Environ. Sci. Technol. 2004, 38, 381-389
Characterization of Airborne and Bulk Particulate from Iron and Steel Manufacturing Facilities STEVEN D. MACHEMER* National Enforcement Investigations Center, United States Environmental Protection Agency, Box 25227, Building 25, Denver Federal Center, Denver, Colorado 80225
Characterization of airborne and bulk particulate material from iron and steel manufacturing facilities, commonly referred to as kish, indicated graphite flakes and graphite flakes associated with spherical iron oxide particles were unique particle characteristics useful in identifying particle emissions from iron and steel manufacturing. Characterization of airborne particulate material collected in receptor areas was consistent with multiple atmospheric release events of kish particles from the local iron and steel facilities into neighboring residential areas. Kish particles deposited in nearby residential areas included an abundance of graphite flakes, tens of micrometers to millimeters in size, and spherical iron oxide particles, submicrometer to tens of micrometers in size. Bulk kish from local iron and steel facilities contained an abundance of similar particles. Approximately 60% of blast furnace kish by volume consisted of spherical iron oxide particles in the respirable size range. Basic oxygen furnace kish contained percent levels of strongly alkaline components such as calcium hydroxide. In addition, concentrations of respirable Mn in airborne particulate in residential areas and at local iron and steel facilities were approximately 1.6 and 53 times the inhalation reference concentration of 0.05 µg/m3 for chronic inhalation exposure of Mn, respectively. Thus, airborne release of kish may pose potential respirable particulate, corrosive, or toxic hazards for human health and/or a corrosive hazard for property and the environment.
Introduction The episodic deposition of airborne particulate onto homes and property and the rapid deterioration of home and auto paint allegedly caused by particulate fallout were common complaints of residents living near iron and steel manufacturing facilities in Sparrows Point, MD. Much of the particulate fallout was described as gray or silvery flakes. Such complaints prompted a preliminary study to determine if particulate from the nearby iron and steel facilities was a contributing source of the airborne particulate deposited in the residential areas and to identify if potential human health or environmental hazards were associated with bulk particulate found at the nearby facilities. To accomplish this, airborne particulate was collected at the iron and steel facilities and in residential areas nearby. In addition, samples of bulk particulate material, commonly referred to as kish, were collected at the iron and steel facilities. Kish was characterized to determine if unique particle characteristics * Corresponding author phone: (303) 462-9108; fax: (303) 4629028; e-mail:
[email protected]. 10.1021/es020897v Not subject to U.S. Copyright. Publ. 2004 Am. Chem. Soc. Published on Web 12/12/2003
existed which could be used to identify airborne kish particles in residential areas around the facilities. Kish was also characterized to identify any potential human health or environmental hazards which might be associated with kish particles. The terms kish and kish graphite refer to the graphite flakes formed during the transfer of molten iron as it cools between manufacturing steps (1-7). As molten iron cools and the solubility of carbon is exceeded, graphite flakes form on the surface of the iron. Kish graphite consists of platelets of graphite, tens to thousands of micrometers across, tens to hundreds of micrometers thick, and often exhibiting pseudohexagonal corners and growth steps. The term kish is sometimes applied more generally to refer to the whole byproduct material produced from molten iron, which includes iron oxide particulate as well as graphite flakes. Iron oxides in kish form from the forced oxidation of iron vapor or fume present at the surface of the melt. Iron oxides in kish consist of ferrous and ferric oxides and occur as a very fine reddish brown powder. This powder may also contain metallic iron. Other elements typically found in kish include Al, Ca, Mg, Mn, Si, and S. Steel industry processes in which kish is formed include blast furnace (BF) and basic oxygen furnace (BOF) processes (7-11). BFs are generally used to produce pig iron from ironbearing materials and coke. Iron ore is the common source of iron, primarily composed of the iron oxides hematite (Fe2O3) and magnetite (Fe3O4) or hydrous iron oxides, such as goethite (FeO(OH)). Coke is produced from coal and provides the reductant for reducing the iron ore and provides heat through combustion with air. Fluxing agents used in BFs include limestone (CaCO3) and dolomite (CaMg(CO3)2), which provide CaO and MgO to balance acid components, such as silicon and aluminum oxides, in the iron ore and coke. CaO and MgO form a slag to remove sulfur and other impurities. Emission control systems for BFs typically consist of an active system of emission-capturing hoods connected to a baghouse by ductwork. BOFs are commonly used to produce steel from pig iron and/or scrap iron through the introduction of oxygen gas and fluxes to remove or decrease the presence of unwanted elements such as C, Mn, P, Si, and S. Fluxing agents used in BOFs include lime (CaO), fluorspar (CaF2), and mill scale (iron oxides) to form a slag with phosphorus and sulfur impurities. Emission control systems for BOFs typically consist of a fume suppression system designed to prevent the formation of iron oxide emissions through the exclusion of oxygen at the surface of the hot metal.
Methods The iron and steel facilities where samples were collected are referred to here as facility 1 and facility 2. Five bulk kish samples from facility 1, two air filters from facility 2, and two air filters and 11 fallout samples from residential areas were collected. Samples of bulk kish from facility 1 consisted of two samples from the BOF area and three samples from the BF area. Airborne particulate was collected on two air filters at facility 2, where iron pigs were cast from hot metal. The BF, BOF, and facility 2 operations were all located on the same site within a few hundred meters of each other. “Fallout” samples consisted of the passive collection of particles from atmospheric deposition in residential areas and were similar distances from the different sources of kish. For over a year, fallout samples and air filters were collected sporadically from widespread sites in residential areas within a 2.5 km radius of the iron and steel facilities, and these samples are VOL. 38, NO. 2, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
381
referred to collectively as “receptor” samples. Receptor samples were located between 2 and 2.5 km from the facilities. Bulk kish samples from facility 1 and exposed air filters from facility 2 are referred to collectively as kish “reference” samples because characteristics of these samples were used for comparison to characteristics of receptor samples. Analyses were conducted on bulk kish from facility 1 to determine particle size distribution; major crystalline species; total carbon and carbonate content; bulk elemental composition; pH; results of the U.S. Environmental Protection Agency’s (EPA) Toxicity Characteristic Leaching Procedure (TCLP) (12); and the character of individual particles, including size, shape, texture, and elemental composition. In addition, the carbon content, elemental composition, and pH of the less than 38 µm size fraction of the bulk kish samples were determined to ascertain the distribution of elements and pH properties relative to particle size in the bulk kish material. Fallout samples and air filters from residential areas and air filters from facility 2 were also analyzed to determine major crystalline species and the character of individual particles. Finally, the concentration of respirable Mn was determined on air filters from facility 2 and the receptor area. Collection of Airborne Particulate onto Air Filters. Airborne particulate was collected onto 47 mm diameter Teflon air filters with a deposit area of 12.6 cm2 using an Airmetrics MiniVol portable air sampler. Particle size separation was achieved for the collection of PM10 (particulate matter with a mass median aerodynamic diameter less than 10 µm) by the installation of a PM10 impaction particle size separator. TSP (total suspended particulate) was collected without the impactor installed. The flow rate and sampling duration for the receptor air filters were 5 L/min for three periods of 8 h on three consecutive days (7.2 m3/sampling cycle). The flow rate and sampling duration for the facility 2 air filters were 5 L/min for 2 h (0.6 m3/sampling cycle). The facility 2 air filters were located inside the facility, approximately 6 m from the active operation. Mechanical Sieving. Splits of bulk kish samples were mechanically sieved using a 400 mesh sieve on a Rotap mechanical sieving machine for 10 min/sample to separate the less than 38 µm size fraction. This particle size separation was conducted to compare elemental concentrations of the less than 38 µm size fractions with those of the bulk samples. Elements occurring in significantly greater concentrations in the less than 38 µm size fractions may also be expected in greater concentrations in the respirable particle size fractions compared to bulk samples. Particle Size Analysis by Laser Diffraction. Particle size distributions of bulk kish samples were determined by laser diffraction using a Malvern particle size analyzer. Bulk kish was prepared by adding 0.2-0.3 g of sample to 20 mL of a sodium hexametaphosphate solution. Sample slurries were allowed to sit overnight and were mildly sonicated for 5 min prior to analysis. Scanning Electron Microscopy (SEM). The size, shape, texture, and elemental composition of individual particles in bulk kish, air filters, and fallout samples were characterized by SEM. SEM stubs with adhesive tape were gently pressed onto the particulate material to be analyzed. Specimens were examined under a JEOL JSM 6400 SEM equipped with a PGT PRISM digital energy-dispersive X-ray spectrometer (EDS). The acceleration voltage was set at 20 keV, the beam current was set between 0.4 and 3.1 nA, and the working distance was 15-16 mm. Optical Microscopy. A Leitz binocular microscope was used to characterize individual particles on slide mounts of bulk kish and airborne particles and on a polished section of bulk BF kish. 382
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 2, 2004
X-ray Diffraction (XRD). Crystalline species in bulk kish, air filters, and individual particles in fallout samples were identified with a Phillips PW 1800 X-ray diffractometer with Cu Ka radiation. Bulk kish was analyzed in bulk sample holders, and particulate on air filters was analyzed directly. Individual submillimeter to millimeter size flake particles collected in fallout samples were transferred to zero background planchets for analysis. As feasible, additional particulate from fallout samples was prepared similarly. Diffraction data were collected at 50 kV and 35 mA from 2° to 80° 2θ using a step width of 0.02° 2θ and a scanning rate of 2.75 s/step. The area irradiated on bulk samples, air filters, and planchets was approximately 1 cm2. A similar XRD analysis technique for air filters may be found in O’Connor and Jaklevic (13). Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES). To determine the elemental composition of bulk kish and the less than 38 µm size fractions, sodium peroxide fusions were conducted on 0.2 g of sample with 2.0 g of Na2O2 and 1 pellet of NaOH in a vitreous carbon crucible. Crucibles were heated to 470 °C for 2 h. After cooling, crucibles were rinsed with a solution of 50 mL of distilled, deionized (DDI) water, 25 mL of HCl, and 2 mL of 30% H2O2. After 1 h, sample solutions were vacuum filtered through Nalgene disposable, 0.8 µm, cellulose nitrate membrane filtration units. ICP-OES was conducted on filtered solutions using a Thermo Jarrell Ash ICAP 61-E with cross-flow nebulization. X-ray Fluorescence (XRF). Cd and Ag in bulk kish and the less than 38 µm size fractions were determined with an Oxford Instruments ED2000 EDS. Similarly, Mn on PM10 air filters from facility 2 and the receptor area was determined with a Kevex 770 EDS. Carbon Analysis. Carbon content in bulk kish and the less than 38 µm size fractions was determined with an Eltra model CS-500 carbon-sulfur determinator with high-temperature combustion. Analyses were conducted for low and high carbon levels at 1350 and 1400 °C, respectively, over 60 s using a vanadium pentoxide accelerator. Carbonate Analysis. Carbonate content in bulk kish was determined by coulometric titration using a Coulometrics CO2 coulometer model 5130 and acidification with 2 N perchloric acid prior to titration. Potentiometry. A Beckman pH meter and Orion glass combination electrode were used to measure the pH of 5 g of bulk kish mixed with 5 mL of DDI water. TCLP Analyses. The TCLP, EPA Method 1311 (12), was performed on bulk kish to determine if kish displayed solid waste characteristics for toxicity. Leachates were analyzed for the TCLP metals As, Ba, Cd, Cr, Pb, Se, and Ag by ICP-OES.
Results Particle Size Distribution of Bulk Kish Samples. Particle size distributions of bulk kish are given in volume percent for particle size ranges representing the volume mean diameter of equivalent spheres (Table 1). The particle size distributions of BOF and BF bulk kish from facility 1 were distinctly different from each other. Although most particles comprising BOF kish were greater than 5-10 µm in size, a potentially significant portion had sizes less than 5 µm, approximately 7 vol %. In contrast, more than 60% of BF kish by volume was comprised of particles less than 5 µm in size, more than half of which, or 37% of the total on average, consisted of particles less than 1 µm. These results are compared to aerodynamic diameter criteria in the Discussion. Crystalline Species in Bulk Kish Samples and Facility 2 Kish on Air Filters. XRD results indicated the presence of crystalline graphite (C), hematite, and magnetite in all five bulk kish samples (Table 2). Maghemite (Fe2O3), with a crystalline structure more similar to that of magnetite than
TABLE 1. Particle Size Distribution (D)a of Bulk Kish from Facility 1 in Volume Percentb kish sample sample type 500 µm 012 013 014 015 016
bulk BOF bulk BOF bulk BF bulk BF bulk BF
3.6 NAc 34.8 39.4 37.3
0.7
2.4
2.6
12.1
13.6
18.2
19.4
17.1
10.3
8.8 9.2 9.3
17.2 15.8 15.6
6.4 5.2 5.5
10.9 8.1 9.2
3.7 4.0 4.7
10.6 10.3 10.3
7.3 7.5 7.5
0.3 0.5 0.6
0.0 0.0 0.0
a D ) geometric diameter; to convert to aerodynamic diameter (D ), multiply by the square root of particle density. Particle densities (g/cm3): a calcite, 2.71; graphite, 2.2; hematite, 5.25; magnetite, 5.17; portlandite, 2.2. b Particle size ranges represent the volume mean diameter of equivalent spheres. The estimated relative standard deviation of the volume mean diameter was (0.5%. c NA ) not analyzed.
TABLE 2. Crystalline Species in Bulk Kish from Facility 1 and Air Filters from Facility 2 kish sample
crystalline speciesa
sample type
012, 013
facility 1, bulk BOF
014, 015, 016 028 029
facility 1, bulk BF facility 2 air filters,b PM10c facility 2 air filters,b TSPd
graphite (33-35%), hematite (10-30%), magnetite (10-30%), calcite (3-4%), portlandite (7-8%) graphite (4-5%), hematite (20-40%), magnetite (20-40%), maghemite (20-40%) hematite (40-60%), magnetite (40-60%) hematite (40-60%), magnetite (40-60%)
a Approximate abundances (wt %) were estimated based on results of XRD and the chemical analyses. b Mass loadings of PM 10 and TSP air filters from facility 2 are not comparable because they were not collected contemporaneously. c PM10 mass loading ) 130 µg/cm2. d TSP mass loading ) 57 µg/cm2.
TABLE 3. Crystalline Species in Receptor Sample receptor sample
sample type
017 018 019
fallout fallout fallout
020 021 022 023 024 025 038 039 044 045
fallout fallout fallout fallout fallout fallout air filter - TSPa,b air filter - PM10a,c fallout fallout
crystalline species graphite graphite graphite, hematite, calcite graphite graphite graphite graphite graphite, calcite graphite none detected none detected graphite, calcite graphite
a Mass loadings of PM and TSP air filters from receptor areas are 10 not comparable because they were not collected contemporaneously. b TSP mass loading ) 57 µg/cm2. c PM 2 10 mass loading ) 9.0 µg/cm .
that of hematite, was also present in BF kish. For BOF kish, XRD results also indicated the presence of calcite (CaCO3) and portlandite (Ca(OH)2). XRD results for air filters from facility 2 showed the presence of magnetite and hematite. Crystalline Species in Receptor Samples. In all fallout samples, individual submillimeter to millimeter size, silvery flakes with a metallic luster were identified by XRD as graphite (Table 3). Some fallout samples also contained hematite and calcite. Due to low mass loadings, no crystalline species were identified in receptor air filters. Estimates indicated that no more than 7-8 µg/cm2 were present for any crystalline phase likely to have been deposited on receptor air filters. However, much less for any given phase was likely to have been present. This was well below the approximately 17-18 µg/cm2 estimated for each of the iron oxide phases, hematite and magnetite, identified on the facility 2 TSP air filter. A typical minimum abundance for detecting many crystalline phases by XRD is approximately 20 µg (14). However, the minimum amount of material required to detect a given phase varies considerably depending on the crystalline structure of a given phase, the particular assemblage of phases present, and the substrate on which the material is analyzed. Respirable Mn Concentrations from Facility 2 and Receptor Air Filters. The concentration of respirable Mn on
TABLE 4. Carbon and Carbonate Content in Bulk Kish and Carbon Content in the
