Southern Research Institute, Birmingham, Alabama
Subcontractor- ARCADIS Geraghty & Miller

The overall goal of this project is to test the effectiveness of calcium-based sorbents and oxidizing agents for controlling mercury emissions from coal-fired power plant boilers.

ARCADIS Geraghty & Miller, with EPA support, has developed calcium-based sorbents to remove SO₂ and mercury simultaneously. The sorbents consist of hydrated lime (Ca(OH)₂) and an added oxidant and a silica-modified calcium (CaSiO₃) with an added oxidant. The mercury capacity in ug Hg/g sorbent for the two sorbents is 20 and 110-150, respectively, verses a mercury capacity for the current standard sorbent, activated carbon, of 70-100. The advantages of a lime based sorbent verses carbon is lower cost, simultaneous removal of sulfur, and allowance of ash to be utilized for a cement additive.

Southern Research Institute is testing the calcium-based sorbents in the Combustion Research Facility in Birmington, Alabama (jointly owned by Southern Research Institute and Southern Company Services). The testing is being done in a pilot-scale six million Btu/hour furnace equivalent to 1.75 MW thermal or about 0.6 MW electrical. The facility consists of a coal crushing and milling area, a coal feeding system, a vertical refractory-lined furnace, a single up-fired burner, a horizontal convective pass with air-cooled tube banks, a series of heat exchangers, an electrostatic precipitator, a pulse jet baghouse, and a packed column scrubber. The pilot system also has a fluidized bed Rapid Absorption Process (RAP) reactor in which solids can be recycled from the downstream fabric filter. For this project, the facility was modified to accommodate feeding and metering equipment for the sorbents and a sampling system to accommodate a continuous mercury monitor. The mercury monitor used is a PSA Sir Galahad model, which provides real-time measurements of speciated (elemental and total) mercury concentration. Sorbent effectiveness is being evaluated with a variety of coal types, operating conditions, and emission control system options for particulate control and NOx reduction.

The oxidant enhanced lime and silica-lime sorbents were tested in February, 2002. A Choctaw America high volatile bituminous type A coal was used at a rate of 115 kg/hr (260 lb/hr) with 15% overfire air and 3% furnace exit oxygen, resulting in a flue gas flow rate of approximately 1150 SCFM. The sorbent was duct injected at rates of 5.4 and 6.8 kg/hr (12 and 15 lbs/hr) into a vertical u-section of flue duct with two seconds of residence time before the sorbent was collected on the bags of the fabric filter. The sorbent injection rate was chosen to obtain a stoichiometry of 1.4, calcium in the sorbent to sulfur in flue gas ratio. The actual ratios of calcium to sulfur for the lime based sorbent was ~1.5, and that for the silica-lime based sorbent was closer to 1.0. ARCADIS had difficulty in making a sufficient quantity of the silica-lime sorbent for this test, so it was necessary to feed the sorbent at a lower feedrate in order to complete the test. Both sorbents contained ~5% oxidant by weight. Hence, the oxidant concentration was in vast excess of the mercury concentration in the flue gas. The sorbent injection temperature was about 150 degrees C (300 F) followed by water injection two feet downstream of the injection port that cooled the sorbent and flue gas down to 90 C (195 F) within 10 feet of the injection port. Flue gas temperature at the baghouse outlet was between 77 and 88 C (170 and 190 F).

Three days of testing were done, with mercury measurements performed by both the PSA monitor and the Ontario Hydro method each day. The first day was coal only, the second hydrated lime with oxidant, and the third silica-lime with oxidant. The PSA mercury monitor measured mercury concentrations upstream of the sorbent injection location and downstream of the baghouse. The Ontario Hydro method was used downstream of the baghouse. Results are in the following table.

The oxidant did not decrease the amount of elemental mercury, in fact it increased it relative to what was already occurring with the ash in the baghouse. The total mercury was decreased somewhat. The sorbent also removed SO₂ from the flue gas. The best results were with the silica-lime+oxidant, which decreased the SO₂ concentration from 750 ppm to 200-300 ppm. The sorbents failure to oxidize elemental mercury may have been due to the temperature of injection, which was higher than what it was tested on the bench scale, or the presence of NOx, CO, or flyash, which was not present in the bench scale experiments. Further bench scale tests are being conducted to elucidate why the ARCADIS sorbents did not work as expected.

Another series of tests was conducted in April with hydrated lime with no added oxidant using the same coal as that used for the ARCADIS sorbents. Lime was injected at both 1150 C and 710 C. The injection temperature did not effect the ability of the lime to capture mercury, however, at the higher injection temperature, SO₂ concentration was reduced from 750 to ~350 ppm. No SO₂ was captured at the lower injection temperature. Water injection is necessary to capture SO₂ with hydrated lime at lower temperatures, as was done with the ARCADIS sorbents. All mercury measurements were done with the PSA continuous emissions monitor. The following table shows the results from this test.

The biggest difference between this test and the one with the ARCADIS sorbents was the baghouse temperature. In the ARCADIS tests, water was injected which cooled the baghouse to 77-88 C, and in the lime only test the baghouse was at 132 C.

Benchscale Results
Since the oxidant did not perform as expected, another series of bench-scale results was done to determine the reason. The simulated flue gas was modified from that originally used to include NO and CO. Ash from the coal that was tested at SRI was also added, since it is known to enhance mercury oxidation. Two temperatures were tested, 80 C and 140 C. The results showed that the NO severely degraded the performance of the sorbent and sorbent performance was better at the higher temperature. This is in contrast to the expected behavior of mercury on sorbents, which typically adsorbs better at lower temperature due to a condensation reaction on the surface. For this reason, the flue gas is typically cooled with water injection to enhance adsorption, which was done on the pilot scale tests of the Arcadis sorbents. The reason for the enhanced performance at higher temperature is probably due to more oxidation occurring on the ash at the elevated temperature and oxidized mercury adsorbs better than elemental mercury. Further tests will not be done with the Arcadis proprietary sorbent. Other calcium-based sorbents such as lime will be used with alternative oxidants such as injected chlorine compounds, sodium tetra-sulfide, or high iron oxide fly ash from the bituminuous coal that showed good oxidation. These oxidants will be tested with a lime sorbent on a subbituminous Powder River Basin coal that characteristically has a high fraction of elemental mercury.