The Technology Identification Subgroup used a multi-step process to accomplish feasibility testing of promising mercury removal pretreatment technologies as follows:

- Identify prospective vendors of mercury removal technologies.

- Make initial vendor contact to solicit interest.

- Review technical summaries submitted by interested vendors.

- Select vendors to invite for interviews.

- Conduct vendor interviews.

- Invite vendors to perform bench-scale feasibility tests.

- Develop a feasibility test protocol.

- Develop a QA/QC protocol and independently analyze raw wastewater and treated wastewater samples.

- Have participating vendors perform bench-scale feasibility tests using actual wastewater samples.

- Review feasibility test reports and develop questions.

- Receive revised feasibility test reports.

- Prepare the final Subgroup report.

The Subgroup prepared a mailing list for the above twelve vendors and sent each vendor a letter to solicit their interest in participating in the Bench-scale Feasibility Testing Project. Copies of the vendor mailing list and the letter appear in Appendix C. The letter summarized the objectives of the project and requested that interested vendors submit a letter expressing their interest in participating in the Project along with a technical summary of their technology (e.g., description of the principle of operation, the species of mercury removed, analytical data from any laboratory or field tests).

Of the twelve vendors contacted, the following seven vendors elected to be interviewed by the Subgroup: ATA Technologies Corporation, Barnebey & Sutcliffe Corporation, ICET, Inc., KDF Fluid Treatment, Inc., SolmeteX, Inc., U.S. Filter/Memtek Division, and the B.G. Wickberg Company, Inc. Since Soils N.V. is a Belgium-based firm, it was not interviewed. Of the seven vendors interviewed, five elected to participate in the testing project. Soils N.V. also elected to participate in the project making a total of six vendors that would conduct bench-scale feasibility tests of their mercury removal technologies.

Vendor Interviews

Vendor interviews were completed in December 1996 and in January and April 1997. Each vendor was given a schedule of the proposed interview dates, a statement of goals and objectives for the interview, and an agenda for the interview. Refer to Appendix D for a copy of the Agenda for Vendor Interviews. The vendors were asked to limit their presentations to 45 minutes. They were also asked to be prepared to answer technical questions during a 15 minute question-and-answer period that would follow their presentation. The following are summaries of the interviews:

Aero-Terra-Aqua (ATA) Technologies Corporation

ATA’s technology is a chemically enhanced sorbent marketed as AQUA-FIX ^TM. The adsorbent is designed to remove dissolved and ionic forms of metals. The removal process consists of three separate sorption steps:

1. adsorption of the metal ion onto the adsorbent’s high surface area,

2. ionic interaction by absorption to polar sites, and

3. chemical bonding of the ions through chelation and ion exchange.

The combination of these processes allows for rapid and substantial metal removal. The adsorbent has been effective in the presence of surfactants and strong chelating agents. The AQUA-FIX ^TM beads can be regenerated by rinsing with a dilute mineral acid. The use of filtering devices (such as particulate filters and activated carbon columns) to remove solids and organic compounds before the adsorbent bed has significantly extended the bed life between backwashings and regenerations.

From ATA’s experience in applications of their adsorbent technology, the following was found to optimize performance:

1. Increasing retention times (contact) improves metal removal kinetics,

2. Reduction of competing ions and enhancement of metal uptake can be achieved by a two-stage serially-operated AQUA-FIX ^TM system,

3. Reduction of organic mercury concentrations with carbon pretreatment can enhance mercury reduction,

4. Frequent column backwashing may be necessary for high solids-bearing wastewater streams, if prefiltration is not done.

5. The optimum pH range for the adsorbent is 5.0 - 8.0 S.U.

Barnebey & Sutcliffe Corporation

Barnebey & Sutcliffe Corporation (B & SC) stated that it is one of only four activated carbon manufacturers in the U.S. and that it makes 80 percent of its products from coconut shells and 20 percent from coal. While not as hard as coconut shell-based carbon, coal-based carbon offers a greater number of macro-pores for adsorption of higher molecular weight compounds such as pesticides. The coal-based carbon also serves as a good substrate for sulfur impregnation at high temperature with a sulfur loading of about 18 percent by weight for optimum performance. In contrast to other sulfur-impregnation methods, the B & SC process does not use any solvent. The B & SC sulfur-impregnated product is called "CB-II."

Several points discussed during B&SC’s presentation included the potential for bacteriological fouling of activated carbon and a method to maximize bed loading of mercury. Bacterial growth can occur during periods of no wastewater flow through the bed of activated carbon. The means to prevent biofouling is simply to continue to recirculate wastewater through the system during any extended system "off" period.

To achieve maximum use of the media, two carbon columns are often piped in series. This configuration offers a "roughing" column followed by a "polishing" column. The first column is used until its entire bed reaches full saturation with contaminant, when the column is removed and replaced. The second column removes contaminants escaping the first column (called contaminant breakthrough).

Chlorine and organics can compete with mercury for attachment to adsorption sites in the carbon. B & SC recommended that a full-scale system for mercury removal include an additional column containing standard activated carbon. This third column would be located before the two CB-II columns and would increase the use of the CB-II medium for mercury removal.

To achieve low effluent mercury concentrations, hydraulic loading of a bed of CB-II is typically 3 to 5 gpm per square foot, with a superficial residence time of 12 to 15 minutes. Periodic backwashing of the bed is recommended which may create particulate mercury concerns for the backwash wastewater. Therefore, the backwash would need to be directed back to the main treatment system holding tank for further treatment. Interestingly, it was stated that the sulfur-impregnated carbon performs best at more alkaline pH levels.

ICET, Inc.

ICET is a young research company. ICET has been developing sorbent technologies and is ready to market its first products. The company’s expertise is in surface modification of activated carbon and sand to achieve sorbents capable of removing a wide range of heavy metals from wastewater. One company product suggested as a pretreatment step for heavy metals removal prior to mercury removal is a hydroxyapatite-coated sand.

Features of ICET hydroxyapatite-coated sand include:

1. increased surface area, coatings absorb up to 40-60% of its weight in metals,

2. high selectivity and efficiency, dependent on the form used,

3. no requirement for set up or conditioning of the sand bed,

4. easy removal of the metal from the spent medium, and

5. the hydroxyapatite coating is renewable in situ allowing the reuse of the sand medium.

Features of ICET activated carbon-based sorbents include:

1. high mercury adsorption capacity,

2. flexibility with respect to mercury levels and flow rates,

3. operation at alkaline or neutral pH,

4. ambient temperature operation, and

5. the ability to recover and recycle mercury.

The potential for wastewater matrix interferences because of the variety of organics, trace metals and viable bacterial organisms present in hospital wastestreams was discussed. ICET recognized that to design a successful pretreatment system for hospital wastewater streams, these factors would need to be considered.

ICET proposed to set up an ambitious bench-scale test system that would continuously pump wastewater through a test system at controlled flow rates. Two flow rates would be examined 1,000 cc/min (120 bed volumes per hour) and 500 cc/min (60 bed volumes per hour). The test system would consist of a prefilter (to remove organic and inorganic particulate), a bed of activated carbon (for organics and additional particulate removal), an ICET media bed (selected for heavy metal removal), and finally an ICET media bed (selected for final mercury adsorption). The wastestream would be sampled after each step of filtration/adsorption and the final product would be collected in 100 ml to 200 ml aliquots at predetermined intervals for analytical testing. The question of funding this test program was not resolved at the time of the interview.

KDF Fluid Treatment, Inc.

KDF’s product, a proprietary alloy of copper and zinc called KDF55, creates a galvanic cell when exposed to water. The galvanic reaction is a standard copper-zinc cell and operates because the two chemically dissimilar metals are in direct electrical contact with water. Through this galvanic cell, removal of mercury ions occurs as a metal replacement process, with a copper-mercury amalgam being created on the KDF55 surface. Zinc ions are replaced by the mercury and are released into the effluent stream during the formation of the copper-mercury amalgam. Effluent zinc levels would have to be considered for most facilities in the MWRA service area, since the MWRA has a zinc discharge limit of 1.0 mg/L.

As presented, KDF’s media seems to work well with ionic forms of mercury. In a study done by the New Jersey Department of Environmental Protection (NJDEP) on contaminated groundwater, the effluent standard was 2 µg/L. Speciation tests of the contaminated groundwater showed that the mercury was 92 percent inorganic, probably in the chloride form, and 8 percent organic (e.g., methyl mercury) with a total mercury concentration of 10 to 30 µg/L. During the study, treatment by the KDF media produced an effluent of 0.5 µg/L mercury. KDF media was then selected by the NJDEP for full-scale systems. After five years, 200 systems are in operation at about 300 gallons per day each without any removal or replacement of the original beds of media.

SolmeteX, Inc.

SolmeteX has developed a medium over the last two years called Keyle:X^TM that is highly selective for ionic forms of mercury. The medium applies the technology of selective chromatography (borrowed from the biotech industry). Since the medium offers much higher selectivity and concentration (loading) factors than other adsorbents, such as ion exchange resins, mercury recovery from the spent medium is possible. Typical saturation loading of Keyle:X^TM media is claimed to be 38 to 45 percent mercury by weight. Because of the higher selectivity of the media, the physical size of the Solmetex system is smaller than for other adsorbents. The smaller sizes of the Solmetex systems provide the opportunity for "point-of-use" systems. "Point-of-use" systems can be part of a larger strategy to prevent mercury contamination from reaching large volumes of wastewater.

Saturated Keyle:X^TM medium can be distinguished by its change in color from yellow to black as the saturation front moves down the column. SolmeteX manufactures its cartridges in clear PVC so that the color change can be easily observed. A user would send cartridges of fully saturated medium to a reclaimer where they would be burned for recovery of the mercury.

SolmeteX recently found that proper mercury speciation was critical to the success of Keyle:X^TM in removing mercury from medical incinerator scrubber blowdown. As a pretreatment step, SolmeteX now does oxidation with hypochlorite solution at a pH of 6.5 to 6.7 to assure that the mercury is converted to soluble ions that can be removed by the medium. Hypochlorite dosing is enough to yield 1-2 milligrams per liter (mg/L) of residual chlorine. Chlorine demand is highly variable depending upon the waste stream. SolmeteX is considering using peroxide or chlorine dioxide (which can be electrolytically generated on-site) as the oxidizing agent. The Keyle:X^TM media is not adversely affected by oxidizing agents.

Since other heavy metals are not removed to any great degree, heavy metals removal (if needed) would have to be done as a separate pretreatment step in an upstream column filled with a different medium. Heavy loadings of particulate or oil and grease would require upstream removal.

SolmeteX is currently operating a test system on the wastewater stream from a fume scrubber at a medical waste incinerator. The incinerator and scrubber operate one day per week, during which the flow rate through the SolmeteX system is set at 1.3 gallons per minute. For this system, SolmeteX uses two-stage cartridge prefiltration: 10 microns nominal followed by 1 micron absolute. Then, two cartridges of the Keyle:X^TM medium are used in series. The flow through each Keyle:X^TM cartridge is one bed volume per minute. Each medium cartridge has a life of about six months, with the second cartridge replacing the first cartridge every three months as a new second cartridge is installed.

With its initial use of hypochlorite at this site, SolmeteX observed a pronounced but temporary increase in the influent mercury concentration. The increase could have been caused by a release of mercury from particulate matter that had adhered to piping surfaces or settled within low points of the system. In five recent runs, two effluent samples had non-detectable mercury (< 0.2 ug/L (ppb)) and all effluent samples had <1.0 ug/L (ppb) of mercury.

U.S. Filter/Memtek Division

In an installation at a battery manufacturing plant, a Memtek system for a wastewater stream containing 20 to 30 µg/L of mercury produces an effluent at about 0.2 µg/L mercury. Memtek has also conducted pilot test studies on scrubber wastewater generated by coal-fired power generating facilities. Mercury was a targeted metal in this wastewater stream. Memtek’s conclusion in these studies was that their chemistry and microfiltration system could reduce mercury and other trace metal contaminants to target levels.

It was proposed that the filtrate water could be polished with an ion exchange resin column to remove any residual mercury not precipitated or removed in the membrane microfiltration system. The physical configuration of the complete system can be designed to fit existing space limitations without compromising system efficiency. Both the ion exchange bed regenerant liquid and the sludge cake would have to be disposed of as regulated wastes.

B.G. Wickberg Company, Inc.

This company markets systems using Mersorb^TM, a sulfur-impregnated activated carbon. Mercury reacts with the sulfur to form mercuric sulfide that is quite stable and insoluble. The activated carbon material, when saturated with mercuric sulfide, may be disposed as regulated waste (if applicable) or sent to a refinery for mercury recovery. The company claims that the spent carbon will pass the hazardous waste test known as the TCLP test, allowing disposal as a federally unregulated waste. Also, the company stated that sulfur-impregnated activated carbon will not release adsorbed mercury if subjected to temperature or pH changes.

The B.G. Wickberg Company has experience using this product on scrubber system wastewater streams from medical waste incinerators and laboratory wastes. Organic, elemental and ionic forms of mercury are easily removed, but complexed mercury removal has been dependent on the stability of the mercury complex present. In incinerator wastewater streams, mercury has a great tendency to bind to particulate matter. Particulate filtration in combination with the adsorbent was found to reduce mercury concentrations in the incinerator wastewater streams effectively.

Selection of Test Wastewater

The Technology Identification Subgroup realized that there were significant limits on both resources and time for the feasibility testing project. Since the vendors would be asked to conduct all test work without charge, the Technology Identification Subgroup decided that only one type of wastewater would be used in the project.

The Technology Identification Subgroup reviewed the WWC Subgroup sampling program results for the five types of facilities studied by the WWC Subgroup (incinerators, power plants, hospital clinical laboratories and hospital research laboratories). The review suggested that the largest mercury concentrations were from clinical and research laboratories. The clinical laboratory used for this study showed parameter concentrations that were equal to the overall average of the research laboratories parameter concentrations. Both the clinical and the research laboratories showed identical parameters, except that the parameter concentrations for the research laboratories were more variable. The Subgroup decided, therefore, that only the clinical laboratory wastewater would be used for the testing project.

A local hospital agreed to provide samples of their clinical laboratory wastewater for the project. At this facility, clinical laboratory wastewater is currently collected into holding tanks for offsite disposal. Since the wastewater is collected over a period of several days, the holding tanks served to produce composites of the wastewater. Moreover, since the sampling effort for the feasibility testing project involved collection of five gallon samples, the holding tanks also simplified the collection process. Most important, the wastewater had a fairly consistent mercury concentration between 11 and 90 µg/L (ppb).

Analytical and Mercury Speciation Testing

To verify that an adequate mercury level was present and to partially characterize the specific clinical laboratory wastewater, the Technology Identification Subgroup decided to perform analytical testing of raw wastewater samples collected for the Bench-scale Feasibility Testing Project. Representative samples were tested for total mercury and Priority Pollutant Metals.

Total mercury concentration in the wastewater samples was determined by the MWRA Central Laboratory using EPA Method 245.1. This EPA spectrophotometric method is the analytical method of choice because most federal and state regulations address total mercury concentrations in water and wastewater. The method detection limit for the Laboratory was 0.05 µg/L (ppb). The Priority Pollutant Metals analyses were done to help the participating vendors to determine whether any metals were present at levels that could interfere with their mercury removal processes.

As mentioned earlier regarding species of mercury in wastewater, some mercury removal technologies have been fully effective for only specific species of mercury. Therefore, mercury speciation testing of wastewater samples can provide valuable insight into the various mercury species that may be present in a wastewater proposed for pretreatment. To simplify the process of mercury speciation testing for the project, the Subgroup decided to determine only the amount of particulate mercury in representative samples of the clinical laboratory wastewater.

The raw clinical laboratory wastewater samples intended for analytical testing were collected at the same time that five gallon test samples were collected for overnight shipment to the participating vendors. All the raw wastewater analytical tests were done by the MWRA Central Laboratory on a two-day turnaround basis so that the resulting data could be sent to participating vendors before the start of their bench-scale tests.

Feasibility Testing and QA/QC Protocols

The clinical laboratory wastewater sample collections began on February 21, 1997. The last sample collection occurred on June 13, 1997. The sample collections were made by experienced MWRA Sampling Associates. Five gallon sample containers were packed in ice-filled coolers for overnight shipment to each participating vendor. Each vendor had an opportunity to specify the desired number of five gallon sample containers for its bench-scale tests. The vendors were asked to handle the samples and conduct the bench-scale feasibility tests according to a detailed written protocol. A copy of this document, entitled "Scope of Work, Feasibility Testing" appears in Appendix E.

In an attempt to put all participating vendors on an equal level, the Scope of Work required that several quality control and quality assurance (QA/QC) measures be employed during the testing process from sample collection to the final reporting of data. The Subgroup selected these measures to ensure the integrity, reliability and reproducibility of the test data.

The following is a summary of the QA/QC measures specified in the Scope of Work and some basic reasoning behind the QA/QC goal:

• Sample containers for analytical testing were provided by the Subgroup and were pre-rinsed with nitric acid to ensure that no initial mercury contamination was present. To ensure that mercury contamination was neither initially present nor introduced during test work, the participating vendors were required to complete a mercury analysis of their high purity deionized water (20 samples) and to provide three procedural blanks (i.e., samples of high purity water were carried through the same handling procedure and process as actual test samples). The deionized water samples and procedural blanks were analyzed for mercury contamination by the MWRA Central Laboratory.

• To ensure the integrity of test samples, and to verify claimed mercury and other metals reductions, the vendors were asked to supply split samples of wastewater through each stage of the treatment process to their laboratory and to the Subgroup for analysis by the MWRA Central Laboratory.

• As a final measure of consistency, the vendors were asked to provide a detailed report on their findings in a standard format consisting of:

- an introduction

- test materials, procedures, and experimental protocols

- pretreatment considerations

- test results

- full scale considerations (cost estimates, space requirements)

- discussion / conclusions

- appendices (including analytical test reports).

As outlined above, the vendor-submitted samples of deionized water and procedural blanks were analyzed for mercury contamination by the MWRA Central Laboratory. The analyses showed that test samples were free of initial mercury contamination and also that mercury contamination was not introduced during the feasibility test work of the vendors. Analytical test results of the submitted deionized water samples and procedural blanks are available from the MWRA upon request.

For vendor-submitted split samples of treated wastewater, mercury analyses by the MWRA Central Laboratory served to verify nearly all corresponding vendor analyses. Refer to Appendix A for tables that were developed to summarize and compare MWRA Central Laboratory analytical test data and vendor analytical test data.

For Soils N.V., however, the tables show that there were differences in analytical test results for the submitted samples of both raw and treated wastewater. For example, before Soils N.V. began bench-scale testing, it took five small samples from the five gallon raw wastewater test sample and found an average mercury concentration of 13.6 µg/L (ppb). In contrast, the MWRA Central Laboratory found a higher mercury concentration of 24.9 µg/L (ppb) in a raw wastewater sample collected at the same time as the test sample.

In its feasibility testing report, Soils N.V. claimed to follow the requirements of EPA Method 245.1 during their analytical work. They attributed the difference in mercury analytical results to the difficulty in taking a representative sample from the five gallon test sample container because of heavy particulate in the test sample and to the high fraction of mercury likely held by the particulate.

They were not aware, however, of differences in analytical test results of mercury concentrations for the split samples of treated wastewater that they had submitted to the MWRA Central Laboratory. In contrast to the higher mercury concentration measured in the raw wastewater before it was shipped to Soils N.V. in Belgium, the MWRA Central Laboratory found lower mercury concentrations than did the vendor for the split samples of treated wastewater shipped to the United States.

We believe that the specific character of the "before and after" differences suggest that mercury may have been lost from the raw and treated wastewater sample containers by means of evaporation during the lengthy periods of reduced atmospheric pressure for both the East-bound and West-bound trans-Atlantic flights. As a result, because the Soils N.V. analytical tests were done on samples that were not subject to overseas shipment, we have used the feasibility test data and removals performance values of Soils N.V. in this Report. Refer to the summary and comparison tables of Appendix A for further details.

Feasibility Testing Project Results

Refer to the following Table 2 for an overall summary of the results from the Bench-scale Feasibility Testing Project. The results suggest, for samples of one clinical laboratory wastewater stream, that five different pretreatment technologies showed test mercury removal efficiencies varying from approximately 44 percent to 99.7 percent, with some final test mercury concentrations at very low µg/L (ppb) levels. Moreover, for certain test runs on the clinical laboratory wastewater, some technologies appeared to achieve the feasibility test goal of 1.0 µg/L (ppb) effluent mercury.

As mentioned above, the Technology Identification Subgroup developed tables to summarize MWRA Central Laboratory analytical test data and vendor analytical test data for the bench-scale feasibility tests of each participating vendor. The summary tables are provided in Appendix A. For copies of individual vendor reports on their bench-scale feasibility tests, refer to Appendix F.

BENCH-SCALE FEASIBILITY TESTING PROJECT