Article pubs.acs.org/est
Influence of Limestone Characteristics on Mercury Re-emission in WFGD Systems Raquel Ochoa-González, Mercedes Díaz-Somoano,* and M. Rosa Martínez-Tarazona Instituto Nacional del Carbón (INCAR), CSIC. C/Francisco Pintado Fé, 26, 33011, Oviedo, Spain ABSTRACT: This work evaluates the influence of the effect of the properties of limestones on their reactivity and the reemission of mercury under typical wet scrubber conditions. The influence of the composition, particle size, and porosity of limestones on their reactivity and the effect of sorbent concentration, pH, redox potential, and the sulphite and iron content of the slurry on Hg0 re-emission was assessed. A small particle size, a high porosity and a low magnesium content increased the high reactivity of the limestones. Moreover, it was found that the higher the reactivity of the sample the greater the amount of mercury captured in the scrubber. Although sulphite ions did not cause the re-emission of mercury from the suspensions of the gypsums, the limestones enriched in iron increased Hg0 re-emission under low oxygen conditions. It was observed that the low pH values of the gypsum suspensions favored the cocapture of mercury because Fe2+ formation was avoided. The partitioning of the mercury in the byproducts of the scrubber depended on the impurities of the limestones rather than on their particle size. No leaching of mercury from the gypsum samples occurred suggesting that mercury was either tightly bound to the impurities of the limestone or was transformed into insoluble mercury species.
■
inlet, which in turn depends on the coal rank, the flue gas composition,10−14 the presence of a selective catalytic reduction (SCR) system in the installation and the type of the catalyst.15,16 If SCR units are installed, the capture of Hg2+ in WFGD systems is increased up to 89%.17 It is well-known that gas and slurry composition, oxidation modes, operating temperature, and the pH of the slurry affect the efficiency of mercury removal in scrubbers.18 The chemical reduction of the dissolved mercury, which is known as reemission, reduces this efficiency.19,20 However, the reaction mechanisms and the specific conditions under which the reemission of mercury arises are not fully understood. Several studies have identified the reductive effect of SO2 via its intermediate soluble species, that is, SO32−, HSO3−, as the main influencing parameter.21 The pH and temperature are also relevant factors in most cases.22,23 In addition, the effect of pH on the coremoval of mercury in WFGD systems has still not been fully explained. Some authors have observed a greater retention of mercury at high pH values.20,22,24 pH values higher than 7.0 would cause the oxidation of sulphite and bisulphite favoring reaction between sulfate and mercury species.22 However, Wu et al. (2010) have observed the opposite effect in suspensions containing calcium sulphite.25 Moreover, few studies at laboratory scale dealt with the addition of calcium carbonate to facilitate pH control.24 Some authors have
INTRODUCTION The wet flue gas desulphurization (WFGD) or scrubbing process, in which SO2 and other acid components are removed from the combustion flue gas using a limestone or lime slurry, is a well-established technology for flue gas cleaning in coal fired power plants. Limestone reactivity indicates the rate of providing alkalinity and reacting with the SO2 in the flue gas. This parameter is necessary to estimate the amount of limestone that needs to be added for a given pH to be maintained in the SO2 absorber.1 Limestone properties play a fundamental role in the sustainable production of energy from coal because limestones with high SO2 removal rates minimize the amount of reagent that needs to be used. Recently research has been performed on the relationship between the properties of limestones and their reactivity.2−6 There is now a convincing body of evidence that shows that the rate-controlling steps in the scrubbing process depend on the particle size and porosity of the limestones.3,4,6 A relationship has been established between limestone dissolution rates and particle size.3 In addition, the apparent porosity of the alkali sorbents shows a positive correlation with their reactivity.6 Although scrubbers are designed to capture sulfur species, they may also be effective in reducing the emissions of mercury from coal fired power plants.7,8 Oxidised mercury species (Hg2+), which are soluble in water, are removed in WFGD systems which exhibit efficiencies ranging from 40 to 90%.7,9,10 In general, insoluble elemental mercury (Hg0) passes through WFGD systems without being altered by chemical processes. Consequently, the coremoval of mercury in these systems depends on the speciation of gaseous mercury at the WFGD © 2013 American Chemical Society
Received: Revised: Accepted: Published: 2974
October 9, 2012 February 25, 2013 February 25, 2013 February 25, 2013 dx.doi.org/10.1021/es304090e | Environ. Sci. Technol. 2013, 47, 2974−2981
Environmental Science & Technology
Article
suggested that minor components from limestone dissolution such as Fe, Sn, Cu, Mn, Ni, and Co may reduce Hg2+ species.26 In general, there is still a scarcity of data with respect to the retention of mercury, the chemical and physical characteristics of limestones, and the partitioning of mercury in scrubber byproducts. A better understanding of the limestone constituents, their reactivity, pH, sulphite, and metal content would contribute to the optimization of the WFGD process for the cocapture of mercury. The aim of this work was to investigate the effect of different types of natural limestones on their reactivity with sulfur and on the re-emission of Hg0. The effects of sorbent content, sulphite concentration, pH, redox potential, and the contribution of metals are discussed. The partitioning of the mercury was also studied to establish the behavior and the affinity of mercury species for gypsum or liquid phase in wet scrubber conditions.
CaCO3(s) + H 2SO4 (aq) → CaSO4 (s) + CO2 (g) + H 2O (1)
Mercury Retention Tests and Analysis. Mercury Retention System. To simulate the equilibrium processes in the scrubbers, a lab-scale device was used as reported elsewhere.23 The reactor consisted of a three-necked roundbottom glass flask fitted with a thermostat system to ensure a constant temperature of 40 °C. A calibration gas generator system (HovaCAL, IAS GmbH) coupled to an evaporator was employed to generate oxidized mercury species in gas phase by means of the evaporation of a 1 μg mL−1 mercury solution in 10 mmol L−1 hydrochloric acid. Hydrochloric acid was added to cause the mercury to stabilize in the solution as HgCl2 and generate chlorine species in the flue gas. The gypsum slurry was prepared by adding sulphuric acid to the limestones. WFGD units normally operate with a percentage of solids of 10−15%. In this study the solids content was 10 times less than what is usually employed in WFGD suspensions. The initial concentration of the limestone in the slurry was fixed at 1% to generate gypsums whose mercury concentrations would be high enough for analysis. To evaluate the effect of the pH, it was adjusted continuously by adding 0.1 N H2SO4 with a titrator (Mettler Toledo DL53). The redox potential was measured with an Orion electrode 9678BNWP and values were time recorded continuously using an Orion Meter (Model 720A+). A simulated flue gas containing nitrogen and 50 μg m−3 of Hg2+ was passed through the slurry solution. The Hg0 generated in the scrubbing solution was extracted by using 3 L min−1 of nitrogen. A continuous mercury emission monitor VM 3000 (Mercury instruments) was employed to monitor the Hg0 concentration in the flue gas at the outlet of the reactor. The VM 3000 was based on the UV absorption of the Hg0 and its detection limit was 0.1 μg m−3. The experiments were performed with a duration of 150 min in duplicate. Mercury Analysis. The mercury retained in the gypsums and in the aqueous phase of the slurries was determined. The mercury content in the original limestone, gypsum, and water samples generated in the lab-scale experiments was determined using a LECO Automatic Mercury Analyzer (AMA 254). This analyzer had a detection limit of 0.01 ng and the mercury was measured by atomic absorption at 253.7 nm. The gypsum samples were dried for 48 h at 40 °C prior to analysis to avoid the decomposition of the samples and loss of the retained mercury. Mercury Stability of the Gypsum Samples. The stability of the gypsum samples obtained in the tests was evaluated in water. The leaching tests were performed following the UNEEN 12457−2 standard by stirring the sample in water for 24 h. The suspension produced in the experiments was filtered and the dissolved mercury was analyzed by means of the analyzer AMA 254. In this study, 0.5 g of sample was mixed with 20 mL of water and all of the tests were carried out in duplicate.
■
EXPERIMENTAL SECTION Limestone Samples and Characterization. Three limestone samples obtained from natural deposits were tested. The samples, which were labeled CaP, CaC, and CaB, were characterized both physically and chemically. Calcium carbonate reagent (CaR) was employed as a reference material in the tests. Calcium (Ca), magnesium (Mg), aluminum (Al), silicon (Si), phosphorus (P), sodium (Na), and potassium (K) contents were determined by X-ray fluorescence in a fluorescence spectrometer Bruker SRS3000 after melting the samples with Li2B4O7 and LiF at 1200 °C. For the analysis of trace metals, the samples were previously digested with HNO3 in a microwave oven and the solutions were analyzed in an ICP-MS 7700× Agilent device equipped with a He collision cell to prevent matrix interferences. The crystalline species were identified by X-ray diffraction (XRD) (Bruker D8 Advance) and the morphological study was performed by scanning electron microscopy (SEM) using a FEI Quanta 650 FEG apparatus. The particle size was determined in a Coulter Counter apparatus (Beckman Coulter LS 13 320). The apparent density (ρHg) was obtained by mercury porosimetry in a Micromeritics Autopore IV 9500 apparatus. For the porosimetry analysis, the samples were outgassed at 120 °C during 2 h. The true density (ρHe) was determined with a Micromeritics AccuPyc 1330 helium pycnometer after the samples had been degasified at 120 °C for 12 h. The apparent porosity (ε) was calculated by means of the following equation: ⎛ ρHg ⎞ ⎟⎟100 ε = ⎜⎜1 − ρHe ⎠ ⎝
The chemical reactivity of the limestones was determined through neutralizing the gypsum slurries with a diluted solution of 0.5 N H2SO4 at a pH value of 5.0 ± 0.1 at 50 °C for 6 h, using a Mettler Toledo DL53 titrator. The slurry containing 1.00 g of limestone and 1.00 g of CaSO4·2H2O of analytical grade was stirred at 400 rpm. The addition of CaSO4·2H2O provided additional ions to the slurry. When the pH was higher than the set value, the peristaltic pump of the titrator added the necessary amount of acid to reduce the pH. The conversion of the limestones (X), that is, the mass fraction of carbonates dissolved over a period of 6 h, was determined from the stoichiometric ratio of the sulphuric acid added to that consumed every 9 min, according to the following reaction:
■
RESULTS AND DISCUSSION
Characterization of the Limestone Samples. Table 1 shows the concentrations of the major and minor components of the CaC, CaB, CaP, and CaR limestones, the Ca content within a range of 39 to 40%. The minor elements, such as Mg and Al were detected at concentrations lower than 0.3%, whereas Si reached values between 0.7 and 0.9% in CaC, CaP, and CaB. Mg is usually present as dolomite (CaMg(CO3)2) and Si as quartz (SiO 2). Iron (Fe) was detected in 2975
dx.doi.org/10.1021/es304090e | Environ. Sci. Technol. 2013, 47, 2974−2981
Environmental Science & Technology
Article
Table 1. Chemical Composition of the Limestone Samples (d.b) Ca (%) Mg (%) Al (%) Si (%) P (%) Na (%) K (%) Fe (ppm) Sr (ppm) Mn (ppm) Zn (ppm) Pb (ppm) Co (ppm) Cu (ppm) As (ppm) Cr (ppm) Sn (ppm) Ni (ppm) Sb (ppm) Hg (ppm)
CaC
CaP
CaB
CaR
CaC-45
CaP-45
39.1 0.27 0.09 0.94
