Access the full text.
Sign up today, get DeepDyve free for 14 days.
Abstract The most abundant and typical reducing agent for oxidized mercury in the slurry of wet flue-gas desulphurization (FGD) is the absorbed sulphur dioxide (SO2), which is present as different species of bisulphite or sulphite, depending on the pH of the slurry. In this study, two different measurement principles for continuous sulphite measurement in the slurry of lab-scale FGD were investigated to check their feasibility to be implemented in a wet FGD. The first method is based on light absorbance at the characteristic wavelength of sulphite measurement using a spectrophotometer and, in the second method, sulphite is measured as sulphur dioxide using a gas sensor. In addition, the correlation of sulphite concentration and mercury (Hg) in the slurry can be shown by measuring sulphite semi-continuously. It was concluded that using a spectrophotometer leads to distorted results. In contrast, measuring sulphite as SO2 in the gas phase proved to be more selective. The implementation of the measurement technique in the lab-scale FGD showed promising results for sulphite measurement. Thus, the correlation of Hg and sulphite concentration could be shown at different synthetic slurries containing different halides. Using a slurry without halides demonstrated the ambivalent influence of sulphite in reactions involving Hg, in which sulphite acts as a ligand for Hg complexes as well as a reducing agent, depending on the existing concentration. However, in the presence of halides, the role of sulphite was less significant. Graphical Abstract Open in new tabDownload slide flue-gas cleaning, flue-gas desulphurization, mercury re-emission, sulphite measurement Introduction In coal-fired power plants and most fossil-fuelled industrial processes, mercury (Hg) is set free during combustion and can be emitted into the environment. Hg is ubiquitous due to its persistence and long-range atmospheric transport. Besides natural Hg emissions during volcanic eruptions or re-emission of soil and water, anthropogenic Hg emissions from industrial processes and power generation contribute a major share to the Hg concentration in the atmosphere [1, 2]. Because of its toxicity to human health and nature, governments all over the globe are implementing regulations and emission limits, striving for lower emissions [3, 4]. In many power plants, already installed air-pollutant cleaning devices (APCDs) contribute to Hg removal additionally to their original tasks. As a constituent of fuel, Hg is released as elemental mercury (Hg0) during the combustion process . At lower temperatures after the boiler, Hg0 can be oxidized in the presence of oxidizing agents such as hydrogen halides to oxidized mercury (Hg2+) compounds (e.g. HgCl2, HgBr2 and HgI2) homogeneously and heterogeneously [6–8]. The concentration of each hydrogen halide in the flue gas depends on the individual composition of the coal, in which hydrogen chloride (HCl) generally has the largest share. The Hg-oxidation reaction can be promoted in the selective catalytic reduction (SCR) catalyst, whose main role is NOx reduction with the use of a reducing agent such as ammonia. The heterogeneous oxidation of Hg in the SCR catalyst results in a higher share of Hg2+ downstream of the SCR catalyst, which has been shown in previous studies to be beneficial for Hg removal in subsequent APCDs [9, 10]. Accordingly, Hg2+ is readily adsorbed on fly-ash particles or activated carbon-based sorbents, while Hg0 shows inhibited adsorption behaviour . Particles are removed from flue gas in electrostatic precipitators (ESPs) or fabric filters and, consequently, adsorbed Hg is simultaneously taken out as particle-bound Hg (Hgp). The final step of the flue-gas-cleaning path is the wet flue-gas desulphurization (FGD) unit. Primarily targeted at removing SO2 from the flue gas, it also has the co-benefit of absorbing all water-soluble components, including Hg2+. Limestone (CaCO3) is one of the most common neutralization agents used in wet FGDs, counteracting the acidification of the scrubber caused by absorbed SO2. The scrubbing solution is brought into contact with the flue gas and the absorption process takes place. By implementing a forced oxidation process in the scrubber, gypsum is produced as a final product, which can be utilized as a construction material. Absorbed SO2 forms sulphurous acid (H2SO3), which dissociates to bisulphite (HSO3–) and sulphite (SO32–), as shown in Equation (1): SO2+3H2O⇌HSO3−+H3O++H2O⇌SO32−+2H3O+(1) From this equation, it can be concluded that the distribution of absorbed SO2 or sulphur in the oxidation state of 4 (S(IV)) depends strongly on the pH of the slurry. Fig. 1 illustrates the existing share of each S(IV) species depending on the pH of the solution at 60°C. It can be seen that, at the typical pH of FGDs (5–6), HSO3– is the most abundant species. Fig. 1: Open in new tabDownload slide Distribution of S(IV) species in an aqueous solution As the absorbed Hg2+ is a soft acid, it can undergo complexation reactions with soft bases according to the hard soft acid base theory . Halides and S(IV) are the existing ligands in the slurry, which can form Hg2+ complexes [13, 14]. Equation (2) represents the formation of halidomercurate complexes, in which X represents any of the Cl, Br and I. Equation (3) shows the possible formation of Hg2+ complexes with sulphite being the ligand. HgX2(aq)+2X(aq)−⇌HgX3(aq)−+X(aq)−⇌HgX4(aq)2−(2) Hg2++2SO32−⇌ Hg(SO3)+ SO32−⇌ [Hg(SO3)2]2−(3) The formation of heteroleptic complexes that contain different types of ligands as shown in Equations (4) and (5) is also a possibility: HgCl2+SO32−⇌[Cl2HgSO3]2−(4) [HgCl4]2−+2Br−⇌[HgBr2Cl2]2−+2Cl−(5) The probability of the formation of these complexes depends on the concentration of the respective ligand and their formation constant. For halidomercurate complexes, this means the higher the concentration of the halide, the higher the coordination number of the complex. The formation constant among different halides increases in the order of Cl < Br < I . However, it has to be noted that the complex of [Hg(SO3)2]2– has a fairly high formation constant, which increases the possibility of its existence at high concentrations of sulphite. In addition, SO32– is a softer base than Cl– and Br–, wherefore it can replace them in a complex when existing in adequate concentrations . In contrast to complexes, the bivalent mercury compounds (HgX2) have covalent bonds that are not dissociated in water. Thus, depending on their Henry coefficient, they can evaporate back to the gas phase, resulting in oxidized Hg re-emission. The Henry coefficient of bivalent mercury compounds increases in the order of HgI2 < HgBr2 < HgCl2, which shows the highest volatility for HgI2 . Therefore, a high concentration of halides enables the formation of complexes with high coordination numbers and prohibits HgX2 re-emission by stripping from the aqueous phase. It has been shown that FGDs can achieve high Hg-removal efficiencies. However, it has been observed that, under unfavourable operating conditions, previously absorbed Hg2+ can be reduced to Hg0 and subsequently re-emitted into the clean gas [18–20]. This occurs when the slurry has a reducing environment and depends on the half-cell potential of the Hg2+ complexes and the reducing agent. In a typical wet FGD, S(IV) is the most abundant reducing agent for Hg2+ complexes and the oxidation-reduction (redox) reaction may occur as shown in Equation (6): HgX2+SO32−+3H2O →Hg0+SO42−+2X−+2H3O+(6) It has to be mentioned that, for halidomercurate complexes, the driving force of reduction reactions increases in the order of I < Br < Cl and, for each specific halide, the lower the coordination number, the higher the driving force of the reaction . Evidently, sulphite has an ambiguous effect on Hg behaviour as either a complexation partner or a reducing agent. A deeper understanding of the influence and interplay of sulphite on Hg offers new possibilities for monitoring and controlling the Hg emissions of fossil-fuelled power plants. The approach of reducing Hg emissions by measuring and regulating sulphite concentrations has already been suggested in some works [22, 23]. The state-of-the-art sulphite-measurement method is iodometric titration. The principle is a back-titration of iodine as the oxidation agent with thiosulphate as the reducing agent. Unfortunately, this measurement method has some drawbacks. First, the samples can only be tested consecutively and not continuously. Furthermore, samples must be taken from the slurry, filtrated and titrated, which gives the dissolved sulphite in the filtrate enough time to be partially oxidized and results in a lower estimation of the real values. The preparation of the reagents has to be done meticulously and requires a laboratory. In addition, the process is not selective and counts all reducing agents in the solution as sulphite . Besides titration, sulphite concentration can be measured using other methods. An exact measurement principle is the measurement of sulphite concentration using ion chromatography. In this process, due to the time required between extracting the sample and measurement, specific chemicals have to be used to stabilize the sample and prevent its oxidation . This method requires an equipped laboratory and cannot be carried out continuously. Another continuous measuring method is using a sensor that sends a series of voltages into the slurry via two electrodes and evaluates the response currents. With these data, the correlating sulphite concentration can be calculated . In this study, two different principles for continuous sulphite measurement and the possibility of their implementation in lab-scale FGD are investigated. Furthermore, the correlation between sulphite concentration and Hg behaviour in the presence of different halides in the slurry is discussed. 1Methodology 1.1Description of the lab-scale wet FGD unit All the measurements were carried out in a lab-scale wet-limestone FGD unit. The lab-scale FGD unit was a fully automated, continuously working plant consisting of a gas-preparation section and a scrubber. A schematic of the set-up is shown in Fig. 2. Fig. 2: Open in new tabDownload slide Lab-scale FGD set-up and two different sulphite-measurement methods (dashed line: heating of the gas-preparation system) The flue gas was prepared synthetically as a mixture of gases that replicated the composition of real flue gas to some extent. The synthetic flue gas consisted of 15 vol.-% CO2, 3.5 vol.-% O2 and balance N2, and was injected with a flow rate of 4 l STP, dry/min. Furthermore, 3000 mg/m3STP, dry SO2 and 50 µg/m3STP, dry Hg were added to the flue gas using N2 as the carrier gas for Hg. The volume flows of the gases were controlled and regulated using mass-flow controllers. In addition to these gases, diluted hydrogen chloride (HCl) was added by a continuously dosing peristaltic pump providing 7 vol.-% humidity and 50 mg/m3STP, dry HCl concentration in the flue gas. The gas mixture was heated to 300°C using regulated heating bands and passed through a SCR catalyst to oxidize up to 50% of the Hg. Afterwards, the gas mixture entered at the bottom of the absorber at a temperature of ~120°C. The whole plant was made of glass to prevent interaction with Hg. The synthetic slurry was prepared with a volume of 1.5 l, which was heated to 60°C and stirred throughout the test. A peristaltic pump fed the slurry to the top of the absorber where it absorbed SO2 and Hg2+ while meeting the flue gas counter-currently. The slurry was collected at the bottom of the absorber and was transferred to an external sump via a siphon connection. This special connection was a customization of the lab-scale FGD unit that cannot be found in power plants and permits only liquids to pass. Thus, all Hg measured in the external sump was Hg that was absorbed at first and then re-emitted due to the conditions in the slurry. The FGD unit operated at a liquid-to-gas ratio of 20 l/m3. As the oxygen in the flue gas was not sufficient for complete oxidation of the sulphite, air was injected into the external sump, providing the needed oxygen for forced oxidation. Redox potential, pH and temperature were monitored in the external sump as well as at the bottom of the absorber. By adding calcium carbonate (CaCO3), a proportional–integral–derivative (PID) controller ensured a pH of 5.6 in the external sump. Additionally, dissolved oxygen was measured in the external sump. The basic slurry was composed of 10 wt.-% CaSO4∙2H2O, 300 µg/l Hg2+ and water. In different tests, the complexity was increased by adding 10 g/l chloride, 1 g/l bromide or 0.1 g/l iodide. The addition of halides enabled the investigation of each halide’s effect. Gas measurement could be carried out at the inlet and outlet of the absorber as well as at the external sump. Dissolved sulphite measurement was conducted using two different measurement principles by extracting slurry samples from the external sump. Both methods are shown in Fig. 2 and their measurement principle is explained later in detail. 1.2Analysis methods SO2 and CO2 concentrations were monitored at the absorber inlet and outlet using a non-dispersive infrared gas (NDIR) analyser and oxygen was measured using the paramagnetic principle. Hg concentration was measured at the inlet and outlet of the absorber and in the external sump using a continuous mercury analyser ‘Lumex RA 915 AMFG’ working with cold vapour atomic absorption spectroscopy with Zeemann-background correction. The analyser detects only Hg0. Thus, to measure the total amount of Hg (HgT), which is the sum of elemental and oxidized Hg, in the gas phase, it was necessary to reduce all Hg2+ to the elemental state by mixing the measurement gas with a reducing solution. To determine the ratio of Hg2+ to Hg0, an ion exchange resin was installed after the sampling point, selectively trapping all Hg2+ and consequently measuring solely Hg0. Redox potential was measured in this work using a silver/silver-chloride electrode. However, the values provided in the results section are corrected values using a specific factor and show the potential difference of the slurry to the standard hydrogen electrode. Sulphite was measured in the slurry using two different methods that were calibrated using iodometric titration as the state-of-the-art measurement. Due to the low water solubility of elemental iodine in water, it had to be produced by the reaction of iodide and iodate in an acidic environment by adding HCl in the titration vessel : IO3−+5I−+6H+→ 3I2+3H2O(7) As the reaction only takes place in an acidic environment, the solution of iodide iodate remains stable for a long time as long as there is no addition of acid to the solution. For the experiments in this study, a 0.05-mol/l iodine solution using potassium iodide and potassium iodate was prepared and, for the titration solution, sodium thiosulphate was used. The thiosulphate solution had to be prepared daily, as it is not stable over extended periods. Starch solution served as a colour indicator for the titration. The sample had to be filtrated and titrated immediately after sampling, as sulphite is oxidized continuously by atmospheric oxygen. Thus, long handling times result in the measurement of smaller sulphite concentrations than were initially present in the sample. After preparing the sample, the iodide/iodate solution was added and, by the addition of HCl, iodine was produced. At this point, the existing sulphite in the solution was oxidized using the produced iodine: I2+SO32−+H2O → 2I−+2H++SO42−(8) Afterwards, the remaining amount of iodine was titrated using thiosulphate solution as shown in Equation (9) and the sulphite concentration can be calculated: 2S2O32−+ I2→S4O62−+ 2I−(9) 2Results and discussion 2.1Measurement of sulphite using a spectrophotometer The measurement principle of the UV/VIS spectrophotometer is based on light absorption by molecules at a specific wavelength. By measuring the absorbed light at the characteristic wavelength, the concentration of the specific molecule can be calculated. A previous study investigated measuring sulphite using a spectrophotometer by adding different additives to make colourful complexes with sulphite and thereby enabling measurement of the concentration in the visible-light range [28, 29]. Measuring the light intensity of sulphite directly using a UV spectrophotometer has been also investigated in some studies [30, 31]. The implementation of this method in the lab-scale FGD unit and the possibility of its continuous operation are examined in this study. In order to measure sulphite continuously in the lab-scale FGD unit, a sample was extracted and passed through a filter in order to remove the solid fraction of the slurry. Afterwards, it was fed to the UV/VIS spectrophotometer for the analysis. In this way, the concentration of sulphite could be calculated using the Beer-Lambert law. As the pH of the slurry in the FGD unit was 5.6, according to Fig. 1, the dissolved SO2 or S(IV) existed mostly as HSO3– and a small fraction of SO32–. In order to measure the total S(IV) concentration in the slurry at one specific wavelength, S(IV) must be present in only one form. Thus, the pH of the slurry was shifted using a buffer solution. The choice of a suitable buffer solution and the proper mixing ratio have been investigated in a previous study , which showed the dependency of the absorbance maximum of different sulphite samples on the pH of the solution. The spectrum of the components present in the suspension was investigated individually to check for possible spectral interferences in the sulphite measurement. For this reason, three samples were prepared, each containing one of the following components: sulphite, chloride or mercury with concentrations of 130 mg/l, 10 g/l and 100 µg/l, respectively. Each sample was measured in combination with the buffer solution at a 1:1 ratio. Fig. 3 represents the spectra of these samples between 190 to 250 nm. It can be seen that the maximum wavelengths for sulphite, chloride and mercury are at 205, 204 and 207 nm, respectively. To measure sulphite in the presence of these components, the same slurry matrix without sulphite was used as the baseline measurement for the spectrophotometer. In this way, only sulphite ought to be measured. Fig. 3: Open in new tabDownload slide Spectrum of 130 mg/l sulphite, 100 µg/l mercury and 10 g/l chloride in aqueous solution (190–250 nm) The results obtained from this study reveal the detrimental effect of chloride on the measurement of sulphite using a spectrophotometer. As mentioned before, the maximum wavelengths for sulphite, mercury and chloride are close to each other and, despite using the same slurry as the baseline in order to eliminate the measurement of these components in the analysis, the measurement of sulphite in the slurry containing chloride was not successful. The results show the sensitivity of the spectrophotometric measurement to the components present in the slurry. Therefore, the implementation of this method in a real system with slurry with a more complex matrix is not feasible. 2.2Measurement of sulphite using an SO2gas sensor Taking the equilibrium reaction of S(IV) in aqueous solutions and the pH dependency of this reaction as shown in Fig. 1 into account, an acidic environment shifts the equilibrium to the side of dissolved SO2 gas. Thus, SO2 can be stripped from the sample and measured subsequently using a gas analyser. After mixing with H2SO4, the sample was pumped into a reactor, in which the dissolved SO2 was stripped using air. Afterwards, the SO2-containing airflow passed a SO2 gas sensor that measured the SO2 concentration in the gas flow. The concentration of dissolved sulphite can be calculated according to the SO2 gas concentration. Due to the nature of the slurry, which contained solid fractions, the sample was filtered and fed to the mixing zone using a pump. As the extraction of the sample decreased the volume of the already small slurry, the measurement duration was limited and sulphite was measured semi-continuously. It must be mentioned that the goal of this study is to prove the measurement concept and the feasibility of the principle to be used in full-scale FGD units, in which the problem of a limited sample is not an issue. Before starting the measurement, the measurement principle was calibrated using different concentrations of Na2SO3 solution, whose concentration was measured using iodometric titration. It must be noted that the influence of chloride and mercury on the measurement of the sulphite by the gas sensor was investigated using solutions with the same concentration of chloride and mercury as the synthetic slurry and the selective measurement of sulphite was proven. 2.3Measurement of sulphite in the lab-scale FGD unit As no reliable results could be obtained using the spectrophotometer in slurries containing halides, only the feasibility of measuring sulphite using the SO2 gas sensor was investigated in the lab-scale FGD. In order to trigger a change in the sulphite concentration, the SO2 concentration in the flue gas entering the FGD unit was changed in different steps from 3000 to 5000 mg/m3 and back to 3000 mg/m3 to simulate a sudden increase as well as decrease in the dissolved S(IV) concentration. As sulphite is the main reducing agent for Hg2+ in the slurry, which results in the re-emission of Hg0, it is important to examine the influence of its sudden change on Hg re-emission as well. Thus, the dynamic behaviour of Hg re-emission by making a sudden change in sulphite concentration is also addressed. As mentioned before, Hg re-emission is the total Hg concentration, which is measured at the top of the external sump. At first, a synthetic slurry containing no halides was chosen in order to define and understand the direct relationship between Hg and sulphite in the slurry without having other halidomercurate complexes interfering with the reactions involving mercury. Afterwards, the influence of each halide in the slurry on sulphite measurements as well as Hg re-emission by a sudden change in SO2 concentration was investigated in order to show the role of halides in preventing Hg re-emission. Therefore, four tests were carried out with four different synthetic slurries containing no halides or one of the following: 10 g/l chloride, 1 g/l bromide or 0.1 g/l iodide. For each measurement, the first figure shows the dynamic measurement of the redox potential and dissolved oxygen according to the change in SO2 concentration and the second figure reveals the continuous Hg re-emission and its correlation with the measured sulphite concentration. Unlike all other parameters, sulphite could only be measured semi-continuously, as the small slurry volume posed a limitation to the amount of sample extraction. Figs 4 and 5 illustrate the online measurement results of the slurry containing no halides. At the beginning of the experiment, the redox potential had a value of 320 mV, dissolved oxygen at ~1 ppm, sulphite concentration was measured at ~10 mg/l and HgT re-emission above the external sump showed a concentration of ~60 µg/m3. A change in SO2 concentration to 5000 mg/m3 resulted in a change in all the parameters after some minutes. The dissolved oxygen and redox potential declined, while the sulphite concentration and HgT re-emission rose. It has to be noted that the changes were not simultaneous, but dissolved oxygen and HgT re-emission started first. After some minutes, dissolved oxygen decreased to zero and, at this moment, the redox potential started to fall with a simultaneous increase in the sulphite concentration. The accumulating amount of sulphite consumed the dissolved oxygen in the slurry and the injected oxidation air did not suffice to oxidize the high amount of absorbed SO2. Thus, the concentration of S(IV) increased rapidly, which resulted in higher Hg re-emission from the slurry. Eight minutes after the change in SO2, the concentration of HgT re-emission reached the maximum value of ~355 µg/m3, followed by a decrease. By keeping the same conditions, the system stabilized and HgT re-emission dropped slowly to even lower values than the starting concentration with a simultaneous increase in the sulphite concentration to its maximum at 115 mg/l. Fig. 4: Open in new tabDownload slide Dynamic behaviour of dissolved oxygen and redox potential by change in SO2 concentration in a slurry containing no halides Fig. 5: Open in new tabDownload slide Dynamic behaviour of total Hg re-emission and sulphite concentration by change in SO2 concentration in a slurry containing no halides These results could lead to a conclusion on the ambivalent influence of sulphite in reactions involving Hg. As there was no halide present in this experiment, sulphite played an important role as a ligand for the Hg-complex formation in addition to its reducing-agent role. An increase in SO2 and subsequently sulphite resulted in a sharp increase in Hg2+ reduction; however, by further increasing the sulphite concentration, HgT re-emission decreased due to the formation of Hg-sulphite complexes. It can be concluded that, in the absence of halides, a higher concentration of sulphite results in a larger share of Hg2+ as the mercuric disulphite complex [Hg(SO3)2]2–. This complex has a slower decomposition rate to Hg0 than the monosulphite complex HgSO3 that forms at low sulphite concentrations. The conclusion is in accordance to the results obtained in another study , which suggests the positive effect of a higher sulphite concentration on the stability of Hg complexes in the absence of halides. By changing the SO2 concentration back to 3000 mg/m3, the redox potential started to rise and reached a steady value as soon as dissolved oxygen reached values above zero. Exactly at this point, HgT re-emission fell again to lower values. It has to be mentioned that a time difference was observed between the change in SO2 concentration and the ensuing increase in dissolved oxygen in the slurry, subsequently reaching steady state. During this time, all the excess sulphite in the slurry was being oxidized and the dissolved oxygen showed an increase as soon as there was more oxygen dissolved than the necessary amount required for the oxidation of existing sulphite in the slurry. It could be seen that, before reaching equilibrium, a decrease in the sulphite concentration at this step increased Hg2+ reduction first. Hence, the system revealed a second peak when the sulphite concentration decreased again to ~13 mg/l, which is almost the same range of concentrations as when the Hg re-emission showed the first peak. The latter can be due to the decomposition of Hg(SO3)22– to HgSO3, which has a higher decomposition rate and formation of Hg0. The study of Chang  reveals the same trend for Hg re-emission; however, the experiments have been carried out in a bench system without including Ca2+ in the slurry. It can be concluded that, in this system, at this range of sulphite concentrations, the reduction in Hg2+ is more favourable due to the formation of HgSO3 and its fast decomposition. The same procedure was repeated for slurries containing 10 g/l chloride, 1 g/l bromide and 0.1 g/l iodide, and Figs 6–11 represent the online measurement results, respectively. Fig. 6: Open in new tabDownload slide Dynamic behaviour of dissolved oxygen and redox potential by change in SO2 concentration in a slurry containing 10 g/l chloride Fig. 7: Open in new tabDownload slide Dynamic behaviour of total Hg re-emission and sulphite concentration by change in SO2 concentration in a slurry containing 10 g/l chloride Fig. 8: Open in new tabDownload slide Dynamic behaviour of dissolved oxygen and redox potential by change in SO2 concentration in a slurry containing 1 g/l bromide Fig. 9: Open in new tabDownload slide Dynamic behaviour of total Hg re-emission and sulphite concentration by change in SO2 concentration in a slurry containing 1 g/l bromide Fig. 10: Open in new tabDownload slide Dynamic behaviour of dissolved oxygen and redox potential by change in SO2 concentration in a slurry containing 0.1 g/l iodide Fig. 11: Open in new tabDownload slide Dynamic behaviour of total Hg re-emission and sulphite concentration by change in SO2 concentration in a slurry containing 0.1 g/l iodide The steady-state values of the redox potential at 3000 mg/m3 SO2 ranged from 260 mV for the iodide-containing slurry to 406 mV for the slurry containing chloride. In each case, dissolved oxygen was ~1 ppm and sulphite concentration ~10 mg/l. For slurries containing chloride, bromide and iodide, Hg re-emission was 14, 6 and 11 µg/m3, respectively, lower than the value of the slurry without halides. The results are in agreement with previous studies [34, 21], as the presence of halides plays an important role in the formation of Hg complexes. As mentioned previously, the reaction rate of Hg2+ reduction for the various halidomercurate complexes is related to their half-cell standard electrode potential and is in the order of Cl > Br > I. It has to be considered that the measured Hg re-emission in this study referred to the total Hg re-emission. Comparing the total Hg re-emission, the slurry containing iodide showed higher re-emission than the slurry with bromide. The distribution of Hg0 and Hg2+, which has been investigated in a previous study , revealed that >90% of the total Hg re-emitted consisted of Hg2+ species in the presence of iodide, most likely HgI2, due to its highest vapour pressure. By increasing the SO2 concentration to 5000 mg/m3 in the flue gas entering the FGD unit, the same trend for all parameters was repeated as in the slurry without any halide. In all of the slurries, the dissolved oxygen and redox potential declined, and the sulphite concentration and HgT re-emission increased. The fall in the redox potential started as soon as the amount of dissolved oxygen reached zero. Depending on the existing halides in the slurry and the formation of halidomercurate complexes, the height of the Hg peak varied and its form was less pronounced. Comparing all of the graphs, it can be seen that the highest peak in the HgT re-emission concentration belongs to the slurry containing no halides at ~355 µg/m3, followed by the slurries containing chloride, bromide and iodide at 108, 78 and 33 µg/m3, respectively. Even though the slurries containing bromide and iodide showed a lower redox potential representing a reducing environment, the peak of the re-emitted Hg by increasing the S(IV) concentration was lower due to the presence of more stable Hg complexes. In the case of the iodide-containing slurry, the change in the redox potential during the whole experiment was insignificant. As the amount of oxidation air was kept constant, the increase in HgT re-emission was due to the reduction in Hg2+. Thus, the height of the Hg re-emission peak for the slurries containing halides is in the order of Cl > Br > I, as explained before. After the peak in Hg re-emission, as for the slurry without halides, the Hg re-emission decreased for the slurries containing chloride and bromide. However, the decrease was not as significant as in the no-halide-containing slurry. Hg behaviour in the slurry containing iodide differed slightly from all other slurries and was generally much lower. As the decrease in Hg re-emission was observed with the simultaneous increase in sulphite concentration and the minimum amount of Hg re-emission at 5000 mg/m3 was reached, when the sulphite reached its maximum concentration, the role of the sulphite in the complex formation can be concluded. However, the importance of sulphite in forming complexes differs depending on the composition of the slurry. As halides are suitable ligands for Hg as well, the increase in the sulphite concentration in the slurries containing halides showed a smaller effect on Hg complexation after the peak. For the chloride-containing slurry as shown in Fig. 7, sulphite played a role in the complex formation as the Hg re-emission decreased slightly with a higher sulphite concentration. However, for slurries containing bromide and iodide, this effect cannot be observed. This can be explained according to the formation constant of different halidomercurate complexes. When considering the homoleptic halidomercurate complexes, their formation constants are in the following order : [HgI4]2– > [Hg(SO3)2]2– > [HgBr4]2– > [HgCl4]2– > HgSO3 meaning that the possibility of the formation when all the ligands are present is the highest for iodomercurate, which is the reason why sulphite does not play a role in forming complexes with Hg when iodide is present. However, the formation constant depends on the concentration of the present ligands and considering the heteroleptic mercury complexes brings more complexity to the system as well. In general, it can be seen that the presence of halides influences the role of sulphite as a ligand as well as a reducing agent for Hg. The increase in sulphite concentration was stopped as soon as the SO2 concentration was changed back to 3000 mg/m3 and, after a short time, the accumulated sulphite in the slurry started to oxidize and thus a decrease in the sulphite concentration could be observed. The redox potential of the slurry rose slowly and steadied as soon as the dissolved oxygen concentration reached values above zero. At that point, a sharp drop in HgT re-emission took place. The maximum sulphite concentration depends on the duration of the flue-gas injection containing 5000 mg/m3 SO2. By keeping the system at this condition for longer, the accumulation of sulphite in the slurry would be higher. For all measurements, the system was running with a high SO2 concentration for the duration of 40–60 minutes and the maximum sulphite concentration was measured in the range of 115–130 mg/l for all the slurries besides the iodide-containing slurry, which reached 70 mg/l sulphite after 40 minutes of measurement. It must be noted that, due to the reduction in the volume of the slurry for the sample of the sulphite measurement, the change in SO2 back to 3000 mg/m3 did not necessarily bring the system to a similar steady-state condition as at the beginning of the measurement, especially in the case of the chloride-containing slurry. However, the goal of this study is to prove the sulphite-measurement concept and this problem only applies to the small lab-scale system. Concluding from all the results, the amount of sulphite in the slurry has to be kept within the range of ≤10 mg/l, regardless of the slurry composition, in order to prevent high Hg re-emission. However, this value is only in the case of the studied lab-scale FGD unit and has to be determined for full-scale FGD units. The concentration of the sulphite in the slurry can be controlled at the desired value by precise injection of oxidation air to prevent sudden peaks of Hg re-emission due to sudden changes in the FGD parameters. 3Summary and conclusion Monitoring and regulating the sulphite concentration in FGD slurries of power plants is a viable possibility to control Hg emissions and to prevent re-emission peaks. This study presents two methods for measuring the sulphite and checks their feasibility to be used in a lab-scale firing system. For the first method, a spectrophotometer was used to measure the absorbance of light at the characteristic wavelength of sulphite. In pre-tests, measuring sulphite in solutions without any other components was successfully tested. However, adding chloride, which is a major component in FGD slurries, the analysis was distorted. Thus, the measurement of sulphite in slurries containing chloride was impossible and therefore it is not a feasible option for the analysis of FGD slurries from power plants. The second method using a SO2 gas sensor proved to be more selective. By shifting the pH of the samples to <2, the dissolved S(IV) species were converted to SO2, stripped from the sample and measured using a SO2 gas analyser. The sensor was implemented in the set-up of a lab-scale FGD unit and tested with slurries containing either no halides, chloride, bromide or iodide. The strongest interaction of sulphite and Hg was observed in the slurry without halides. In this case, sulphite is the only partner for the formation of complexes that stabilize Hg in the slurry. However, at concentrations of ~10 mg/l, sulphite acted as a reducing agent, resulting in high Hg re-emission peaks. The ambiguous role of sulphite as a reducing agent and complexation partner was obvious. In all tests with halide-containing slurries, Hg re-emission was generally on a lower level than in the absence of halides. Here, a correlation between sulphite concentration and Hg behaviour was less pronounced, as halides act as strong partners for complex formation. In the case of iodide, no relation was apparent between Hg and sulphite. This can be explained by the formation constants of the various halidomercurate complexes and the hard soft acid base theory. The results prove the sulphite measurement with the used principle as a promising method to measure sulphite continuously, especially in full-scale FGD units, where the amount of the required sample is negligible compared to the volume of the slurry and a continuous measurement can be implemented. Conflict of Interest None declared References  UNEP . Global Mercury Assessment 2013: Sources, Emissions, Releases and Environmental Transport , Geneva: United Nations Environment Programme, 2013 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC  Yudovich YE , Ketris MP. Mercury in coal: a review. Part 2. Coal use and environmental problems . Int J Coal Geol 2005 , 62 : 135 – 65 . Google Scholar Crossref Search ADS WorldCat  Thierry L , Ferrería de la Fuente , JF, Frederik N, et al. Best Available Techniques (BAT) Reference Document for Large Combustion Plants . Industrial Emissions Directive 2010/75/EU (Integrated Pollution Prevention and Control) , Luxembourg: European Commission, 2017 . Google Scholar OpenURL Placeholder Text WorldCat  National Emission Standards for Hazardous Air Pollutants From Coal- and Oil-Fired Electric Utility Steam Generating Units and Standards of Performance for Fossil-Fuel-Fired Electric Utility, Industrial-Commercial-Institutional, and Small Industrial-Commercial-Institutional Steam Generating Units . Final Rule; 40 US Code of Federal Regulations Parts 60 and 63 , Washington, DC : United State Environmental Protection Agency (EPA) , 2016 . OpenURL Placeholder Text WorldCat  Tang T , Xu J, Lu R, et al. Enhanced Hg2+ removal and Hg0 re-emission control from wet fuel gas desulfurization liquors with additives . Fuel 2010 , 89 : 3613 – 7 . Google Scholar Crossref Search ADS WorldCat  van Otten B , Buitrago PA, Senior CL, et al. Gas-Phase oxidation of mercury by bromine and chlorine in flue gas . Energy Fuels 2011 , 25 : 3530 – 6 . Google Scholar Crossref Search ADS WorldCat  Chi Y , Yan N, Qu Z, et al. The performance of iodine on the removal of elemental mercury from the simulated coal-fired flue gas . J Hazard Mater 2009 , 166 : 776 – 81 . Google Scholar Crossref Search ADS PubMed WorldCat  Wilcox J , Rupp E, Ying SC, et al. Mercury adsorption and oxidation in coal combustion and gasification processes . Int J Coal Geol 2012 , 90–91 ( Suppl C ): 4 – 20 . Google Scholar Crossref Search ADS WorldCat  Zhou Z-J , Liu X-W, Zhao B, et al. Effects of existing energy saving and air pollution control devices on mercury removal in coal-fired power plants . Fuel Process Technol 2015 , 131 : 99 – 108 . Google Scholar Crossref Search ADS WorldCat  Wu C , Cao Y, Dong Z, et al. Evaluation of mercury speciation and removal through air pollution control devices of a 190 MW boiler . J Environ Sci (China) 2010 , 22 : 277 – 82 . Google Scholar Crossref Search ADS PubMed WorldCat  Wang S , Wu H. Environmental-benign utilisation of fly ash as low-cost adsorbents . J. Hazard. Mater 2006 , 136 : 482 – 501 . Google Scholar Crossref Search ADS PubMed WorldCat  Pearson RG . Hard and soft acids and bases . J Am Chem Soc 1963 , 85 : 3533 – 3539 . Google Scholar Crossref Search ADS WorldCat  Griffiths TR , Anderson RA. The electronic spectra of the mixed mercury dihalides. Part 1. Computational procedures for calculating spectra, for a new route to equilibrium and formation constants, and the resolved spectra . J Chem Soc, Dalton Transactions 1980 : 205 – 8 . doi: 10.1039/DT9800000205. Google Scholar OpenURL Placeholder Text WorldCat  Griffiths TR , Anderson RA. The electronic spectra of the mixed mercury dihalides. Part 2. Identification, equilibrium and formation constants, and assignment of transitions . J Chem. Soc, Dalton Transactions 1980 : 209 – 15 . 10.1039/DT9800000209. Google Scholar OpenURL Placeholder Text WorldCat  Powell KJ , Brown PL, Byrne RH, et al. Chemical speciation of environmentally significant heavy metals with inorganic ligands. Part 1. The Hg2+– Cl–, OH–, CO32–, SO42–, and PO43– aqueous systems (IUPAC Technical Report) . Pure Appl Chem 2005 : 739 – 800 . doi: 10.1351/pac200577040739. Google Scholar OpenURL Placeholder Text WorldCat  Lin C-J , Pehkonen SO. Aqueous phase reactions of mercury with free radicals and chlorine: implications for atmospheric mercury chemistry . Chemosphere 1999 , 38 : 1253 – 63 . Google Scholar Crossref Search ADS WorldCat  Riedel E , Alsfasser R, Janiak C, et al. Moderne Anorganische Chemie, 3rd edn . Berlin: Walter de Gruyter , 2007 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC  Heidel B , Hilber M, Scheffknecht G. Impact of additives for enhanced sulfur dioxide removal on re-emissions of mercury in wet flue gas desulfurization . Appl Energy 2014 , 114 ( Suppl C ): 485 – 91 . Google Scholar Crossref Search ADS WorldCat  Omine N , Romero CE, Kikkawa H, et al. Study of elemental mercury re-emission in a simulated wet scrubber . Fuel 2012 , 91 : 93 – 101 . Google Scholar Crossref Search ADS WorldCat  Wo J , Zhang M, Cheng X, et al. Hg2+ reduction and re-emission from simulated wet flue gas desulfurization liquors . J Hazard Mater 2009 , 172 : 1106 – 10 . Google Scholar Crossref Search ADS PubMed WorldCat  Heidel B , Klein B. Reemission of elemental mercury and mercury halides in wet flue gas desulfurization . Int J Coal Geol 2017 , 170 : 28 – 34 . Google Scholar Crossref Search ADS WorldCat  Appelo PA , Laslo DJ Sulfite Control to Reduce Mercury Re-emission (US 8,828,341 B1) , 2014 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC  Laslo DJ Oxidation Control for Improved Flue Gas Desulfurization Performance (US 9,321,006 B2) , 2016 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC  Metrohm Application Bulletin Nr. 225/2 d . http://products.metrohm.com (2 April 2020, date last accessed) .  Deutsche Institut für Normung e.V . Bestimmung der gelösten Anionen mittels Ionenchromatographie. Teil 3: Bestimmung von Chromat, Iodid, Sulfit, Thiocyanat und Thiosulfat (EN ISO 10304-3) , 1997 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC  Larsson M , Brogaard FJ, Winquist F Sulphite Sensor and Method for Measuring Sulphite Concentration in a Substance (EP 2,579,032 A1) , 2015 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC  VGB PowerTech e. V . VGB Merkblatt M701: Analyse von REA-Gips , Essen: VGB PowerTech , 2008 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC  Abdel-Latif MS . New spectrophotometric method for sulfite determination . Anal Lett 1994 , 27 : 2601 – 14 . Google Scholar Crossref Search ADS WorldCat  Singla M , Singh M, Walia MS, et al. An optoelectronic instrument for the determination of sulphite in beverages . Food Chem 2006 , 97 : 737 – 41 . Google Scholar Crossref Search ADS WorldCat  Heidel B , Thorwarth H, Scheffknecht G Spektralfotometrische Bestimmung der Sulfitkonzentration in Suspensionen von nassen Rauchgasentschwefelungsanlagen . Hamburg: VGB Konferenz Chemie im Kraftwerk , 2012 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC  Golding RM . Ultraviolet absorption studies of the bisulphite-pyrosulphite equilibrium . J Chem Soc 1960 : 3711 – 3716 . 10.1039/JR9600003711. Google Scholar OpenURL Placeholder Text WorldCat  Blythe GM , Currie J, DeBerry DW Bench-scale Kinetics Study of Mercury Reactions in FGD Liquors: Final Report. Cooperative Agreement No: DE-FC26-04NT42314 . Austin, TX : URS Group , 2008 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC  Chang L , Zhao Y, Li H, et al. Effect of sulfite on divalent mercury reduction and re-emission in a simulated desulfurization aqueous solution . Fuel Process Technol 2017 , 165 : 138 – 44 . Google Scholar Crossref Search ADS WorldCat  Bittig M Zum Einfluss unterschiedlicher Liganden auf die Quecksilberabscheidung in absorptiven Abgasreinigungsstufen . Dissertation. Universität Duisburg-Essen , 2010 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC  Masoomi I , Heidel B, Schmid MO, et al. Effect of additives on mercury partitioning in wet-limestone flue-gas desulfurization . Clean Energy 2020 , 4 : 132 – 41 . Google Scholar Crossref Search ADS WorldCat  Heidel B Wechselwirkungen bei der Abscheidung von Schwerfeldioxid und Quecksilber durch nasse Rauchgasentschwefelungsanlagen . Dissertation. Universität Stuttgart , 2015 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC © The Author(s) 2020. Published by Oxford University Press on behalf of National Institute of Clean-and-Low-Carbon Energy This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact email@example.com © The Author(s) 2020. Published by Oxford University Press on behalf of National Institute of Clean-and-Low-Carbon Energy
Clean Energy – Oxford University Press
Published: Dec 31, 2020
Access the full text.
Sign up today, get DeepDyve free for 14 days.