This paper explores characterizing coals for environmental quality and whether or not this procedure is possible or desirable. It examines the influence of coal quality on the emission of pollutants to the atmosphere. Some pollutants, as represented by sulfur dioxide, are formed virtually stoichiometrically from their precursors in coal. However, even in these cases, other coal properties such as ash composition can affect the amounts emitted from the combustor. Other pollutants are affected by coal quality in much more subtle ways. Although the nitrogen in the coal is the precursor of most of the nitrogen oxides, the nitrogen content is not the only factor that influences how much of it is converted to NOx rather than inoffensive nitrogen. Particulate emissions reflect the ash content of the coal, but their properties can also determine the effectiveness of the electrostatic precipitators used to control them. Most trace elements, mercury and selenium excepted, are emitted on the particulate matter. Their properties and that of the coal and mineral hosts can affect how they are partitioned among the various size fractions of the ash. Finally, the effect of coal quality on the emissions of carbon dioxide, the major "greenhouse gas" is briefly considered.
This paper is largely based on work reviewed by Davidson (2000), who focuses on five air pollutants produced by coal. First, sulfur emissions that should be expected to reflect the sulfur content of the coal are discussed. Second, the much more subtle effects of fuel quality on the formation and emission of nitrogen oxides are addressed. At the outset, it needs to be stressed that the emissions of nitrogen oxides are strongly dependent on the combustion conditions. Third, effects on particulate emissions and specifically their control by electrostatic precipitators are considered. Then the emissions of trace elements, entrained on particulates or in the gas phase, are examined. Finally, the effect of coal quality on the emissions of carbon dioxide, the major "greenhouse gas," will be briefly considered.
Coal properties are normally determined by standard analytical methods. In general, proximate analysis of coal provides values for the weight percentages of moisture, ash, volatile matter, and fixed carbon. Ultimate analysis provides values for the quantitative elemental analysis of carbon, oxygen, hydrogen, nitrogen, and total sulfur. Other analyses and tests that are often performed include heating value, hardgrove grindability index (HGI), ash analysis, and ash fusion temperatures. In addition, it is possible to determine forms of sulfur, chlorine, trace elements, maceral composition, and ash resistivity.
The measurable coal properties in themselves can be indicators of other less well defined parameters that indicate coal quality. For example, coal rank is the degree or stage the coal has reached in its coalification -- its geochemical maturity -- but it is not an objective property. One of the commonly used measures of rank is the volatile matter (VM), which is determined by heating the coal under controlled conditions. The percentage loss of mass less the percentage of moisture gives the proportion of volatile matter. In the American Society for Testing Materials (ASTM) standard method, the sample is heated to 950°C (~1740°F), well below the temperatures normally encountered in pulverized coal combustion (the test was designed primarily for assessing a coal's suitability for coking). Another measure, fixed carbon (FC), is neither fixed nor carbon, because it contains appreciable amounts of sulfur, hydrogen, nitrogen, and oxygen and is sensitive to carbonization conditions. It is the notional solid residue, usually determined by difference, of the remainder of the coal when the proximate analysis values of moisture, ash, and volatile matter have been deducted from the original mass. It incorporates all the error, bias, and scatter involved in the proximate analysis. The fuel ratio is the ratio FC/VM and is, essentially, an inverse measure of volatility.
It is obvious that coal properties and quality should be major influences on the quantities of pollutants emitted to the atmosphere. For example, the sulfur content in the coal will determine the amounts of sulfur dioxide produced by combustion of that coal. But for others the influences are less clear cut since the combustion chemistry is more complex. Nitrogen can be emitted as nitric oxide and nitrogen dioxide, contributors to acidification; nitrous oxide, implicated in the enhanced greenhouse effect; or nitrogen, the harmless major constituent of the atmosphere.
For the most part, the extent of sulfur emissions in large-scale pulverized coal combustion is straightforward; nearly all the sulfur in the coal is converted to sulfur dioxide. The only coal properties that greatly affect the emission of SO2 are the total sulfur content and the ash, and the amount captured by the ash is only a small part of the total. Most of the SO2 is either emitted or captured by flue gas desulfurization. High-chlorine coals, however, can affect the operation of the desulfurization equipment.
The emissions of sulfur dioxide are directly related to the sulfur content of coal.
Figure 1. Formation and fate of nitrogen oxides during combustion of coal (Wójtowicz and others, 1993) |
By using combustion modifications (primary measures), it is possible to control or prevent the formation of "thermal Nox." The remaining emission of NOx is almost entirely owing to the nitrogen in the coal. Fuel nitrogen does indeed have a major influence on the amount of NOx emitted. Other coal properties that affect the production of NOx are the volatile matter, the fuel ratio (FC/VM), and the carbon-to-hydrogen ratio.
Recognition of the importance of the fuel nitrogen content has led to some proposals that coals be classified by their nitrogen content. For example, Okamoto (1997) suggested that "good quality coal" contains less than 1.5 percent N, with "average quality coal" containing 1.5 to 2.0 percent, and "low quality coal" containing more than 2.0 percent.
But, as figure 2 shows, the fuel nitrogen content of coal is only indirectly related to the NOx emissions; the level of NOx emissions from combustion of coal cannot be anticipated on the basis of nitrogen content alone.
Figure 2. Predicted NOx (top) and actual NOx from pilot-scale tests (bottom) versus nitrogen content (Bennett, 1996). |
The upper part of the figure shows the CQIM prediction for a wide range of coals tested at the Australian Coal Industry Research Laboratories (ACIRL). The CQIM NOx prediction used in 1996 had a strong relationship with the nitrogen content of the coal. However, ACIRL also measured NOx emissions from the same coals tested in a pulverized coal boiler simulation furnace. The lower part of the figure shows no correlation between the nitrogen content of the coal and the measured NOx emissions. It was conceded that the results from the test furnace could not be directly related to full-scale units because there are significant differences, in terms of NOx formation, between the thermal environment and the mixing conditions at pilot scale and full scale. Because the pilot-scale results were all determined under the same conditions, the heat release and excess air impacts would be essentially the same, leaving fuel nitrogen as the only variable.
The influence of coal properties on NOx emissions, despite an intensive research effort, has
not yet been well enough defined to allow the prediction of NOx emission levels from these properties alone. Many factors influence NOx formation. The roles and interaction of factors such as nitrogen partitioning between the volatiles and chars, nitrogen functionality, coal rank, coal nitrogen content, and volatile content in the formation of NOx are not well understood. The combustion conditions are also extremely important; because even small changes in operating parameters may be equal to or greater than the fuel parameter effects, correlating data from plants with different firing configurations should be avoided. Data from plants with different firing configurations should be treated as distinct data sets.
Volatile matter is clearly an important influence on NOx formation, but the scatter of published data has led several research groups to seek better measures than those determined under the standard proximate analysis conditions.
The new techniques being developed are designed to be more representative of combustion conditions. At more representative temperatures, the nitrogen contents of the produced chars could provide an indication of utility boiler NOx levels. The further development of techniques such as high-temperature wire mesh and the increased direct comparison of laboratory-scale studies with larger plant studies, together with close attention to coal characterization, should allow more accurate and reliable correlations to be developed for the wider range of worldwide coals that are being used for power generation. However, because these methods are not yet standardized, slightly different experimental conditions can lead to differing results.
Although the fuel properties are definitely important, perhaps too much work has focused on the fuel itself; it is the interaction between the fuel and the power station that is important. Figure 3 shows the NOx emissions over 12 months for a 660-MW unit in Australia. Emissions are load dependent, reducing with load, but there is a great deal of scatter even at full load. The power station takes a single coal from a single mine, so the figure is showing plant-related effects rather than coal-related effects.
Figure 3. 10-minute averaged NOx emissions over 12 months, 660 MW unit (Lowe, 1999). |
Some of the factors that impact on combustion performance and NOx are:
All these factors will change over the life of the plant, between maintenance overhauls, and through the year.
There is still a need for further research into the chemistry of NOx formation if powerplant operators are to have the ability to finely tune boilers to the lowest possible level of NOx emissions while still achieving good combustion performance for a wide range of coals. This knowledge is especially important for stations firing imported coals from various countries.
The effect of boiler configuration has become more important with the development of low NOx burners. Low NOx burners are designed to control fuel and air in order to create larger and more branched flames.
Finally, owing to the complexity of the NOx formation process and its dependence on site-specific combustion conditions, a successful predictor for NOx emissions should be based on changes in NOx emissions (relative to a baseline) resulting from coal property changes. The successful predictor might be based on coal properties measured by current standard analytical methods, advanced analytical methods, or models, but it seems unlikely that absolute NOx levels can, as yet, be predicted with any confidence.
The emission of particulate matter is largely a consequence not only of the coal's properties but also of the efficiency of the system used to control particulate emissions. The most commonly used systems are electrostatic precipitators, followed by fabric filters (baghouses). The efficiency of fabric filters is largely determined by the particle size of the fly ash, but electrostatic precipitator (ESP) efficiency is affected by other ash properties.
It is clear that the coal properties affecting the operation of an ESP will be mainly those of the mineral matter in the coal, because it is the minerals that are converted to ash. The fly ash properties that, in turn, influence the ESP operation include:
Most methods used in predicting ESP performance include ash composition as a major factor in predicting the fly ash resistivity. However, any predictions of ESP performance will be specific for the ESP installed and cannot be generalized on the basis of the coal-quality parameters alone. Prediction of fly ash precipitation is an inexact science; both pilot plant testing and electrical simulation studies remain important in determining the ease of precipitation of fly ash in practice.
The emission of particulates is also important because they have been implicated in trace element emissions via the fine particles that are not collected by ESPs. This issue will be addressed next.
Of the eleven trace elements that have been identified as potential hazardous air pollutants (HAPs), all, with the exception of mercury and selenium, are found mainly on solids that are efficiently trapped by electrostatic precipitators (ESPs). The retention efficiencies for individual trace elements can be approximately the same as the overall collection efficiency of the ESP that can exceed 99.9 percent. Despite this, because the efficiencies of ESPs are generally lowest in the 0.1-1.5 micron particle size range there has been concern that trace elements may escape ESPs if they are 'preferentially enriched' (on a mass basis) on these fine particles. It is unclear at present whether it is the particle size itself or the chemical composition of the particle which gives rise to potential health hazards.
Figure 4. Partitioning of trace elements during coal combustion (Benson and others, 1997) |
In order to predict trace-element emissions, we need to know the distribution of the trace elements in coal to the different size fractions of ash in the combustion products -- partitioning. This partitioning is dependent on many factors, including the size distribution of the coal, the combustion conditions, the forms of occurrence of the trace elements in the coal, and the interaction of the different elements both during and after combustion.
One of the principal factors governing the partitioning of trace elements is their volatility and the volatility of their compounds. However, volatility alone is not sufficient to predict emissions because it neglects other factors, especially the interactions among the mineral species in the flue gas.
Trace-element emissions cannot be estimated from conventional standard analyses of coal because these do not include trace-element determinations. There are also large uncertainties concerning their forms and associations within the mineral and organic matter of coal.
Lowe and others (1993) listed three methods of estimating trace-metal emissions from combustion:
They noted that the U.S. Environmental Protection Agency had used generalized emission factors to summarize trace-metal emissions. These emission factors simply express the amount of a species that will be emitted from a boiler in terms of mass per unit of energy produced. It was pointed out that most researchers consider the use of generalized trace-metal emission factors to be very inaccurate because the approach does not account for variations in fuel composition. It is possible to "predict" that the quantity of trace metals in the flue gas will be greater than 100 percent of the quantity present in the coal. The European Environment Agency uses enrichment factors in its calculation method for heavy-metal emissions. The default values for metals emitted in gaseous form (fg) are 0.5 wt percent for arsenic, 90 wt percent for mercury, and 15 wt percent for selenium.
However, given the lack of uniform trace-element concentrations and distributions in different coals, the use of default values should be viewed with utmost caution. However, if the trace-element concentrations in a specific coal are known, it should be possible to predict the emissions of many of them. Predictions of the emissions of the more volatile elements such as mercury, selenium, and arsenic may be more difficult and less accurate.
Improving coal quality is the first step to improving thermal efficiency and hence reducing greenhouse gas emissions (Smith, in press). The rank of coal is an important factor because higher rank coals produce less CO2 per unit of energy than lower rank coals, especially if they are measured on a net calorific value or lower heating value (LHV) basis that does not include the latent heat of water vapor.
The moisture in coal is a reason for using LHV rather than HHV, and CO2 emissions could be abated if high-moisture coals were dried.
The ash content of coal usually reduces the heat available to the boiler and thus causes CO2 emissions per unit of energy to increase. Carbonate minerals also decompose to release CO2 directly. One of the main reasons for the low efficiency of coal-fired powerplants in China is the high ash content (~23 percent) of the indigenous coals. The situation in India is similar. An IEA Coal Research study (Couch, 2000) has pointed out that there are well over 4,000 coal-fired boilers that could be operated more efficiently by improving coal quality either by drying or beneficiation.
Sulfur in coal contributes to the calorific value (1percent S = ~0.1 MJ/kg) with no additional CO2. However, emissions of SO2 are not exactly welcome either. Ironically, when western subbituminous coal is substituted for bituminous coal in the United States, sulfur dioxide emissions may have been abated, although carbon dioxide emissions may have increased by 6 to 8 percent (Quick and Glick, 2000).
There is evidence that inertinite in U.S. coals produces more CO2 than the vitrinite (Quick and Glick, 2000). However, it would be unwise to assume that this finding would apply to coals from other countries, given the large variations in properties within maceral groups. Not all inertinites are inert, and not all vitrinites are reactive. This disparity is a consequence of the formal definition of macerals and the variations among coals from different parts of the world.
Bennett, P. A., 1996, Evaluation of CQIM, in A practical workshop on impact of coal quality on thermal coal utilisation: CRC for Black Coal Utilisation, Paper 20, 14 p.
Benson, S. A., Pavlish, J. H., Zygarlicke, C. J., and Miller, S. J., 1997, Trace element behavior in coal combustion systems: Effects of coal quality on powerplants:, Palo Alto, Calif., Electric Power Research Institute, 11 p.
Couch, G. R., 2000, Opportunities for coal preparation to lower emissions: London, IEA Coal Research, CCC/30, 46 p.
Davidson, R. M., 2000, How coal properties influence emissions: London, IEA Coal Research, England, 56 p.
Lowe, A., 1999, Australian power station NOx overview, in Proceedings: The future of pulverised coal firing: Newcastle, NSW, Australia, CRC for Black Coal Utilisation, 12 p.
Lowe, P. A., St. John, B., and Breed, W. S., 1993, Trace metal emissions from coal fired boilers, in Proceedings of the tenth annual international Pittsburgh coal conference, Chiang S-H, ed., University of Pittsburgh, Pittsburgh, Pennsylvania, p. 795-800.
Okamoto, A., 1997, Evaluation of thermal coal quality (Japan): Coal and Safety, vol. 11, p. 18-32.
Quick, J. C., and Glick, D. C., 2000, Carbon dioxide from coal combustion: variation with rank of US coal: Fuel, vol. 79, no. 7, p. 803-812.
Smith, I., in press, Potential for economic greenhouse gas reduction in coal-fired power generation: London, IEA Coal Research,
Wójtowicz, M. A., Pels, J. R., and Moulijn, J. A., 1993, Combustion of coal as a source of N2O emission: Fuel Processing Technology, vol. 34, no. 1, p. 1-71.