F, and Fine Particles from

Sep 30, 2009 - Another source is the boiler where the formation reaction takes place at the surface of fly ash particles which contain soot (11, 12). ...
7 downloads 11 Views 623KB Size
Download PDF
Environ. Sci. Technol. 2009, 43, 8308–8314

Simultaneous Removal of Mercury, PCDD/F, and Fine Particles from Flue Gas J E N S K O R E L L , * ,† H A N N S - R . P A U R , † HELMUT SEIFERT,† AND SVEN ANDERSSON‡ Forschungszentrum Karlsruhe, Institute for Technical Chemistry, Thermal Waste Treatment Division (ITC-TAB), P.O. Box 3640, D-76021 Karlsruhe, Germany and Go¨taverken Miljo¨ AB, P.O. Box 8876, SE-402 72 Go¨teborg, Sweden

Received May 8, 2009. Revised manuscript received September 17, 2009. Accepted September 18, 2009.

A multifunctional scrubber (MFS) has been developed to reduce the complexity of flue gas cleaning plants. The MFS integrates an oxidizing scrubber equipped with a dioxin-absorbing tower packing material and a space charge electrostatic precipitator. All these processes have been previously developed at Forschungszentrum Karlsruhe. In the described multifunctional scrubber, mercury, sulfur dioxide, hydrogen chloride, polychlorinated dibenzo-p-dioxins (PCDD), polychlorinated dibenzofurans (PCDF), and submicrometer particles are removed simultaneously. A MFS pilot plant with a flue gas volume flow of 250 m3/h has been installed in a slipstream of a waste incineration pilot plant. Pilot scale testing was performed to measure mercury, particles, and PCDD/F in the raw and clean gas. After optimization of the process these three flue gas components were separated from the flue gas in the range 87-97%.

1. Introduction The flue gas cleaning in waste incineration plants is a decisive cost factor, which shall be reduced by integration of flue gas cleaning stages. In flue gas cleaning systems of waste incineration plants, several process steps are needed to keep to the emission limits. Separation processes like dry particle separation, gas absorption (wet or dry), adsorption, and catalytic or thermal conversion are used in various combinations. With the different separation processes, a high pollutant removal efficiency can be achieved with corresponding expenditure of capital and energy (1). To lower the costs of flue gas cleaning plants, more simple and less costly processes with simultaneous removal of several pollutants in one process step are used (1, 2). An example for the integration of several cleaning steps into one apparatus is the NID (novel integrated desulphurization) process. It consists of an upward directed entrained flow reactor with lime and activated carbon injection and a fabric filter. This process is able to capture SO2, HCl, HF, HBr, heavy metals, dioxins, and fine particles. A quite similar process is the socalled Circoclean process which uses a Venturi shaped absorber with lime and activated carbon injection operated as a circulating fluidized bed (CFB). A fabric filter is installed * Corresponding author phone: +49 7247 82 3832; fax: +49 7247 82 4303; e-mail: [email protected]. † Forschungszentrum Karlsruhe. ‡ Go¨taverken Miljo¨ AB. 8308

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 21, 2009

downstream of the CFB absorber. Another process with the same principle is the Turbosorp process (2). Mercury is one of the most hazardous environmental substances that can cause severe health problems. It is emitted in gaseous form by different thermal processes and spread on a global scale by atmospheric transport. The worldwide mercury flows are documented in the United Nations Environment Programme “Global Mercury Assessment” (3) and the “Position Paper on Mercury” of the European Commission (4). According to Pacyna (5) twothirds of the mercury emissions from anthropogenic sources in the year 2000 occurred from the combustion of fossil fuels, especially in power plants and residential boilers. The contribution of waste incineration to the global anthropogenic mercury emissions is relatively small (3%). The mercury chemistry in waste incineration flue gases has been investigated by several authors (6-8). Mercury can be removed from flue gases by adsorption on activated carbon (9) or by absorption in wet scrubbers (6, 7, 10) Dioxins are formed during the combustion of chlorinecontaining fuels in municipal solid waste incinerators, coal fired power plants, and other thermal processes. Depending on the fuel composition and the burnout of the fuel bed high dioxin concentrations can be formed especially at disturbed combustion conditions. Another source is the boiler where the formation reaction takes place at the surface of fly ash particles which contain soot (11, 12). Fine particles with a diameter below 10 µm (PM10) are emitted from combustion and industrial production processes. Stack measurements at combustion plants for waste and biomass show high number concentrations of particles around 100 nm. As the PM10-fraction is associated with adverse health effects (13, 14), even more stringent emission limits may be necessary in the future. This paper describes the reduction of the number of flue gas cleaning devices in waste incineration plants by integration of three different processes into a multifunctional scrubber. MercOx process. MercOx (Mercury Oxidation) is a wet scrubbing process which applies an acidic mixture of hydrogen peroxide and an additive as a scrubbing liquid by which elemental mercury is oxidized into a water-soluble form (15). The additive supports the oxidation and prevents the absorbed oxidized mercury from being reemitted from the scrubbing liquid. SO2 is absorbed in the scrubber and oxidized to sulfuric acid. HCl is also retained in the scrubber. The wastewater of the scrubber contains oxidized mercury, sulfuric acid, HCl, and particulates. After neutralization the mercury can be precipitated in stable sulfide form, which is suitable for final storage. The amount of hydrogen peroxide dosed into the scrubber depends on the SO2 inlet concentration and the additive concentration is kept on a constant level. ADIOX Process. The ADIOX process applies a special tower packing material to separate PCDD/F from the flue gas (16). Tower packing material made of pure polypropylene will lead to an absorption-desorption equilibrium of PCDD/F during operation. This results in increased PCDD/F clean gas concentrations after operational disturbances or start up phases of the incineration plant (memory effect). The ADIOX material contains carbon particles which adsorb the PCDD/F and thus prevent their desorption. After use (normally several years) this tower packing material can be disposed of by incineration resulting in complete thermal decomposition of the PCDD/F. 10.1021/es901289g CCC: $40.75

 2009 American Chemical Society

Published on Web 09/30/2009

FIGURE 1. Flowchart of the multifunctional scrubber pilot plant. In the absorber stage (MercOx) hydrogen peroxide and an additive are added to oxidize and stabilize mercury. Hg, SO2, and HCl are absorbed by the scrubbing liquid, PCDD/F are absorbed by the carbon containing tower packing material (ADIOX). In the aerosol collector stage (CAROLA) particles are charged and precipitated. CAROLA Process. The CAROLA process (Corona induced aerosol removal) is a wet electrostatic precipitator for fine particles based on charging of particles and subsequent deposition on a collector (17, 18). The flue gas flows through an ionization stage, which consists of a nozzle plate and star-shaped electrodes, operated at a voltage up to 20 kV. The particles are charged by a corona discharge and are precipitated in the downstream collector, which consists of a grounded tower packing stage. Due to the intensive charging process the particle collection is achieved by means of space charge of the particles without an external electrical field. The particles can be removed by flushing with water.

2. Experimental Section All flue gas volumes are in m3 which refers to standard conditions (0 °C, 0.1013 MPa, dry base). Pilot Plant. A pilot plant of the multifunctional scrubber (MFS) is installed in the bypass of the waste incineration pilot plant THERESA at Forschungszentrum Karlsruhe. The MFS pilot plant consists of the components quench, absorber and aerosol removal stage (CAROLA process) (Figure 1). The flue gas (flow rate 250 m3/h) enters the quench with a temperature of 180 °C. The head of the quench unit is made from graphite followed by a polyvinylidene fluoride (PVDF) tube with an inner diameter of 150 mm which is heat resistant up to 140 °C. The quench sump, the absorber and the aerosol removal stage are made from polypropylene (PP) which is heat resistant up to 90 °C. The sumps of quench and absorber are connected. In the quench the flue gas is cooled to about 80 °C by the injection of scrubbing liquid. In the following mercury absorber stage also HCl, SO2, and PCDD/F are removed from the flue gas. Downstream of the absorber fine particles are precipitated by the aerosol removal stage (CAROLA unit). The absorber column consists of a PP tube with an inner diameter of 338 mm, the aerosol removal

module is a PP tube with a inner diameter of 475 mm. The absorber stage and the collector part of the aerosol removal module are packed columns filled with Highflow 15-7 (Rauschert Verfahrenstechnik). The packing heights are 1 m in the absorber and 0.8 m in the collector part of the aerosol removal stage. The circulation pump of the absorber stage delivers 1200 L/h of scrubbing liquid, resulting in a liquid/ gas ratio of 5 L/m3. The residence time of the flue gas in the quench stage is 0.3 s, in the absorber stage it is 1.0 s. The packed column of the aerosol removal stage is operated with water to improve the removal of HCl, HgCl2, and H2O2. The residence time in this stage is 2.0 s. The concentration of hydrogen peroxide in the scrubbing liquid was set to 1 g/L. During the measurement campaign 35 samples were taken from the absorber circuit and the hydrogen peroxide concentration was measured by titration. The pH value was set below 0.5 by hydrochloric and sulfuric acid, which is similar to an industrial MercOx scrubber in Sweden. This is maintained by the oxidation and absorption of SO2 to H2SO4 by H2O2. The additive concentration was in the range of 1.4-2.2 g/L. This composition of the scrubbing liquid resulted in a redox potential in the range between 580 and 837 mV. Since the fuels of the pilot waste incinerator that delivered the flue gas for the MFS pilot plant only contain trace amounts of mercury the raw gas of the MFS is spiked with mercury by the injection of HgCl2 solution through a heated tube made of stainless steel. The spiking is operated continuously. The tube of the mercury generator is heated to a temperature of 600 °C, is 700 mm in length and 60 mm in diameter. The solution (HgCl2 concentration: 51 mg/L) is injected into the tube via a spray nozzle with a flow rate of 1 L/h. The produced gas is introduced into the flue gas duct by heated pressurized air. The injection of the solution at 600 °C leads to a partial decomposition of the HgCl2, which provides a mixture of VOL. 43, NO. 21, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8309

TABLE 1. Tests Conducted during the Measurement Campaign at the MFS Pilot Plant Hg speciation raw/ Particle concentration raw/ PCDD/F concentration raw/ clean gas clean gas clean gas test no.

fuel in pilot waste incinerator

1 2 3 4 5 6 7 8 9

fuel oil fuel oil fuel oil + wood chips fuel oil + wood chips fuel oil + wood chips fuel oil + wood chips fuel oil + PVC fuel oil + PVC fuel oil

objective Hg Hg Hg Hg Hg Hg Hg Hg Hg

+ + + + + + + + +

number of samplesa

number of samplesa

number of samplesa

2/2 3/3 3/3 3/3 2/2 3/3 3/3 3/3 2/2

1/1 4/4 3/4 4/4 3/3 3/3 1/2 -/2 2/2

-/-/-/-/-/-/1/1 1/1 -/-

particle removal particle removal particle removal particle removal particle removal particle removal PCDD/F removal PCDD/F removalb particle removal

a Some samples were destroyed by dropping on the floor or other incidents. For this reason the numbers of raw and clean gas samples are not identical in some tests. b Fresh scrubbing liquid.

elemental and oxidized mercury in the raw gas like it occurs in industrial waste incineration plants. The target Hg feed concentration in the raw gas was 150 µg/m3. Measurement Techniques. Two continuous emission monitors (Hg-CEM, Seefelder Messtechnik GmbH) are installed in the raw and clean gas of the MFS. The probes are not heated and separate out the particulates. The lines to the CEMs are heated to 200 °C. The CEMs are calibrated once a day using a built in Hg(0) generator. In these instruments oxidized mercury is reduced to elemental mercury in a thermo catalytic reactor. The flue gas condensate is removed by a cooler and elemental mercury is collected by a gold trap. The gold trap is heated under N2 and the desorbed mercury is analyzed using an atomic absorption spectrometer at 254 nm. The positions of the raw and clean gas samplings are shown in Figure 1. A heated probe is used to extract the flue gas from the duct. After a glass wool filter the flue gas passes the two mercury adsorption tubes which are heated to 120 °C. A dry speciation method is used to determine elemental and oxidized mercury (HgCl2) in the raw and clean gas (7, 19). It consists of two tubes containing selective adsorbents. The material used for adsorption of oxidized mercury is the ionexchange resin Dowex 1 × 8 which is doped with chloride ions by a pretreatment with hydrochloric acid. Anasorb C300, a mixture of metal oxides is used for the adsorption of elemental mercury. Test tubes containing Anasorb C300 are commercially available (SKC Inc.). The gas is dried by silica gel and the gas volume is corrected for standard conditions by a gas sampling system (GS 312, DESAGA GmbH). This sampling system contains in a compact unit components to produce the gas flow, to measure and control the flow rate and the volume, and to condition the gas. The sample gas flow rate is adjusted to 2 dry norm liters per minute, the sampling time is 2 h for each measurement. After sampling the mercury is extracted from the Anasorb C300 adsorbent by HNO3 and HCl. During extraction Anasorb C300 is nearly completely dissolved as described below. The Dowex adsorbent is extracted by a 4-fold extraction with HNO3. The mercury concentration in the resulting solutions is measured by a flow injection mercury system (FIMS, Perkin-Elmer). A number of quality control measures are applied to secure the results of the dry mercury speciation method. The Dowex resin pretreatment and the tube preparation are documented in detail. Each Dowex tube is stored in a plastic packing at 4 °C after preparation. Recovery tests with different mercury loads were conducted for both adsorption materials which show 83% ( 5.2 (n ) 17) recovery for elemental mercury and 96% ( 1.3 (n ) 18) recovery for oxidized mercury. These factors are applied for the correction of the raw measurement data. The recovery tests with Dowex were performed by dosing a certain volume of a HgCl2 solution directly onto the 8310

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 21, 2009

material. The Dowex resin was extracted three times with HNO3 (65%) followed by two times flushing with H2O (bidest.). The Hg concentration in the solution was determined by a flow injection mercury system (FIMS, Perkin-Elmer). Four different Hg concentrations were measured five times each. The recovery tests with Anasorb C300 were performed by sucking a Hg containing synthetic air flow through the sorbent tubes. The Hg containing gas flow was produced by dosing SnCl2 into a HgCl2 solution and stripping the resulting Hg(0). The Anasorb material was digested by HCl and HNO3 followed by determination of the Hg concentration in the solution by FIMS. Four different Hg concentrations were measured five times each. The particle mass concentrations in the raw and clean gas of the MFS were measured according to VDI method 2066 using glass fiber filters. The PCDD/F concentrations were determined according to DIN EN 1948 and subsequent HRGC-MS-analysis of the PCDD/F congeners.

3. Results and Discussion In a measurement campaign (Table 1) the concentrations of mercury, particles, and PCDD/F in the raw and clean gas of the MFS pilot plant were determined on nine consecutive days. The aim was to determine the removal efficiencies for mercury, PCDD/F and particulate matter during the combustion of different fuels in the waste incineration plant. Mercury Removal. In the following paragraph the mercury removal efficiency of the multifunctional scrubber is discussed for the different conditions in the waste incineration plant. Figure 2 shows continuous measurements of the raw and clean gas mercury concentrations at the MFS pilot plant by Hg-CEM. Particulate-bound Hg is not included in the diagram. The Hg tests occurred on 9 days, whereas the HgCl2 spiking in the raw gas was on, i.e., for about 8 h each day. The diagram only shows the raw and clean gas measurements with the Hg spiking on. The times with spiking off (zero Hg in the raw gas) are not included in the diagram. The fluctuations are due to the adsorption and desorption of the injected HgCl2 in the heated steel tube caused by temperature changes and gas composition. The results of the mercury speciation measurements in the raw gas are summarized in Figure 3. The purpose of the adsorption tube samples was to provide information about the Hg speciation, not to compare it to the CEM data. Due to the limited space for the pilot plant the location of the sorbent trap sampling was 55 cm downstream of the HgCl2 solution injection at the opposite side of the flue gas duct. The continuous Hg monitor was connected 65 cm downstream of the HgCl2 solution injection at the same side of the flue gas duct. During tests nos. 1, 2, and 9 (fuel: oil) the share

FIGURE 2. Continuous mercury measurement in the raw and clean gas of the MFS pilot plant measured by two Hg-CEM systems.

FIGURE 3. Mercury species in the MFS raw and clean gas measured by sorbent traps with selective adsorption materials. Depending on the raw gas levels of HCl, SO2, and NOx the shares of oxidized and elemental mercury were different in each sample. The raw gas concentrations of HCl, SO2 and NOx are also shown in the diagram. of oxidized mercury in the raw gas was higher than the one of elemental mercury. The average ratio of Hg(0)/Hg(II) for these experiments is 0.65. Elemental mercury prevailed in the raw gas of tests nos. 3-6 (fuel: oil+wood chips). Here the average ratio of Hg(0)/Hg(II) is 8.12. During the combustion of oil + wood chips the SO2/HCl ratio in the raw gas was 12.77, during the combustion of oil it was 11.29. The higher share of SO2 may support the formation of Hg(0) in the raw gas. The HCl concentration in the raw gas is an indicator for the formation of HgCl2. During the experiments with combustion of oil and oil+wood chips the average HCl raw gas concentration was low (14 and 13 mg/m3h). During tests nos. 7 and 8 (fuel: oil+PVC) the average HCl raw gas concentration was 608 mg/m3 and oxidized mercury was the dominant Hg species. The average ratio of Hg(0)/Hg(II) for these experiments is 0.28. The combustion of PVC results in the formation of HCl which leads to the high share of oxidized mercury. The raw gas concentrations of HCl, SO2, and NOx are also shown in Figure 3. HBr data are not available.

During the experiments with combustion of oil (tests nos. 1, 2, and 9) Hg clean gas concentrations in the range between 6 and 30 µg/m3 have been determined. The average mercury removal efficiency was 76% ( 15%. For tests nos. 3-6 (fuel: oil + wood chips) Hg clean gas concentrations between 5 and 17 µg/m3 have been measured. The lower clean gas concentrations resulted in a higher average mercury removal efficiency of 81% ( 9%. During the combustion of oil and PVC (tests nos. 7 + 8) Hg clean gas concentrations in the range of 11 and 19 µg/m3 have been determined. Since the raw gas concentrations have also been quite low (18-51 µg/m3) the resulting average removal efficiency was only 38% ( 27%. The HCl produced by the combustion of PVC results in a high share of HgCl2 in the raw gas. HgCl2 tends to adsorb at surfaces in the flue gas path at low temperatures. Since the temperature is only 180 °C in the raw gas upstream of the MFS the HgCl2 is likely to be adsorbed in the flue gas line before it reaches the raw gas sampling point. Tests nos. 1 and 2 also showed a high proportion VOL. 43, NO. 21, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8311

FIGURE 4. Particle mass concentrations in the MFS raw gas (black columns) and in the MFS clean gas (white columns). of oxidized mercury but not as high as during PVC combustion (Tests nos. 7 and 8). Particle Removal. In this paragraph the particle removal efficiency of the multifunctional scrubber is discussed for the different conditions in the waste incineration plant. The particle removal of the multifunctional scrubber was determined by filter samples with and without application of high voltage at the ionizer. The particle size distribution was not analyzed. Without high voltage means that the corona discharge is off. The raw gas samples were taken upstream of the quench unit, the clean gas samples were taken downstream of the CAROLA unit (Figure 1). Therefore the particle removal efficiency is composed of the particle removal units quench, absorber and CAROLA unit. Coarse salt particles are separated by forces of inertia, and then the salt is dissolved in the MFS quench and absorber liquids before entering the CAROLA unit. In the quench and absorber, only the coarse particles are removed as these units have a low efficiency for submicrometer particles. The CAROLA unit is mainly responsible for the collection of fine particles which represent only a small share of the total particle mass. This agrees with the observation that about 92% of the particles are removed even when the corona discharge was not in operation. Figure 4 shows the results of the filter samples in the MFS raw and clean gas. The particle size distribution was not determined. During the combustion of fuel oil (tests nos. 1, 2, and 9) raw gas particle mass concentrations in the range between 240 and 540 mg/m3 have been measured. In tests nos. 3-6 (fuel: oil and wood chips) raw gas particle mass concentrations of 154-512 mg/m3 have been determined. The container of raw gas filter no. 8 has been destroyed by incident and thus this filter was not analyzed. The particle load in the raw gas is only partly due the combustion of fuels. Typical particle concentrations downstream of the boiler of the waste incineration pilot plant are 20-60 mg/m3. The main amount of particles (200-500 mg/m3) is generated by the spray dryer downstream of the boiler (salt particles). Tests nos. 1 and 2 were conducted without high voltage at the ionizer (corona discharge off). Clean gas particle mass concentrations between 17 and 61 mg/m3 have been determined resulting in an average removal efficiency of 94% ( 3%. Test no. 9 was conducted with high voltage at the ionizer (corona discharge on). This resulted in clean gas 8312

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 21, 2009

concentrations of 39 and 1 mg/m3 and a mean removal efficiency of 92% ( 11%. Tests nos. 3-6 were conducted with (samples 6-9, 14-16) and without (samples 10-13, 17-19) high voltage at the ionizer. Seven samples have been collected during the high voltage operation. Clean gas particle mass concentrations between 3 and 16 mg/m3 have been determined. The corresponding average removal efficiency was 97% ( 1%. Another seven samples have been collected during operation without high voltage (corona discharge off). Here clean gas particle mass concentrations of 11-59 mg/m3 have been measured resulting in an average removal efficiency of 91% ( 9%. Coarse salt particles are separated in the MFS quench and absorber units before entering the precipitator. This agrees with the observation that about 92% of the particles are removed even when the corona discharge was not in operation. Downstream of the absorber the fine particle fraction is charged and precipitated in the aerosol collector module. This pilot plant installation had no ideal raw gas properties to test the CAROLA process. But the aim of the measurement campaign was to test the complete MFS with all three integrated processes. PCDD/F Removal. Gaseous PCDD/F measurements were conducted during start up and cool down phases of the pilot waste incinerator which are characterized by relatively high PCDD/F concentrations in the raw gas. The two tower packing stages in the MFS pilot plant in the bypass of the pilot waste incinerator had volumes of 0.09 (absorber unit) and 0.14 m3 (CAROLA unit) and were operated at 70 °C. The PCDD/F concentrations at the inlet and outlet of the MFS were determined in two experiments. In each experiment two 4 h raw gas samples were taken simultaneously to one 8 h clean gas sample. The raw gas sampling point was downstream of the pilot incinerators’ boiler (300 °C), the clean gas was sampled downstream of the MFS (70 °C). By the sampling the TEQ value in the flue gas was determined. TEQ means toxicity equivalent. To get the TEQ value the concentrations of the determined dioxin and furane species in the flue gas have to be multiplied with the corresponding equivalent factors followed by adding up the resulting products. The dioxin species with the highest toxicity (2,3,7,8-TCDD) has the equivalent factor 1. At raw gas concentrations of 1.38 and 0.35 ng TEQ/m3, a removal efficiency of 97% was measured.

FIGURE 5. Raw (black columns) and clean (white columns) gas total mercury concentrations at the industrial MercOx/ADIOX scrubber measured by sorbent traps during a test campaign in 2005. Each sample shows the average Hg concentration over a time period of four hours. At a hazardous waste incinerator in Sweden a PCDD/F removal efficiency of 86% was measured (see the Industrial application of the MFS section). In another full scale installation at a Danish waste incinerator the measurements have shown that the PCDD/F concentrations in the clean gas are far below the emission limit of 0.1 ng TEQ/ m3 (20). The high PCDD/F removal efficiency in the MFS pilot plant is presumably due to the long residence time especially in the second tower packing stage (0.14 m3) of the aerosol collector module. Industrial Application of the MFS. Since the year 2000 a MercOx scrubber has been installed in a Swedish waste incineration plant to improve the mercury removal efficiency. The ADIOX tower packing material has been used in this scrubber since 2004 to minimize the PCDD/F memory effect during operational disturbances and start ups of the incineration plant. The incineration plant consists of a rotary kiln line for hazardous waste and a grate furnace line for household waste. The flue gases of both incineration lines first each pass a semidry scrubber with lime milk injection and a bag house filter with activated carbon injection. As the last step the flue gases of both lines (120 000 m3/h) are cleaned in the wet scrubber. In a treatment plant the wastewater of the scrubber (700 L/h) is neutralized and the mercury is precipitated by sulfides and separated by a filter press yielding a precipitate which is suitable for final storage. The cleaned wastewater is recycled and injected into one of the semidry scrubbers. Three measurement campaigns have been conducted at the industrial scrubber in 2004 and 2005 (21). Figure 5 shows mercury measurements conducted with the solid adsorption method described in the Measurement Techniques section. At inlet concentrations between 50 and 150 µg/m3 (Tests nos. 10-13) the outlet concentration is well below 10 µg/m3. When the raw gas concentration is very high (Test no. 15) the mercury concentration in the scrubbing liquid is increasing. This leads to a moderate increase of the clean gas concentration as shown in the following test no. 16. The PCDD/F emissions of the industrial scrubber have been determined by adsorption on XAD resin over a period of two years. According to the measurements 160 mg TEQ have been emitted in 2003 and only 22 mg TEQ in 2004 with the new tower packing material in place. This results in a PCDD/F removal efficiency of 86% by application of ADIOX in the scrubber.

According to preliminary economic estimations flue gas cleaning plants with a multifunctional scrubber seem to have lower investment costs than state-of-the-art systems. Details of the economic evaluation are provided in the Supporting Information. This comparison only refers to the plant technology. To improve the accuracy of these preliminary figures it will be necessary to study existing waste incineration plants with definite raw gas concentrations and boundary conditions.

Acknowledgments We acknowledge the work by Klaus Rieber and Marco Mackert for collecting the experimental data and by Dr. Klaus Jay for the HRGC-MS analysis of PCDD/F congeners.

Supporting Information Available Additional figures showing the waste incineration pilot plant which delivered the flue gas for the MFS pilot plant, the principles of MercOx, ADIOX and CAROLA processes, and an economic evaluation of the MFS. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Schaub, G. Flue gas cleaning with pollutant removal and resource recovery in municipal waste incineration - state-of-the-art and trends. Chem. Ing. Tech. 1996, 68, 1424–1431. (2) Werther, J. Gaseous emissions from waste combustion. In Chemical Industry and Environment V; Ho¨flinger, W., Ed.; Vienna University of Technology: Vienna, 2006, Vol. I. (3) UNEP (2002): Global mercury assessment: http://www.chem. unep.ch/mercury/Report/GMA-report-TOC.htm. (4) EU-Commission (2002): Position paper on mercury: http://europa. eu.int/comm/environment/air/pdf/pp_mercury.pdf#page)5. (5) Pacyna, E. G.; Pacyna, J. M.; Steenhuisen, F.; Wilson, S. Global anthropogenic mercury emission inventory for 2000. Atmos. Environ. 2006, 40, 4048–4063. (6) Vogg, H.; Braun, H.; Metzger, M.; Schneider, J. The specific role of cadmium and mercury in municipal solid waste incineration. Waste Manage. Res. 1986, 4, 65–74. (7) Metzger, M.; Braun, H. In-situ mercury speciation in flue gas by liquid and solid sorption systems. Chemosphere 1987, 821– 832. (8) Hall, B.; Lindqvist, O.; Ljungstro¨m, E. Mercury chemistry in simulated flue gases related to waste incineration conditions. Environ. Sci. Technol. 1990, 24, 108–111. (9) Liu, W.; Vidic, R. D.; Brown, T. D. Impact of flue gas conditions on mercury uptake by sulfur-impregnated activated carbon. Environ. Sci. Technol. 2000, 34, 154–159. VOL. 43, NO. 21, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8313

(10) Van de Velden, M.; Dewil, R.; Baeyens, J.; Josson, L.; Lanssens, P. The distribution of heavy metals during fluidized bed combustion of sludge (FBSC). J. Hazard. Mater. 2008, 151, 96– 102. (11) Everaert, K.; Baeyens, J. The formation and emission of dioxins in large scale thermal processes. Chemosphere 2002, 46, 439– 448. (12) Hunsinger, H.; Jay, K.; Vehlow, J. Formation and destruction of PCDD/F inside a grate furnace. Chemosphere 2002, 46, 1263– 1272. (13) Pope, C. A.; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Keweski, D.; Ito, K.; Thurston, G. D. Lung cancer, cardiopulmonary mortality and long-term exposure to fine particulate air pollution. J. Am. Med. Assoc. 2002, 287, 1132–1141. (14) Mu ¨ lhopt, S.; Paur, H.-R.; Diabate´, S.; Krug, H. F. In vitro testing of inhalable fly ash at the air liquid interface. In Advanced Environmental Monitoring; Kim, Y. J.; Platt, U. , Eds.; Springer: New York, 2007. (15) Korell, J.; Seifert, H.; Paur, H.-R.; Andersson, S.; Bolin, P. Flue gas cleaning with the MercOx process. Chem. Eng. Technol. 2003, 26, 737–740.

8314

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 21, 2009

(16) Andersson, S.; Kreisz, S.; Hunsinger, H. PCDD/F removal from flue gases in wet scrubbers: A novel technique. Organohalogen Compd. 2002, 157–160. (17) Bologa, A.; Paur, H.-R.; Seifert, H.; Wa¨scher, T. Pilot-plant testing of a novel electrostatic collector for sub micrometer particles. IEEE Trans. Ind. Appl. 2005, 41, 882–890. (18) Bologa, A.; Paur, H.-R.; Seifert, H.; Wa¨scher, T.; Woletz, K. Novel wet electrostatic precipitator for collection of fine aerosol. J. Electrost. 2009, 67, 150–153. (19) Korell, J.; Paur, H.-R.; Seifert, H. Further development and quality control of a dry mercury speciation method. International conference on mercury as a global pollutant, Madison, WI, August 6-11, 2006. (20) Andersson, S.; Kreisz, S.; Hunsinger, H. ADIOX for wet scrubbers and dry absorbers. Filtr. Sep. 2005, 22–25. (21) Lo¨thgren, C.-J.; Takaoka, M.; Andersson, S.; Allard, B.; Korell, J. Mercury speciation in flue gases after an oxidative acid wet scrubber. Chem. Eng. Technol. 2007, 131–138.

ES901289G