Fate-of-Mercury in Flue Gas Evaluations: the Critical First Step in Reducing Mercury Emissions from Coal-Fired Utility Boilers

The first phase cap is 38 tons annually and the bulk of these reductions is expected be achieved by taking advantage of “co-benefit” – that is, mercury reductions that are achieved as a result of new controls to reduce sulfur dioxide (SO2) and nitrogen oxides (NOX) emissions, as required by USEPA’s Clean Air Interstate Rule (CAIR). In the second phase, due in 2018, emissions from these units will be capped at 15 tons/year

Maximum co-benefit reduction, however, is not necessarily assured simply by installing SO2 and NOX control devices. Fate-of-mercury studies conducted since the late 1990’s have shown a wide disparity in the way Hg behaves across various air pollution control devices (APCDs), and consequently how much Hg is emitted at the stack, even across boiler systems of identical or similar design. In general, this inconsistency is believed to be the result of the complexity of mercury interactions in modern boiler/flue gas systems. Among the numerous interrelated factors that affect the amount of mercury ultimately emitted: the coal type and composition; ash composition; flue gas temperature and chemistry; concentrations of acid gases; retention time in APCDs; scrubber chemistry and boiler system operational conditions.

Effective Hg removal requires a basic understanding of how mercury acts throughout the specific flue gas stream. The initial source of the mercury, of course, is the fuel source. Mercury typically exists as a trace element in coal on the order of 0.1 ppmw, although this concentration varies considerably between coal types. As coal is burned near 1700°K, the mercury volatilizes to form thermodynamically favored gaseous elemental mercury (Hg0). In the subsequent cooling of the combustion gases, interaction with other combustion products results in a portion of the elemental mercury being converted into gaseous oxidized forms of mercury (Hg2+). From a Hg emissions reduction viewpoint, oxidation is beneficial because Hg2+ compounds are generally water-soluble and are, therefore, more effectively captured by wet scrubbers and, to a lesser extent, other pollution control systems. Conversely, Hg0 is difficult to control and is likely to enter the global atmospheric cycle because of its high vapor pressure and low water solubility. A small portion of Hg0 and Hg2+ at the stack adsorbs onto residual particulates (e.g. fly ash), forming particle-bound mercury (HgP). So, effective control with existing APCDs involves either forming oxidized compounds or particle-binding.

These mercury oxidation/adsorption processes are functions of the aforementioned interrelated factors that are quite complex and only partially understood, making predictive total mercury emission determinations difficult. Therefore, the task of obtaining relevant information regarding the specific behavior of mercury in any given boiler flue gas system entails taking a variety of mercury concentration measurements (speciated, particle-bound and total Hg) along the flue gas stream.

Electric utility APCDs and other devices that have been shown to influence the fate-of-mercury in the flue gas stream typically include some or all of the following:

• NOX control by selective catalytic reduction (SCR). The vanadium and titanium oxides (V2O5, TiO2), used commonly in the vanadia-titania SCR catalyst for catalytic NOx reduction, promote the formation of oxidized mercury (Hg2+).
• Air Pre-heater. Air pre-heaters promote conversion of Hg0  Hg2+ via a precipitous flue gas temperature drop (typically  300o F).
• Particulate control by electrostatic precipitators (ESP’s) or fabric filters (baghouses). Hg removal by particle binding can occur as fluegas is passed over the captured particulate, as both Hg0 and Hg2+ are theoretically affected by the unburned carbon in the flyash.
• SO2 control by wet flue gas desulphurization (FGD). Oxidized mercury (Hg2+) is typically water soluble, and thus can theoretically be removed in wet scrubber systems. Effective removal of Hg2+ in scrubber systems has been shown to be a function of the scrubber chemistry.

In a typical fate-of-mercury study, flue gas sampling locations generally include the inlets and outlets of these devices, providing an understanding of the speciation/removal effect of each device. In most instances, the outlet of one device serves as the inlet for the next device. Sample locations are tested simultaneously, affording a definitive “snapshot” of the fate-of-mercury during the time of sampling. In addition, the fuel source, combustion by-products and APCD additives may be analyzed for a rough mass balance Hg calculation. Materials analyzed concurrent with flue gas sampling may include, but are not limited to: coal (or coal blends), fly ash, bottom ash, scrubber liquor, scrubber makeup water, scrubber solids (lime, gypsum), and other flue gas additives (SO3 control, etc.). Obtained Hg concentration results in the flue gas and related combustion by-products/additives can then evaluated with respect to:

1. Desired mercury emission limits (or compliance with the facility’s regulatory mandated mercury emission limits)
2. The effectiveness of existing APCD’s in oxidizing or adsorbing mercury
3. Operational changes that may influence the Hg control effectiveness of APCD’s
4. Economic effects:
   * Selling or purchasing Hg trade allowances
   * Disposal costs of Hg laden combustion by-products
   * Cost/benefit analysis of additional mercury control equipment

Until recently, however, obtaining accurate flue gas samples was a costly, time-consuming and difficult process. Historically, the USEPA essentially recognized only one method for obtaining speciated Hg flue gas samples, namely ASTM D 6784 – 02, commonly referred to as the Ontario Hydro Method (OHM). This method is summarized as follows:

A sample is withdrawn from the flue gas stream isokinetically through a probe/filter system, maintained at 120°C or the flue gas temperature, whichever is greater, followed by a series of impingers in an ice bath. Particle-bound mercury is collected in the front half of the sampling train. Oxidized mercury is collected in impingers containing a chilled aqueous potassium chloride solution. Elemental mercury is collected in subsequent impingers (one impinger containing a chilled aqueous acidic solution of hydrogen peroxide and three impingers containing chilled aqueous acidic solutions of potassium permanganate). Samples are recovered, digested, and then analyzed for mercury using cold-vapor atomic absorption (CVAAS) or fluorescence spectroscopy (CVAFS).

Schematic of Mercury-Sampling Train in the Method 5 Configuration

What may not be obvious from this innocuous method summary is the requirement for extreme field and laboratory QA/QC, as the potential for contamination is significant. Consider:

• The level of Hg in the flue gas being sampled is very low; it can be less than 0.5 g/m3. This concentration is sufficiently low so that an OHM sample could be contaminated by breathing on the collected solutions (amalgam teeth fillings), or if a mercury thermometer has ever been broken at the analytical laboratory.
• All sampling train surfaces (that come into contact with the sampled fluegas), or any of the reagents (that must be made up on-site) present potential for contamination. Essentially, the OHM is a laboratory-quality procedure being conducted in the field (often a very dirty environment).

One of the OHM chemical reagents, potassium permanganate is considered by USEPA as a hazardous material, and it has the potential to off-gas. This condition imposes considerable difficulties for shipping.

In addition, the OHM has an inherent positive bias of particulate-bound Hg. This stems from the fact that, over the typical 2 to 3 hour test run, there is often a considerable build-up of particulates on the filter, which has the tendency to strip off both Hg0 and Hg2+ as the flue gas is pulled over the filter cake. Suffice it to say that the OHM currently is not viewed favorably by the utility industry at large.

Fortunately, there now is a better Hg-in-flue-gas sampling method. Recent advances in speciated mercury sorbent trap technology, pioneered by Frontier Geosciences, Inc., have culminated in the development of the Flue Gas Absorbent Mercury Speciation (FAMS™) method. The FAMS™ method relies on sequential selective capture to separate and quantify Hg0, Hg2+ and HgP. A known, precise volume (±0.1 liter) of gas is pulled through the FAMS™ sorbent train. The temperature of the FAMS™ sorbent train is kept at 95 ± 5° C during sampling to avoid water condensation in the trap.

The FAMS™ method has undergone numerous intercomparison studies with the OHM, including a USEPA and Frontier Geosciences sponsored Performance Based Measurement System (PBMS) validation study conducted in 2001 at the Energy Environmental Research Center (EERC) that led to the following conclusion: “Based on the results of this PBMS, we conclude the FAMS Method is equivalent to the ASTM approved OH Method and a therefore a valid method for the determination of total Hg, HgP, gaseous Hg2+ and Hg0 concentrations in fluegas matrices. Considering many factors, including simplicity, lack of hazardous solutions in the field, precision, sensitivity, accuracy and cost, the FAMS method has many advantages that make it a viable choice for the measurement of fluegas Hg speciation.”

Field sampling techniques also play an important role in obtaining quality sorbent trap mercury measurements. Air Quality Services (AQS) embraced the FAMS™ approach over three years ago, and has significantly enhanced the field sampling tasks by developing specifically engineered sorbent trap sampling equipment. This equipment utilizes a quartz probe liner, proprietary (patent pending) controlled-temperature multiple-sample probe, silica-gel water knockout, mass flow meters with totalizers, and fine adjustment diaphragm pumps. This equipment has proved to be adaptable to sampling locations with temperatures as high as 700oF (SCR inlet) and as low as 120oF (wet stacks), providing precise sampled flue gas volume determinations under controlled temperature conditions. The multiple sample probe design allows for precision comparison of collected flue gas samples, a QA procedure embodied in both 40 CFR 75 Appendix K and USEPA Method 30B.

The beauty of the FAMS™/AQS Hg sampling approach lies in its simplicity. Because the low Hg background traps are prepared under strict QA/QC procedures at the laboratory, there is minimal potential for field contamination. Individually numbered traps are simply uncapped, put on line, leak-tested and inserted into the probe (which is pre-heated and mounted to an air-tight flange at the sampling port). After sampling, the traps are leak-tested again, removed from the sampling train, capped and shipped off to the lab along with appropriate field data sheets and chain-of-custody forms. No exposed glassware, no on-site solution preparation, no rinses, no clogging filters, and no hazardous materials. Testing can normally be completed within one or two days with minimal disruption of plant operations.

Once received at Frontier Geosciences’ dedicated trace metals laboratory, the traps are digested and analyzed by CVAFS according to the procedures established in EPA Test Method 1631, Revision C. Detection levels for the three Hg species (in g/m3) are HgP = 0.006, Hg0 = 0.0027, Hg2+ = 0.0025 and total Hg = 0.0045, which is essentially in the low parts-per-trillion range. Aliquots of each digestion are archived, and can be re-analyzed up to 40 times.

In summary, the FAMS™/AQS Hg sampling approach provides a relatively economical (much less costly compared to the historical Ontario Hydro method) means of obtaining quality-assured, highly accurate Hg-in-flue-gas concentration measurements with minimal potential for field and/or laboratory contamination. In addition, CVAFS analysis of the fuel source, combustion by-products and APCD additives may be analyzed for a rough mass balance Hg calculation. Once obtained, the results of these analyses provide the basis for sound decision making regarding the operational economics of coal-fired utility boiler systems as it pertains to mercury emissions control.