Binding Waste Anthracite Fines with Si-Containing Materials as an

ARTICLE pubs.acs.org/est

Binding Waste Anthracite Fines with Si-Containing Materials as an Alternative Fuel for Foundry Cupola Furnaces He Huang,† John T. Fox,† Fred S. Cannon,*,† Sridhar Komarneni,‡ Joseph Kulik,‡ and Jim Furness§ †

Department of Civil and Environment Engineering and ‡Materials Research Institute and Department of Crop and Soil Sciences, The Pennsylvania State University, University Park, Pennsylvania 16802, United States § Furness-Newburge Inc., Versailles, Kentucky 40383, United States

bS Supporting Information ABSTRACT: An alternative fuel to replace foundry coke in cupolas was developed from waste anthracite fines. Waste anthracite fines were briquetted with Si-containing materials and treated in carbothermal (combination of heat and carbon) conditions that simulated the cupola preheat zone to form silicon carbide nanowires (SCNWs). SCNWs can provide hot crushing strengths, which are important in cupola operations. Lab-scale experiments confirmed that the redox level of the Sisource significantly affected the formation of SiC. With zerovalent silicon, SCNWs were formed within the anthracite pellets. Although amorphous Si (þ4) plus anthracite formed SiC, these conditions did not transform the SiC into nanowires. Moreover, under the test conditions, SiC was not formed between crystallized Si (þ4) and anthracite. In a full-scale demonstration, bricks made from anthracite fines and zerovalent silicon successfully replaced a part of the foundry coke in a full-scale cupola. In addition to saving in fuel cost, replacing coke by waste anthracite fines can reduce energy consumption and CO2 and other pollution associated with conventional coking.

1. INTRODUCTION About half of the iron poured in America is derived from cupolas.1 The carbon source in foundry cupolas requires high energy content and a fast burning rate.1-4 Moreover, the solid fuel should maintain it's structural integrity in different zones in the cupola (see details in the Supporting Information). Foundry coke, which is made through extensive pyrolysis, has been used in cupolas for a couple of centuries. However, only bituminous coals that thermally fuse are currently processed for coke. The prolonged pyrolysis (1000 °C, 16-36 h) required for coking consumes a tremendous amount of energy and produces carbon dioxide and other pollutions.5 The price of foundry coke increased 400% from 2002 to 2010. Currently, the foundry industry has not found a suitable alternative fuel for cupolas that alleviates the energy requirement for coking while also providing fast burning and structural integrity. In this study, a replacement fuel was developed by binding waste anthracite fines, which often represent an unusable residue from the anthracite mining process, that have been discarded in the valleys of eastern Pennsylvania’s coal region over the past centuries. Anthracite, by nature (without pyrolysis), offers similar chemical properties as coke, as show in the Supporting Information (Table S-1). However, anthracite fines must be first briquetted into bricks. Such bricks must maintain a structural integrity in the cupola preheat and melt zones. Innovative solutions may come from thoughtful and creative analysis on the flow of materials and energy in the industrial r 2011 American Chemical Society

operation.6-11 When materials flow in a cupola was evaluated (Figure S-1, Supporting Information), it was noticed that silicon metal is conventionally added into the cupola to improve the iron quality. Since silicon is charged into the cupola anyway, it would be useful to bind together the silicon and anthracite fines with an organic binder and feed these all together into the cupola as bricks. The organic binders could provide the ambient-temperature strength. Then, the silicon metal could react with anthracite in the preheat zone to form SiC, especially silicon carbide nanowires (SCNWs) to hold the anthracite fines together while the bricks subsequently burn in the melt zone. SCNWs have been used by others to reinforce ceramic composites.12,13 An even more ambitious approach would be to bind Si (þ4) materials with the anthracite fines and produce SiC by utilizing the high temperature reducing condition in the preheat zone. Then, the usage of silicon metal, which is also produced through an energy intense process, could be reduced. The authors herein made anthracite pellets from recipes that employed different Si-containing materials. These were tested for their unconfined compressive strength (UCS) at both ambient and pyrolyzed (at 1400 °C) conditions. The effect of different Received: August 6, 2010 Accepted: February 8, 2011 Revised: February 8, 2011 Published: March 02, 2011 3062

dx.doi.org/10.1021/es102684k | Environ. Sci. Technol. 2011, 45, 3062–3068

Environmental Science & Technology

ARTICLE

Table 1. Strengths of Anthracite Pellets Bound with an Array of Binders and Additives Recipes carbon source 100 g of anthracite fines þ1 g of collagen

foundry cokeb a

other additives

unconfined compressive

drop shatter

strength (UCS) before pyrolysis (kPa)

remaining before pyrolysis (%)

UCS after pyrolysis (kPa)

no other additives

1234

76

0

1 g of fructose

1095

99

0

3 g of sodium silicate (0.7 g as silicon)

N/Da

90

0


2455

99

0


2342

98

∼28

1 g of kaolinite (0.22 g as silicon) 10 g of kaolinite (2.2 g as silicon)

N/D 1340

91 93

0 ∼7

2 g of silicon

1230

77

69

5 g of silicon

N/D

80

345

10 g of silicon

1250

91

690

1 g of fructose, 10 g of silicon

1300

99

720

no additives

2758

93

2758

N/D: not determined. b For UCS after pyrolysis, the coke also experienced the same pyrolysis process as did the bindered pellets.

Si-containing materials on the postpyrolysis strength was analyzed on the basis of the difference in their morphologies and crystal structures after pyrolysis. The cause of the differences is discussed on the basis of the redox potential and surface tension. The authors collaborated with Ward Foundry (Blossburg, PA) and conducted a full-scale demonstration of these briquetted anthracite-silicon bricks in a cupola that exhibited favorable results.

2. EXPERIMENTAL SECTION 2.1. Raw Materials. Anthracite fines used in this research were obtained from Jeddo Coal Company (Wilkes-Barre, PA). These were Jeddo’s “#5” anthracite fines (identified as “anthracite fines” herein). Sieve analysis shows that most of the anthracite fines are within the range of U.S. mesh # 10-80 (2000-177 μm). For some experiments, these coal fines were crushed into powders that could pass through a U.S. mesh #100 (150 μm) sieve (identified as “anthracite powder” herein). The proximate analysis and the heat content of the anthracite fines were compared with a typical foundry coke in Table S-1, Supporting Information. The foundry coke was that which has been used by Ward Manufacturing (Blossburg, PA). The collagen-based binder was provided by Entelechy, representing Hormel Foods Company (Austin, MN), and it was received as a dry, small, granular form. Fructose (98%-102%) was obtained from Mallinckrodt Baker, Inc. (Phillipsburg, NJ). Materials that contain silicon at three different redox levels were tested. These levels included metallic Si (0), amorphous Si (þ4), and crystallized Si (þ4). Elemental silicon lumps (10 cm and smaller, 98.4% purity) were purchased from Alfa Aesar (Ward Hill, MA). The lumps were crushed into powders (less than #100 mesh) before being added into the anthracite. The Amorphous Si (þ4) material was sodium silicate (Na2SiO3) solution, and it was provided by J.B. DeVenne Inc. (Berea, OH). The crystallized Si (þ4) was kaolinite powder (KGa-1, Al2Si2O5(OH)4) that was obtained from Washington County, GA. 2.2. Anthracite Pellet Preparation. Anthracite fines were dried at 105 °C overnight to remove the moisture content. The collagen binder (1 g, typically) was dissolved in 12 g of water at 70 °C to form a gelatin solution. Fructose, when added, was dissolved into the water with the collagen binder. When the Sicontaining material was in powder form, it was mixed with the anthracite fines first. If instead the Si-containing material was a

solution (i.e., sodium silicate), it was added in the gelatin solution before being mixed with the anthracite fines. The final mixture was packed into a cylindrical mold (2.86 cm diameter, 4.76 cm long) with 275 kPa (40 psi) pressure applied on both ends. Finally, the pellet was extruded from the mold and cured under ambient conditions. During the curing, evaporation released 10 g of the 12 initial grams of water. At least three anthracite pellets were produced from the same batch for replications. 2.3. Mechanical Strength Tests. At room temperature, both the unconfined compressive strength (UCS) test and the drop shatter test 14 were performed. The UCS protocol employed a Simpson-Gerosa machine, and the drop shatter test results depict the % remaining after dropping a pellet 1.83 meters, 10 times (see Supporting Information for details). The results listed herein are for duplicate analyses, and UCS values for duplicates were generally within 7% of one another, while the drop shatter results were within 1-10 percentage points of one another. 2.4. Pyrolysis. Anthracite pellets were pyrolyzed in a LindbergBlue tube furnace from Thermo Scientific (Newington, NH). A horizontal alumina tube was employed. A slow nitrogen gas flow (∼2 standard cubic centimeters per minute) was used to prevent the anthracite from burning. First, the furnace was ramped up at 3 °C/min to 1400 °C (similar to temperatures in the cupola preheat zone), where it remained for 2 h. Then, the furnace was cooled down to room temperature at 3 °C/min. This 3 °C/min was the fastest rate that this heater could go to these very high temperatures. This is slower than the full-scale conditions in a cupola, where the temperature raises from ambient to 1400 °C within 45 min. 2.5. Characterizations. Scanning electron microscopy (SEM) analysis was performed on an FEI Quanta 200 Environmental SEM. The instrument was operated under low-vacuum conditions (10-103 Pa) using a Gaseous SE detector. The high voltage was set at 20 kV, and the spot size of the electron beam was set at 4 nm. A transmission electron microscope (TEM, Model 2010, JEOL, Tokyo, Japan) was used to determine the morphology and particle size, for electron diffraction. Ambient temperature X-ray diffraction analysis (XRD) patterns were obtained via a PANalytical X’Pert Pro MPD diffractometer. The diffraction patterns were collected for two-theta between 5 and 70 degrees. The pellets were crushed into powders by the ball mill, and the powders were placed into a special aluminum sample holder. Thus, the powder XRD pattern 3063

dx.doi.org/10.1021/es102684k |Environ. Sci. Technol. 2011, 45, 3062–3068


ARTICLE

Figure 1. Crystal structure changes of pellets that contained 100 g of anthracite fines, with various Si-containing materials (10 g), bindered with 1 g of collagen, after pyrolysis at 1400 °C.

offered a bulk measurement and represented the average situation in the pyrolyzed pellets.

3. RESULTS AND DISCUSSION 3.1. Effect of Pellet Composition on Mechanical Strength. Anthracite pellets were prepared with an array of recipes and protocols. These included 100 g of anthracite fines, 1 g of collagen, 0-1 g of fructose, and 0-10 g of silicon, kaolinite, or sodium silicate. The UCS and drop shatter retentions have been presented in Table 1. The recipes that employed 10 g of silicon metal powder and 1 g of fructose achieved the most favorable combination of UCS before and after pyrolysis (1300 kPa before and 720 kPa after pyrolysis). Sodium silicate addition (20 g) offered considerable UCS (2344 kPa) before pyrolysis but only 28 kPa after pyrolysis. The pellets that include kaolinite offered

scant strength after pyrolysis. For all three Si sources, less Si corresponded to lower strengths. 3.2. Crystal Structure Changes within Anthracite Pellets with Different Si-Containing Materials. As detected by the XRD (Figure 1), the raw anthracite fines tested in this study contained muscovite [KAl2(AlSi3O10)(F,OH)2], kaolinite, and quartz (SiO2). These materials also contained crystallized Si (þ4). After pyrolysis, muscovite and kaolinite were converted into mullite (Al6Si2O13). Pyrolysis of the anthracite fines alone yielded no SiC. When 10 g of kaolinite powder was bindered to 100 g of anthracite fines with 1 g of collagen, kaolinite was converted mostly to mullite after pyrolysis. This kaolinite-anthracite combination did not yield SiC. With sodium silicate addition, pyrolysis did yield beta-SiC (3CSiC), per XRD (Figure 1e,f). The pyrolysis condition incurred a net 3064



ARTICLE

Figure 2. SEM images of anthracite pellets with different Si-containing materials after the pyrolysis at 1400 °C. Pellets with 10 g of kaolinite powders (a and b); pellets with 10 g of solid sodium silicate (c and d); pellets with 10 g of silicon powders (e and f). Insets in each picture are the EDS spectra from respective locations.

reduction of the silicon from Si (þ4) to Si (0). The amorphous structure of the sodium silicate facilitated this reaction. The formation of SiC after pyrolysis became greater when silicon powder (10 g) was used as the Si source (Figure 1g,h). After pyrolysis, the silicon metal XRD response was depleted. Strong diffraction peaks from 3C-SiC were observed. Zero-valent silicon and carbon reacted with each other to form SiC. Possible reaction pathways for this reaction have been discussed in the Supporting Information. 3.3. After Pyrolysis: The Morphologies of Anthracite Pellets Containing Various Si-Containing Materials. SEM images (Figure 2a) showed that kaolinite powders were sintered into larger ceramic features that spanned across neighboring anthracite particles. Figure 2b shows many spheres on the anthracite surface generated from the alteration of kaolinite to mullite and silica. With sodium silicate, the SEM images of the pyrolyzed products exhibited spongelike ceramic structures (Figure 2c,d) that were reminiscent of their sodium silicate precursor structures. The ceramic blocks were not SiC, as indicated by energy-dispersive spectrometry (EDS; Figure 2d). No SCNW appeared in the SEM images.

When 10 g of silicon was included, the pyrolysis created a labyrinth of SCNWs. These nanowires grew to 20 μm generally and sometimes to 100 μm. The nanowires altered the color of the pellets from black to light green. These nanowires were often 20-50 nm in diameter, and they were coated by an amorphous thin layer of silicon oxide (about 2 nm thick) as shown in the high resolution TEM (HRTEM) image in Figure 3. Figure 3 also indicates the nanowires were highly crystallized. The space between two planes in Figure 3 was about 0.25 nm which is concurrent with the calculated value in the 3C-SiC structure (0.251 nm). The stacking faults confirmed that the nanowires were grown by stacking the (111) lattice plane in the [111] direction.15-17 3.4. Discussion on Different Binding Mechanisms. The differences in morphology and crystal structure confirmed that each of the three Si-containing materials exhibited a distinct binding mechanism, relative to their ability to hold the anthracite fines together after pyrolysis. The redox level of the silicon significantly affected the development of the binding systems. 3065



ARTICLE

Figure 4. Different amounts of powdered anthracite and silicon: effect on the unconfined compressive strength (UCS) after pyrolysis at 1400 °C.

3.5. Improving SCNW Binding Strength by Reducing Anthracite Grain Size. The anthracite pellets could be packed Figure 3. HRTEM of the SCNW formed at 1400 °C within the anthracite pellets with 9% silicon powder.

The crystallized Si (þ4) in kaolinite could not react with the anthracite to form SiC under the conditions tested in this study. Instead, the kaolinite powders became sintered into larger ceramic features, via mechanisms that are similar to the process of making porcelain from kaolinite. Also, numerous spheres formed on the anthracite surface, indicating that the kaolinite alteration product experienced very high surface tension on the anthracite surfaces when it was still in the liquid phase at high temperature. The contact angles were almost 180° for the spheres. That is, the liquid-air surface tension was almost equal to the liquid-solid surface tension. This achieved negligible adhesive strength between the liquid and the solid, in this case, between the kaolinite alternation product and the solid anthracite surface. Therefore, the strength of the pyrolzyed anthracite pellets was very weak. The amorphous Si (þ4) in the sodium silicate solution was able to form some SiC with the anthracite at 1400 °C (per XRD). However, SEM showed no SCNW. For Si (þ4), a considerable reduction was needed for silicon to enter the vapor phase (as SiO), so as to participate in growing SCNWs via chemical vapor deposition.18,19 Although SiC (not as nanowires) can be produced at temperatures lower than 1400 °C from some amorphous silica materials,20-22 temperatures higher than 1400 °C were required to produce SCNWs from amorphous silica.23,24 Furthermore, there was no evidence that the SiC generated from the amorphous Si (þ4) provided binding strength for the pyrolyzed anthracite pellets. The ceramic bridges that held the anthracite particles together were most likely the remnants of the solid sodium silicate, as indicated by EDS. The zerovalent silicon powders completely reacted with the anthracite at 1400 °C and grew an abundance of SCNWs from the anthracite surfaces. SiC has very high mechanical strength and resistance to high temperatures.25,26 Furthermore, the SCNWs were grown from the anthracite surface and attached to the anthracite firmly through the formation of SiC between carbon and silicon.

more tightly by decreasing the grain size. In this way, the gaps between the anthracite particles became small enough for the SCNWs to directly connect two neighboring anthracite particles. Figure 4 shows the postpyrolysis strength of anthracite pellets that were made from different contents of powdered anthracite and silicon. With the same silicon content, anthracite pellets made totally from the powdered anthracite were 5 times stronger than the anthracite pellets made totally from the original anthracite fines. With partial powdered anthracite, the postpyrolysis strength of the anthracite pellets was enhanced. The anthracite pellets made from 50% powdered and 50% original anthracite were only slightly weaker than those made from 100% powdered anthracite. Figures 2e and 5 exhibited these distinctions incurred by anthracite particle size. When the anthracite pellets were made from the original anthracite fines only, there were many relatively large void spaces between the anthracite particles. SCNWs often were not long enough to span from one anthracite grain to another. In this case, SCNWs could only bind the anthracite particles at the vicinities of their touching points (Figure 5a,c). However, the void spaces between the anthracite powders were much smaller (Figure 5b). The resultant SCNWs mostly spanned between the grains (Figure 5d). Even when the anthracite source included 50% fines and 50% powder, the smaller grains could fit within the voids between the larger grains to make the void spaces smaller (Supporting Information). For anthracite briquettes used in the cupola, the desirable UCS would be about 700-1300 kPa, which would be required to accommodate the burden of scrap iron and carbon that falls on top of it (Supporting Information). On the basis of the work that has been presented herein, these desirable strengths could be achieved when half of the anthracite was powdered, and 5 g of silicon was included with 100 g of anthracite. Ongoing Penn State work is yet further appraising the most optimal recipes and protocols for making these briquettes for full-scale substitution of coke. 3.6. Full-Scale Test of Briquetted Anthracite Fine Bricks in an Operating Foundry Cupola. Anthracite fines were combined and briquetted with collagen binder, fructose additive, and silicon powders were fed into an operating foundry cupola to replace a part of the foundry coke. All together, 500 pounds of the anthracite bricks were prepared. Each brick was 5.75 in. 3066



ARTICLE

Figure 5. Different morphologies of pyrolyzed anthracite pellets (with 10 g of silicon) made from different grain sizes. (a) Made from anthracite fines (mesh # 10 80). (b) Made from anthracite powders (

Binding Waste Anthracite Fines with Si-Containing Materials as an

Recommend Documents