Daryl L. Roberts and Tom E. Broderick
ADA Technologies
Dennis Sparger
Sandia National Laboratory
ABSTRACT
ADA Technologies is developing a sorbent-based process that recovers mercury from flue gases, a process that regenerates the sorbent and recovers the mercury in a form suitable for further distillation and ultimate recycle. Because of these attributes of the process, ADA Technologies has adopted the name "Mercu-RE" to describe its process.
In the laboratory, the uptake of mercury and mercuric chloride has been tested with sorbents purchased commercially and prepared by ADA Technologies. In addition, the sorbents have been imbibed into the body of ceramic filters to simultaneously remove mercury vapors and particulate matter from flue gases. In these tests, the mercury uptake efficiency has varied between 65% and 99%. The efficiency of uptake was unaffected by the presence of eight volume percent water vapor and HC1 concentrations of 50, 100, and 200 ppmv. In our early laboratory work, selected sorbents and filters were regenerated thermally and put through five cycles of sorption and regeneration. No loss of performance was observed through these cycles.
Proof-of-concept field testing to date has included a two actual cubic foot per minute (ACFM) unit at IT Corporation (Oak Ridge, TN) and a 50 ACFM unit at Science Applications International Corporation in Idaho Falls, ID. In these tests, the process removed over 99% of the mercury. The sorbent was not regenerated in the two ACFM test but was regenerated three times during the 50 ACFM test (one week duration).
Current work is emphasizing the long-term durability of the sorbent. Fundamental studies of the sorbent have shown that on some sorbents the noble metal crystallites that comprise the active ingredient are thermally stable after 2160 hours of exposure to the regeneration temperature of 700°F. This time period represents about 540 days of commercial operation (four hours of regeneration per day of operation). These thermal stability tests are continuing, and long-term testing in the presence of aggressive flue gas components is planned for FY '97.
INTRODUCTION
Mercury contamination is problematic in many DOE site remediation and waste disposal efforts (Kvartek et al., 1994; Perona and Brown, 1993; Taylor, et al., 1995). The most visible of these efforts is that of the Oak Ridge Y-12 plant from which approximately 350 metric tons of mercury were discharged to the environment between 1953 and 1963 (Turner, et al., 1984). These discharges have led to a wide-ranging variety of soils and sediments that need remediation. In addition, DOE has identified 38,000 m3 of mixed low-level and transuranic wastes in its waste inventory that contain radioactive mercury contamination, including some radioactive elemental mercury.
Retorting and vitrification are two leading options for treating mercury-containing soils and wastes. These processes create a mercury-containing off-gas that can contain a mercury concentration equal to the vapor pressure of elemental mercury at the flue gas temperature (approximately 3500 ppmv or 20 grams per cubic meter at 300°F). Even the off-gas of a condenser operating at the relatively low temperature of 50°F can have a mercury partial pressure of 5x10-4 torr or 5,800 µg/m3. Although there is no national standard for the concentration of mercury in all exhaust gases, the Environmental Protection Agency regulates municipal waste incinerators to a limit of 85 µg/m3. With respect to workplace or public exposure, the Threshold Limiting Value for mercury is 50 µg/m3, and the Immediately Dangerous to Life and Health (IDLH) limit is 28,000 µg/m3. The IDLH value is the vapor pressure of mercury at approximately 80°F which is why any fugitive liquid mercury in a lab, school, or office building is regarded as an immediate threat. Finally, mercury exists in many chemical forms, some of which are equally volatile to elemental mercury (e.g., the vapor pressure of mercuric chloride at 300°F is 2 torr compared to 2.7 torr for elemental mercury vapor).
Only municipal waste incinerators are governed by national regulations today, and the best available control technology specified for municipal waste incinerators is the injection of activated carbon. Improvements to this method are needed since 100 to 10,000 pounds of activated carbon are required to remove one pound of mercury from the flue gas depending on the concentration and speciation of mercury in the flue gas. The mercury-contaminated carbon becomes part of the ash collected by particle control devices. Recent studies at ADA Technologies and other laboratories indicates that mercury re-volatilizes from carbon, although the rate is slow at room temperature.
ADA Technologies has been developing a mercury control technology that aims to remove mercury from flue gases and recover elemental liquid mercury as a by-product with no solid wastes. The following paper reviews the status of the control technology and gives an example of a field test wherein surrogate, mercury-containing wastes were melted in a plasma hearth pilot unit under the sponsorship of the Department of Energy, Office of Environmental Management.
BASIS OF ADA'S MERCURY CONTROL TECHNOLOGY
The underlying principle of ADA's regenerable mercury removal process (Mercu-RE process) is the sorption of mercury and its common compounds by noble metals and the regeneration of these metals by thermal desorption of the mercury. This principle has been exploited for mercury analysis and environmental sampling (Dumarey, et al., 1985; Slemr, et al., 1979). Simply stated, mercury and its common compounds sorb to the noble metals and are desorbed as elemental mercury vapor at an elevated temperature. In analytical chemistry or environmental sampling work, this regenerability essentially imparts an ability to concentrate the mercury from the low levels found in most samples. It is then easily measured, either gravimetrically or by UV absorption.
In laboratory work to date, aimed at waste incinerator applications, ADA has devised sorbents, essentially gold dispersed on either activated carbon or alumina support particles, that remove up to 99% of elemental mercury vapor and mercuric chloride vapor at 300°F in a synthetic flue gas containing more than 1000 µg/m3 of mercury, 8% water vapor, and either 100 ppm HC1 or 2000 ppm SO2. Further, upon heating to 500°F to 700°F, the mercury is fully desorbed as mercury vapor, even when the mercury is sorbed as mercuric chloride. We have also imbibed the sorbent into the walls of a filter and performed mercury uptake and regeneration experiments. In sorbent beads or in a filter form, the reversible capacity of the gold for mercury exceeds 100 mg of Hg per gram of gold (10 wt.%).
In a typical experiment, approximately four grams of sorbent are held in an oven, and the synthetic flue gas is passed through the bed with a superficial velocity near 0.5 to 1.0 ft/sec. Typical sorbent pressure drops have been in the range of 0.2" to 5" of water. Typical bed residence times have been in the range 0.1 sec to 1.0 sec. The space velocities, a measure of throughput in catalytic technologies, have been in the range of 10,000 hr-1 to 100,000 hr-1.
Representative results are depicted in Figs. 1 through 4. In Fig. 1, a gold-spiked carbon bed sorbs essentially all of the mercury for more than 24 hours before mercury vapor breaks through. In Fig. 2, a gold-spiked alumina sorbent takes up mercury for approximately 40 hours. The breakthrough shoulder is broad; the capture efficiency falls below 90% after 10 hours but stays above 85% for the next 30 hours.
Fig. 1. Carbon-based sorbent removes
all mercury for 24 hours.
Fig. 2.Alumina-based sorbent removes
85% of mercury for 40 hours.
Fig. 3.Repeated regeneration of
alumina-based sorbent.
Fig. 4. Repeated regeneration of
sorbent imbibed into ceramic filter.
In Fig. 3, mercury is sorbed to over 98% efficiency for 60 minutes by a gold-spiked alumina sorbent, and these beads are repeatedly regenerated by heating to 600°F. In Fig. 4, mercury is sorbed to 99% efficiency for 80 minutes by a ceramic filter into which sorbent has been imbibed, and the filter is repeatedly regenerated by heating to 700°F.
FIELD TESTING AT PLASMA HEARTH THERMAL TREATMENT UNIT
The Mercu-RE control technology was tested at a pilot-scale thermal treatment facility being developed by the Department of Energy. This treatment system is intended to reduce the volumes of hazardous wastes currently containerized and stored at DOE facilities. Large amounts of mercury are frequently found in these stored wastes, necessitating mercury control and measurement for the proper disposal of these wastes.
The thermal treatment system consists of a plasma hearth chamber, a secondary combustor, an acid gas scrubber, a baghouse, and a bank of HEAP filters. Figure 5 shows that the mercury concentrations at the inlet to the sorbent beds were in the range of 10,000 µg/m3 to 60,000 µg/m3, and that the inlet concentrations were much higher than the outlet concentrations. Typical mercury removal efficiencies exceeded 99%. Figure 6 shows concentrations of total and elemental mercury leaving the scrubber over a 4-hour period of time. The "spikes" in concentration correlate to the input of wastes to the thermal treatment unit. The difference in mercury readings between the total and elemental measurements is the concentration of speciated mercury (i.e., the concentration of chemically combined mercury.) The measurement of the mercury concentrations were made with ADA's continuous, speciating mercury analyzer. This analyzer measures the concentration of elemental mercury and of all other mercury-containing compounds simultaneously and in real-time, down to concentrations of 0.2 µg/m3 (Roberts, et al., 1996). The analyzer responded rapidly to changes in mercury concentration. The surrogate waste was fed by inserting a can into the hearth approximately every 20 minutes, resulting in a pulse of mercury in the flue gas.
Fig. 5. Sorbent bed removing mercury
from plasma hearth off-gas.
Fig. 6. Expanded view of outlet
concentrations during plasma hearth test.
During the field test, several measurements were made of mercury concentrations using the Mercury Speciation Sampling Train (MSST) and Draft EPA Method 29. These sample trains were used to determine the concentration of mercury in the uncontrolled flue gas. Mass balance calculations (based on analysis of the feed material and system flue gas flow rates) predicted that the average mercury concentration would be over 10,000 µg/m3. To measure these high concentrations with the CEM, a dilution probe was used to reduce the mercury concentrations by approximately two orders of magnitude. The analyzer readings verified the mass balance predictions. The range of concentrations measured using the analyzer (11,300 to 18,100 µg/m3) agree well with the MSST samples (14,100 to 19,000 µg/m3). Results using Draft Method 29 were well below analyzer results, ranging from 3,700 to 8,600 µg/m3. It is suspected that the Method 29 impinger solutions may have been overloaded given the high mercury concentrations in the flue gas.
THERMAL ENDURANCE TESTING
The crystallites of noble metal dispersed on the microporous sorbent support can, in theory, coagulate by surface diffusion and, in time, compromise the performance of the sorbent. For this reason, we have initiated an accelerated thermal endurance test. The test consists of keeping the sorbent at the regeneration temperature (700°F) for an extended period of time and periodically removing samples to determine the average crystallite size. This test is an accelerated test because in commercial application, the sorbent would be at the regeneration temperature only approximately four hours per day. The crystallite size measurement is made with x-ray diffraction, specifically looking at the line broadening that occurs when the crystallites are smaller than about 100 nm (Cullity, 1978).
We have exposed three sorbents to the regeneration temperature for 90 days, representing 540 days of commercial operation. So far, one of the three sorbents is quite stable, one is approaching a stable crystallite size, and the other needs more data to be sure that it is stable (Fig. 7). The thermal stability of the sorbent appears to be sufficient for commercial operation.
Fig. 7. Stability of noble metal
during exposure to high temperature.
FUTURE PLANS
Thermal endurance itself is only the beginning of the life time challenges the sorbent must face. The repeatable uptake of mercury` through 100 cycles in the laboratory is the next test planned. Assuming that there is one or more sorbents that passes this test with less than 10% degradation in performance, we will then test the sorbent for 100 cycles in the field on off-gas from a plasma hearth facility. In this test, the sorbent will be exposed to potentially condensable poisoning agents such as arsenic, selenium, or vanadium. Other aggressive components will be present also such as NOx, SO2, HC1, and water vapor. These tests are planned for Fall, 1997.
BIBLIOGRAPHY
CULLITY, B. D., Elements of X-Ray Diffraction, Addison-Wiley Publishing Company, Inc., 1978.
DUMAREY, R., R. DAMS, J. HOSTE, "Comparison of the collection and Desorption Efficiency of Activated Charcoal, Silver, and Gold for Determination of Vapor-Phase Atmospheric Mercury," Anal. Chem., 57, 2638-43 (1985).
HADEISHI, T., D. A. CHURCH, R. D. MCLAUGHLIN, B. D. ZAK, M. NAKAMURA, B. CHANG, "Mercury Monitor for Ambient Air," Science, 187,348-9 ( 1971 ).
KVARTEK, E. J., W. H. CARLTON, M. DENHAM, L. ELDRIDGE, N. C. NEWMAN, "Assessment of Mercury in the Savannah River Site Environment," WSRC-TR-94-0218ET, September, 1994.
PERONA, J. J., C. H. BROWN, "A Technology Assessment for Mercury-Containing Mixed Wastes," DOE/MWIP-9, Oak Ridge National Laboratory, March, 1993.
ROBERTS, D. L., R. W. MARMARO, R. J. SCHLAGER, "Continuous Emissions Monitor for Total, Elemental, and Total Speciated Mercury," presented at EPRI-DOE-EPA Joint Workshop on Mercury Measurement and Speciation Methods for Utility Flue Gas, Research Triangle Park, NC, January 24-25, 1996.
SLEMR, F., W. SEILER, C. EBERLING, P. ROGGENDORF, "The Determination of Total Gaseous Mercury in Air at Background Levels," Anal. Chim. Acta, 110, 35-47 (1979).
TAYLOR, P. A., K. T. KLASSON, S. L. CORDER, "Mercury Separation from Aqueous Wastes," presented at the Summer National Meeting of the American Institute of Chemical Engineers, July 30-August 2, 1995.
TURNER, R. R., C. R. OLSEN, W. J. WILCOX, "Environmental Fate of Mercury and Cesium-137 Discharged from Oak Ridge Facilities," presented at 18th Annual Conference on Trace Substances in Environmental Health, Columbia, MO, June 4, 1984.