FINAL TECHNICAL REPORT

September 1, 1996, through August 31, 1997

Project Title: MERCURY ADSORPTION IN SIMULATED FLUE GAS OF COAL- FIRED BOILERS BY OXIDATION

ICCI Project Number: 96-1/2.4A-3

Principal Investigator: G. A. Kudlac, McDermott Technology, Inc. Project Manager: Ron Carty, ICCI

ABSTRACT

It has been reported that a large portion of mercury in coal cannot be removed by conventional flue gas cleanup equipment. However, oxidized mercury can be captured more easily by sorbents than elemental mercury. The objective of this project is to convert elemental mercury to oxidized mercury by using oxidants prior to sorption.

Equilibrium calculations of mercury forms based on the thermodynamic data were performed using the STANJAN computer program. Thermodynamic data of common mercury species was gathered and data files were created. Properties of an Illinois coal were used in the equilibrium calculations of coal combustion. The reactants and their concentrations, and conditions of combustion were specified in the calculations. Six oxidants were preselected and their thermodynamic data was incorporated in the calculations. The specific coal combustion products, including mercury species and their concentrations, were determined from the equilibrium calculations. Based on these calculations, promising oxidants for mercury oxidation were selected for detailed bench-scale mercury adsorption tests. Equipment required for upgrading the existing bench-scale adsorption facility were procured. Two activated carbon samples were selected for sorption tests. Base case (without any oxidants) and preliminary tests for mercury adsorption using the bench-scale adsorption facility were initiated. Several preliminary mercury oxidation tests, including a base case (no sorbent or oxidant) and oxidant screening tests (oxidant present, no sorbent) were completed and the Ontario Hydro scrubbing solutions were submitted for mercury concentration analysis.

All parametric oxidation/adsorption tests using the preselected oxidants and two activated carbon samples have been completed. Testing was conducted over a range of oxidation reactor temperatures, test duration times and oxidant type. Both nitrogen and nitrogen with approximately 1500 ppm SO2 present were used as the carrier gas. All activated carbon samples were analyzed for mercury content using Cold Vapor Atomic Absorption Spectroscopy (CVAAS).

With the exception of hydrogen peroxide and chlorine, little oxidation activity was observed with the evaluated oxidants. Under the most favorable conditions (hydrogen peroxide and 260 C), approximately 20% of the elemental mercury entering the reactor was oxidized and adsorbed, compared with a baseline level of 7% at the same conditions.

EXECUTIVE SUMMARY

Mercury is found to be the most volatile trace metal in coal that cannot be effectively removed by conventional flue gas cleanup systems. However, it has been reported that oxidized mercury showed much higher removal potential than elemental mercury. The objective of the project was to convert elemental mercury to oxidized mercury using oxidants to improve the removal efficiency of total mercury.

The thermodynamic data of the major components of coal, combustion gases, common mercury species, and flue gas was gathered and data files were created using Stanford University's STANJAN computer software. The properties of an Illinois coal (IBC-106) were used in the equilibrium calculations. Theoretical calculations of mercury species under coal-fired flue gas conditions using the STANJAN program were performed. Six oxidants were preselected and included in the equilibrium calculations. The preselected oxidants included chlorine, ozone, hydrogen peroxide, calcium oxide, iron oxide (Fe2O3), and sodium perchlorate. Chlorine was found to be the most reactive oxidant. The other five oxidants showed very little effect on mercury oxidation. The existing bench-scale adsorption facility was upgraded to include a furnace and a reactor chamber.

Two base case tests (with no oxidant or sorbent) using the bench-scale adsorption facility were performed with standard Ontario Hydro impingers downstream of the oxidation reactor. The Ontario Hydro impinger solutions were analyzed to determine the actual mercury concentration in the gas stream at the oxidation reactor inlet for comparison with the stated permeation rate for the elemental mercury permeation tube in an effort to account for potential dropout. Mercury adsorption percentages presented in this report are based on the actual mercury concentration determined from the impinger solution analysis.

Two activated carbon samples were used as the sorbents for the mercury adsorption tests. Removal of mercury using oxidants was carried out under different test conditions. The amount of mercury captured by the sorbents was determined using Cold Vapor Atomic Absorption Spectroscopy (CVAAS). Due to potentially wide variations in mercury concentration in a given carbon sample, the total carbon sample for each test was divided and duplicate analyses performed to yield a total mercury concentration for a given test.

Based on reported results from various field tests investigating the use of various additives as a means of improving mercury control, the oxidants chosen for the bench-scale phase of the project were modified and include: sulfur dioxide (SO2), hydrogen peroxide (H2O2), hydrochloric acid (HCl), chlorine (Cl2), and hydrogen sulfide (H2S). All parametric oxidation/adsorption tests using the preselected oxidants and both of the activated carbon samples have been completed. Testing was conducted over a range of oxidation reactor temperatures, test duration times and oxidant type. Both nitrogen and nitrogen with approximately 1500 ppm SO2 present were used as the carrier gas. Over the range of reactor temperatures tested, hydrogen peroxide was found to be the most active oxidant, followed by chlorine. For the most active conditions (H2O2 as the oxidant, 260 C reactor temperature), approximately 20% of the inlet mercury was oxidized and adsorbed, compared with approximately 7% at the same operating conditions with no oxidant present. The other oxidants (SO2, H2S, and HCl) showed minimal oxidation activity.

Testing was also conducted in an effort to determine the effect of sorbent type on mercury adsorption. Two Calgon activated carbons, one a standard type and one modified with potassium iodide, were used for the comparison testing. Due to unresolved operational and/or analytical problems with the sorbent comparison samples, however, no definitive comparison between the two activated carbons was possible.

McDermott Technology, Inc. assumes no liability with respect to the use of, or for damages resulting from the use of, or makes any warranty or representation regarding any information, apparatus method, or process disclosed in this report.

McDermott Technology, Inc. expressly excludes any and all warranties either expressed or implied, which might arise under law or custom or trade, including without limitation, warranties of merchantability and of fitness for specified or intended purpose.