in part ex- plained by ammonium sulfate and bisulfate formed either in the stack or through the particulate sampling system. SASS samples were analyzed for or- ganic content and inorganic trace elements. Total chromatographable organ- ics, hydrocarbons in the boiling range of 100° to 300 °C (210 to 570 °F), measured 0.023 ng/J (90 ^g/dscm) for the baseline and 0.01 ng/J (40 /jg/dscm) for the NH3 injection test. Organics measured by gravimetry (GRAV) analysis for hydrocar- bons having boiling points greater than 300 °C (>570°F) were 0.29 ng/J (1,300 ng/dscm) for the baseline and 0.059 ng/J (240 ug/dscm) for the NH3 injection test. Infrared spectra of the gravimetric residue suggest the presence of aliphatic hydro- carbons and alcohols for both tests. The XAD-2 extract of the baseline test. which contained the highest organic con- tent, was also subjected to liquid chromatography separation. There were no discernable peaks from fractions 2 through 4. These spectra indicate that, of the 1.2 mg/dscm of organic matter in the total sample, about 70 percent is aliphatic hydrocarbons, 20 percent is alcohols, and 10 percent is carboxylic acids. 3 ------- Gas chromatography/mass spec- trometry analysis of sample extracts was performed to determine the presence and concentration of 58 semivolatile organic priority pollutants. Of these, the only ones detected were naphthalene, phenanthrene, and phenol in amounts corresponding to flue gas concentrations generally <1 Kj/dscm. Results of spark source mass spec- trometry (SSMS) and atomic absorption spectrometry (AAS) indicated that in- organic trace elements were not affected by NH3 injection. Major elements having flue-gas concentrations exceeding 50 mg/dscm for both tests included: sulfur, copper, nickel, silicon, titanium, vanadium, zinc, potassium, cobalt, fluorine, and iron. These emissions are most likely the result of: C. Castaldmi, R. DeRosier, andL. R Water/and are with Acurex Corp , Mountain View. CA 94039. Joseph A. McSorley is the EPA Project Officer (see below) The complete report consists of two volumes, entitled "Environmental Assess- ment of Nf-fa Injection for an Industrial Package Boiler." "Volume I. Technical Results." (Order No. PB 86-159 852/AS; Cost: $16.95) "Volume.ll. Data Supplement." (Order No. PB 86-182 409/AS; Cost: $22.95) The above documents will be available only from: (cost subject to change) National Technical Information Service 5285 Port Royal Road Springfield. VA22161 Telephone. 703-487-4650 The EPA Project Officer can be contacted at: Air and Energy Engineering Research Laboratory U.S. Environmental Protection Agency Research Triangle Park. NC 27711 • Inorganic elements in the fuel oil. • Erosion of metal surfaces by the hot combustion gases in the boiler passes, including the NH3 injection grid. • Erosion of sampling equipment metal parts. Summary Emissions from an industrial boiler fir- ing refinery gas and residual oil were tested with and without the Thermal DeNOx ammonia injection process. N0« reductions of 30 to 60 percent were ob- served depending on the ammonia injec- tion rate, the relative amount of gas and oil fired, and the use of hydrogen injection with the ammonia. The primary emission effect on other species was the formation of ammonium sulfate and the discharge of unreacted ammonia. United States Environmental Protection Agency Center for Environmental Research Information Cincinnati OH 45268 Official Business Penalty for Private Use $300 EPA/600/S7-86/005 0000329 PS U S ENVIR PROTECTION AGENCY CHICAGO ------- |