Process Impacts On Trace Element Speciation

EPA Grant Number: R827649C008
Subproject: this is subproject number 008 , established and managed by the Center Director under grant R827649
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).

Center: Center for Air Toxic Metals® (CATM®)
Center Director: Groenewold, Gerald
Title: Process Impacts On Trace Element Speciation
Investigators: Zygarlicke, Christopher J. , Galbreath, Kevin C. , Mann, Michael D. , Miller, Stanley J. , Schelkoph, Grant L.
Institution: University of North Dakota
EPA Project Officer: Chung, Serena
Project Period: October 15, 1999 through October 14, 2002
Project Amount: Refer to main center abstract for funding details.
RFA: Center for Air Toxic Metals (CATM) (1998) RFA Text |  Recipients Lists
Research Category: Targeted Research


The focus of research this year in this project is on the fundamental transformations of trace element species during coal combustion (Program Area l), with a heavy emphasis on mercury interactions with ash particulate and chlorine. Bench- and pilot-scale combustion testing and advanced sampling and analytical techniques (Program Area 2) were used for experimental studies. Many of the mechanisms of trace element transformation apply to other conversion and combustion systems such as gasifiers and waste incinerators. The information generated is being logged into the CATM Database (Program Area 4), and the experimental data are also being compared with predictive modeling of the formation of mercury and other trace element species in combustion environments. Trace element speciation results are being applied to the development of sorbents and control technologies (Program Area 3) [ 1,2]. Results of this research on mercury interactions with five gas particulate and gaseous chlorine species were presented at the CATM Fourth Annual meeting and will be presented in the 1998 Air and Waste Management Association conference in San Diego.

The primary goal is to examine the fundamental trace metal transformation mechanisms that occur during fuel conversion. The main objectives for this year's work are as follows:

  • Trace metal modes of occurrence in coal. Collaborative research on the mode of occurrence of (species, form, mineral associations) trace elements in coal is being performed in conjunction with a team of international laboratories.

  • Fuel and process impacts on mercury speciation and partitioning. A focused study of the interactions of Hg with flue gas particulate and chlorine gas species and impacts on Hg speciation is being accomplished in this year's work.

  • Model predictions of trace element partitioning. A model formulated under other project work at the EERC for prediction of gasification trace element partitioning by vapor and particulate association is being refined and tested for applications in combustion systems.
  • Approach:

    Trace element mode of occurrence in an Illinois No. 6 coal is being identified using a combination of direct and indirect analytical techniques for determining the trace element forms and associations. Scanning electron microscopy (SEM) and wavelength-dispersive spectroscopy (WDS) were used to selectively examine mineral grains in the coal. Float-sink fractions of ground coal were analyzed using standard chemistry analytical techniques such as atomic absorption spectroscopy (MS) and x-ray fluorescence (XRF). The WDS analysis gave direct association of mineral-bound trace elements and the float-sink analysis gave more of a general inorganic or organic/fine mineral differentiation.

    Mercury speciation tests were performed using a subbituminous PRB Absaloka coal (0.052 +- 0.005 ppm Hg, 50 +- 10 ppm Cl) that was burned at an excess O2(g) of 8.5 mol%, while 50- and 100-ppmv HCl(g) were injected into the furnace at 1330?C. Apparently, the flue gas and/or fly ash produced from Absaloka coal combustion possess intrinsic properties that promote Hg-fly ash sorption. Absaloka was also known to have a history of producing more elemental Hg as a vapor compared to oxidized Hg vapor forms. A simple heated oxygen-nitrogen gas mixture with injections of gaseous elemental Hg and HCl was also employed to check the impacts of the reactor alone on Hg speciation. Hg and Cl speciation analyses of the resulting flue gases were performed at 200? or 250?C using the Radian International method for multimetals collection and Hg speciation. An on-line Hg analyzer was employed to check the transformation of Hg. Baghouse fly ash was also sampled and analyzed for Hg and chlorine bulk quantities. X-ray diffraction analysis of fly ashes were conducted to identify crystalline ash components using methods described by McCarthy and Solem [3]

    Combustion testing using an Illinois No. 6 coal, for which other collaborating data are abundantly available, produced flue gas particulate and vapor trace element species that were partitioned by size and analyzed for trace element concentrations. The data will be compared to predictive modeling results using a modified model called TraceTran. The TraceTran model that already exists at the EERC for gasification systems predicts trace element concentrations as vapor and as particulate forms in various size fractions.


    Trace Element Mode of Occurrence Studies

    A collaborative research effort is under way between the EERC, the USGS, the University of Kentucky, the University of Sheffield (United Kingdom), Imperial College (United Kingdom), CSIRO (Commonwealth Scientific Industrial Research Organization, Australia), BGS, CSIC (Spain), and the Geological Survey of Canada to determine the mode of occurrence of Hg, Se, As, Cd, Pb, Cu, Ni, Sb, Cr, and Zn. It is up to the collaborating laboratory to devise analytical methods determining the modes of occurrence. SEM-WDS analysis of Illinois No. 6 coal shows relative high concentrations of Hg, Ni, Cu, Se, and As in pyrite grains in the coal. Antimony and Cr show pyrite associations as well, but to lesser degrees, and they show some associations with clay and quartz particles. Sink fractions show enrichments of trace elements that have typical sulfide or clay associations. Chemical fractionation work is now being performed to determine organic associations of trace elements.

    Process Impacts on Mercury Speciation

    The primary goal of this investigation was to determine whether chlorination is a dominant Hg transformation mechanism in a real coal combustion flue gas. A secondary goal was to identify the flue gas components and mechanisms accountable for the apparently enhanced Hg sorption characteristics of Absaloka coal fly ash.

    Spike Testing Using Hg'(g) and HCl(g)

    The spiking of 10 µg/m3 Hg0(g) into a simple gas mixture of 8.5 mol% O2 and 91.5 mol% N2 indicated that about 50% of the Hg0(g) spike was transformed to Hg2+X(g) (see figure below). On-line analyzers indicated that the only gaseous components available to react with Hg0(g) were 8.5 mol% O2 and 30 ppmv NO,. Kinetic limitations preclude any significant homogeneous Hg0(g)-O2(g) or Hg0(g)-NOx (g) reactions to account for the formation of Hg2+X(g) [4-71. The formation of Hg2+X(g) must, therefore, involve a heterogeneous or catalytic reaction on ceramic or refractory surfaces in the CEPS. Possible reaction products in the Hg-NOx system include Hg nitrite and nitrate compounds, but they are generally unstable at flue gas temperatures, >200?C, upstream of the sampling location [S, 93. Alternatively, the oxidation of Hg0(g) could involve a heterogeneous reaction with adsorbed (ads) Hg0 or O2 on surfaces or a catalyzed Hg0(g)-02(g) reaction resulting in Hg0(g) (decomposes at 500?C) as the reaction product. The most plausible reaction is:

    Hg0(g,ads) + ?O2(g,ads) ---> Hg0(g) (Reaction 1)

    The formation of Hg0(g) occurs within a temperature range corresponding to the flue gas sampling and Hg0(g) decomposition temperatures, i.e., 200? to 500?C. These temperature conditions exist over a 1.85-m refractory-lined section of the CEPS upstream of the sampling location. Flue gas residence time in this section is estimated to be <0.1 s. Hall and others [4] experimentally investigated the heterogeneous reaction of Hg0(g) with O2(g) in the presence of activated C and fly ash at 20? to 700?C. They concluded that Reaction 1 was not an important heterogeneous reaction because of kinetic limitations and the relatively short residence time in a flue gas duct. An alternative explanation is that Reaction 1 is catalyzed by a component of the refractory used in the combustor. A detailed chemical and mineralogical characterization of refractory from the baghouse-inlet sampling port indicates that corundum (Al2O3), mullite (Al6Si2O13 ), dicalcium silicate (Ca2SiO4 ), anhydrite (CaSO4 ), and r-utile (TiO2) compose the refractory. The CaSO4, component is actually present as an irregular white coating on the refractory surface and is not an essential refractory component. Al203 and TiO2 are known catalysts; however, additional laboratory-scale tests are required to evaluate whether any of these refractory components can catalyze Reaction 1.

    Poor recoveries of Hg0(g) during the 10-µg/m3 Hg0(g) + 100-ppmv HCl spike test (see figure on page 22) provide indirect evidence for the formation of HgCl2(s,l) (melting point 276?C, boiling point 302?C). Apparently, the formation of HgCl2(s,l) involves Hg0(g) as a reactant and not Hg2+X(g) (where X is suspected to be O2-). Cl speciation results, presented in the figure at the top of the next page, indicate that 7 ppmv of Cl,(g) was available in the gas stream to react homogeneously with Hg0(g). Metals such as Al203 are available in the CEPS to catalyze Cl2(g) formation (Reaction l), as indicated by XRD analysis of the refractory. Poor HCl(g) spike recoveries also suggest that Cl was available on surfaces within the combustor to react heterogeneously with Hg0(g).

    Absaloka Coal - HCl(g) Spike Tests

    The Absaloka coal had 7.9 wt% ash and 0.57 wt% sulfur on an as-received basis and 0.52 ppm Hg and 50 ppm chlorine on a dry basis. Hg speciation results for the baseline Absaloka coal test, shown in the figure at the bottom of the next page, are very similar to results obtained from burning the pulverized Absaloka coal in a much larger pilot-scale (580-MJ/hr) combustion system and measuring Hg speciation using several different methods. Hg speciation results are similar despite the fact that pilot-scale testing was performed at a much lower excess O2(g) concentration of 4 mol%. This suggests that excess O2(g) concentration does not significantly affect the Hg speciation of Absaloka coal combustion flue gas. Test results from both combustion systems indicate that Absaloka coal combustion flue gas and/or entrained fly ash possess intrinsic properties which promote Hg sorption as evidenced by relatively high proportions of Hg(p).

    As seen in the following table, a significant depletion in Hg(p), resulting in a relatively low Hg mass balance closure, provides indirect evidence for the formation of HgCl2(s,l) during the Absaloka coal + 100 ppmv-HCl spike test. Chlorine speciation measurements revealed that Cl2(g) was not detected for the Absaloka coal-HCl(g) spike tests. The presence of H2O(g) and SO2(g) apparently inhibited the formation of Cl2(g). Therefore, the formation of HgCl,(s,l) would have to involve HCl(g) or Cl(p) as a reactant. Hg(p) may have reacted heterogeneously with HCl(g) to form HgCl,(s,l) which was then deposited on surfaces within the combustor. This reaction mechanism results in the desorption of Hg(p), implying that Hg is loosely bound to the ash. Alternatively, the presence of HCl(g) may have inhibited Hg-ash sorption and thus the formation of Hg(p). The formation of HgCl2(s,l) would then involve a homogeneous Hg0(g)-HCl(g) or heterogeneous Hg0(g)-Cl(p) reaction. Test results, however, are inconclusive for deciphering the reaction mechanism(s) responsible for the apparent formation of HgCl2(s,l).

    Chlorine analyses of the flue gas and collected fly ash (see figures on page 26), indicate that the injected HCl(g) is scavenged by fly ash. The lack of a positive correlation between Cl and Hg in the bottom figure on page 26 indicates, however, that the Cl retained on fly ash does not create active Cl sites for Hg chemisorption. In addition, the combination of high Cl and unburned C content of the ash produced during the 1 00-ppmv HCl(g) spike test (bottom figure on page 26 and figure on page 27, respectively) did not promote Hg(p) formation. Components other than chlorinated ash and C particles must promote the formation of Hg(p) in Absaloka coal fly ash. XRD analyses of baghouse ash samples were conducted to identify crystalline components that may account for the observed Hg(p)-HCl(g) interaction characteristics. Fly ash produced from Absaloka coal contains lime (CaO) and the CaO-acid gas reaction product CaSO4, as major crystalline phases. Automated SEM analyses by Galbreath and others [l0] indicate that CaO(s) is generally a major component of the fine ash fractions (<2 µm in diameter) of PRB subbituminous coal fly ashes. CaO(s) and portlandite (Ca[OH]2[s]), a hydration product of CaO(s), are effective HgCl2(g) and HCl(g) sorbents [1l-l3]. In bench-scale experiments, Ghorishi and Gullett [13] found that the presence of HCl(g) inhibits the adsorption of HgCl2(g) by Ca(OH)2(s). They hypothesized that this inhibition effect was a result of reactive competition for the available alkaline sites. The inverse relationship between Hg(p) and Cl(p) documented in this investigation is consistent with the hypothesis of Ghorishi and Gullett [13], thus implying that CaO(s) is an important Hg-sorption component of Absaloka fly ash.


    The injection of 10 µg/m3 Hg0(g) into a simple heated gas mixture (8.5 mol% 02, 91.5 mol% N2) indicated that as the gas cooled for approximately 2.5 s from 1250?C to 200?C about 50% of the Hg0(g) spike was transformed to Hg2+X(g), where X is most likely O2-. The formation of Hg0(g), however, is kinetically inhibited at the relatively short residence time available in the combustor [4,6,7]. A catalyzed Hg0(g)-O2(g) reaction involving a refractory metal oxide compound is proposed to explain the apparent formation of Hg0(g). Low recoveries of Hg0(g) during 100-ppmv HCl(g) spike tests into the gas mixture suggest that HgCl2(s,l) was formed and deposited in the combustor. Cl speciation measurements indicate that Cl2(g), an active Hg chlorinating agent, was available to react homogeneously with Hg0(g).

    Combustion testing of a low-Cl (50 ? 10 ppm) Absaloka subbituminous coal at 8.5 mol% excess O2(g) indicated that on average 41%, 19%, and 40% of the total Hg (5.5 ? 0.6 µg/m3) in 250?C flue gas was present as Hg(p), Hg2+X(g), and Hg0(g), respectively. These results are consistent with the relatively high Hg(p) concentrations noted in a much larger-scale 580-MJ/hr combustion system [2], thus corroborating the enhanced Hg sorption capacity of Absaloka coal fly ash. 50- and 100-ppmv HCl(g) spike tests indicate that the Hg sorption capacity of Absaloka coal fly ash is adversely affected by HCl(g), even though Cl is scavenged by ash particles to form Cl(p). Cl2(g) was not detected in the coal combustion flue gas, possibly because of the inhibition effect of H2O(g) and SO2(g). The inverse relationship between Hg(p) and Cl(p) documented in this investigation is analogous to the inhibition effect of HCl(g) on the HgCl2(g) sorption capacity of CaO-based sorbents [13]. The identification of CaO(s) as a dominant ash component and similarities in Hg-Cl-CaO and Hg-Cl-fly ash interactions suggest that CaO(s) is an important Hg-sorption component of Absaloka fly ash. Additional tests are required to confirm this hypothesis and determine whether it applies to other PRB subbituminous coals.

    Trace Element Transformation Modeling

    The CEPS system was set up and configured to run Illinois No. 6 combustion tests with triplicate sampling of flue gas particulate and trace element forms using a five-stage multicyclone fly ash sampler and Ontario Hydro multimetals sampling train. The tests were performed flawlessly, and the ash and impinger samples were analyzed for Hg, Se, Pb, Ni, Cd, and Cr using conventional AAS techniques.

    TraceTran is an EERC model that predicts the concentrations of trace elements in specific particulate size categories and in the vapor phase. The model has been adapted and modified for coal combustion systems, and the algorithms for predicting trace element concentrations in the combustion ash from coal mineral and trace element input data will be verified and calibrated using the Illinois No. 6-CEPS combustion trace element partitioning data. The results from this work were not yet available for this report.


    The form and quantity of trace elements emitted from a fuel energy conversion system are a function of the trace element concentrations, forms, and associations in the fuel; the conversion

    process; and operating conditions. Trace element modes of occurrence have a great bearing on coal-cleaning effectiveness and possibly on the overall combustion transformation mechanisms for certain elements. Arsenic that is associated with an arsenopyrite mineral in a coal pyrite grain may react and transform in its environment differently from arsenic associated with a clay minerals as an accessory. Several elements are being considered in an international study of trace element modes of occurrence.

    The main focus of research this year was in the area of process impacts on Hg speciation. The speciation of Hg is an important factor to consider in understanding its final fate when emitted from a power plant stack, since control options and transport models for Hg rely heavily on its form or species. Chlorine and ash particulate, which are common flue gas constituents in coal combustion flue gas, can have profound impacts on Hg speciation. Hg exists primarily as gaseous elemental species, Hg0(g), and as gaseous or solid inorganic mercuric compounds, Hg2+X (where X is Cl2[g], SO4[s], O[s,g], etc.), in coal combustion flue gas. Hg emissions from coal-fired electric utility boilers can be generally classified into three main forms: particle-associated mercury, Hg(p); gaseous divalent mercury, Hg2+X(g); and gaseous elemental mercury, Hg0(g). Gaseous mercuric chloride (HgCl2[g]) is generally considered to be the dominant Hg2+X(g) form based on theoretical and experimental investigations of Hg speciation in coal and waste combustion flue gases.

    Accurate prediction of Hg speciation and partitioning of other common trace elements is needed for devising air toxic emission control strategies, which are currently under consideration by EPA. Models are being devised in this program to predict particulate and gaseous trace element emissions during combustion.

    Supplemental Keywords:

    Scientific Discipline, Air, Toxics, air toxics, Environmental Chemistry, HAPS, EPCRA, 33/50, Engineering, Chemistry, & Physics, Environmental Engineering, mercury , trace element speciation, Chlorine, mercury, trace elements, trace metal transformation, waste incinerator, mercury & mercury compounds, Mercury Compounds, coal combustion, trace metals

    Progress and Final Reports:

  • 2000
  • 2001
  • Final

  • Main Center Abstract and Reports:

    R827649    Center for Air Toxic Metals® (CATM®)

    Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
    R827649C001 Development And Demonstration Of Trace Metals Database
    R827649C002 Nickel Speciation Of Residual Oil Ash
    R827649C003 Atmospheric Deposition: Air Toxics At Lake Superior
    R827649C004 Novel Approaches For Prevention And Control For Trace Metals
    R827649C005 Wet Scrubber System
    R827649C006 Technology Commercialization And Education
    R827649C007 Development Of Speciation And Sampling Tools For Mercury In Flue Gas
    R827649C008 Process Impacts On Trace Element Speciation
    R827649C009 Mercury Transformations in Coal Combustion Flue Gas
    R827649C010 Nickel, Chromium, and Arsenic Speciation of Ambient Particulate Matter in the Vicinity of an Oil-Fired Utility Boiler
    R827649C011 Transition Metal Speciation of Fossil Fuel Combustion Flue Gases
    R827649C012 Fundamental Study of the Impact of SCR on Mercury Speciation
    R827649C013 Development of Mercury Sampling and Analytical Techniques
    R827649C014 Longer-Term Testing of Continuous Mercury Monitors
    R827649C015 Long-Term Mercury Monitoring at North Dakota Power Plants
    R827649C016 Development of a Laser Absorption Continuous Mercury Monitor
    R827649C017 Development of Mercury Control Technologies
    R827649C018 Developing SCR Technology Options for Mercury Oxidation in Western Fuels
    R827649C019 Modeling Mercury Speciation in Coal Combustion Systems
    R827649C020 Stability of Mercury in Coal Combustion By-Products and Sorbents
    R827649C021 Mercury in Alternative Fuels
    R827649C022 Studies of Mercury Metabolism and Selenium Physiology