Printer Friendly

Optimizing SIC addition for cupola melting: one metalcasting company set out to determine the effects of silicon carbide purity on cupola melting.

Like 60% of the liquid iron produced in the U.S. in 2007, the alloy used to make castings at Grede Holdings LLC's Iron Mountain, Mich., facility is melted in a cupola. About 50% of cupola shops nationwide use steel as their primary charge material. This requires the addition of considerable amounts of carbon and silicon to the melt to produce quality iron.

One common material used to introduce carbon and silicon to ferrous melts is the silicon carbide (SiC) briquette. According to a group of Grede researchers and others, the briquettes generally contain either 36% or 65% SiC and variable amounts of coal or high-sulfur petroleum coke and sand.

"The 36% SiC contains well over double the amount of impurities as the 65% SiC per unit weight of contained SIC," write Adam Buchcuski, Iron Mountain plant manager, Brent Buchcuski, Iron Mountain melt superintendent, Greg Jarski, Iron Mountain plant metallurgist, and James Cree, Grede New Castle plant metallurgist, in their paper "Effects of Varying SIC Purity on Cupola Performance." "Many of these differences can affect cupola operation. The complexity of the effects and the difficulties in making meaningful measurements has led to great confusion in assessing their relative values."

Using funding from the American Foundry Society (AFS), the researchers performed an in-plant study at Grede Iron Mountain to determine the precise differences between two common forms of SiC.


What is the relative value of 36% SiC vs. 65% SiC in terms of melt yield, elemental recovery and melt efficiency?

1 Background

The use of SiC as an alloying agent in cupola melting can produce significant cost savings and reductions in carbon monoxide and carbon dioxide emissions. To achieve the maximum benefits, the differences in the performance of 36% SiC and 65% SiC need to be better understood. Because of the complexity of the two materials, each may perform more effectively in some cupolas but not in others.

Recent research has provided greater understanding of some aspects of the role of SiC in cupola operations, but the work has not specifically examined the differences between 36% SiC and 65% SiC. This study builds on the previous work to provide the information needed by metalcasting facilities to maximize the benefits of the materials in their cupola operations.

2 Procedure

The researchers measured the carbon, silicon, manganese and sulfur concentrations of the selected metallic charge materials. Steel thickness largely determines the amount of iron oxide that will be generated in the cupola, which leads to silicon loss. The frag steel used in the cupola charge for these trials was 0.125-0.25-in. thick.

Coke composition and size were measured, as well as the composition of the limestone. The cupola was 233 in. in height with an unlined hearth diameter of 72 in. and a well depth of 29 in. Prior to the cupola's weekly startup, the coke bed height measured 42 in., resulting in an overall coke column height of 71 in. The iron dam height was 7.5 inches.

The two grades of briquetted SiC, coke and limestone were analyzed by both vendors and an independent commercial lab (Auburn Analytical Labs). All other charge materials were analyzed by vendors only. Briquetted SiC was examined for silicon carbide, combined carbon, free carbon, sulfur and moisture. Coke was analyzed for carbon, ash, moisture, volatile matter and sulfur. Limestone was tested for CaO, MgO, [SiO.sub.2], [[Al.sub.2]0.sub.3], total carbon, lesser constituents and loss on ignition.

The researchers established a method of adding carbon and silicon outside the cupola to correct iron that was out of specification. Corrections to iron composition were made in a holding furnace or ladle. The critical cupola output responses during the performance of each SiC trial were measured and recorded.

The following cupola operating data acquired during the melt trials were used for the analysis of the results:

* Cupola charge records and weights of all charge materials for every individual charge at each point in time (typical time between cupola charges was three to five minutes).

* Continuous electronic cupola data, including blast rate, blast temperature, windbox back pressure, oxygen enrichment flow rate and spout iron temperature.

* Comprehensive spout slag analyses every 15 minutes, including [Si0.sub.2], CaO, [[Al.sub.2]0.sbu.3], MgO, MnO, FeO, [TiO.sub.2] sulfur and carbon.

* Cupola spout/trough iron temperature by continuous infrared pyrometer and immersion pyrometer (target spout iron temperature was 2,700F [1,482C]).

* Cupola thermal analysis by tellurium cups for carbon, silicon and carbon equivalent every five minutes with matching spout and holding furnace spectrometer samples.

* Stack gas analyses, including %CO, %[CO.sub.2] and temperature every 15 minutes.

* Tuyere back pressure and temperatures of water-cooled cupola shell measurements.

* Climatic data from the Kingsford (Iron Mountain), Mich., National Weather Service station every hour.

* Cupola iron thermal analysis by plain cups for metallurgical thumbprint every 15 minutes (eutectic point was the primary metric).


Slag analyses were a crucial component of the mass balances used to calculate melt yields and elemental recovery. Slag samples were collected in real-time but analyzed later by an outside laboratory. A time lag of 45 minutes between cupola charging and spout iron was used for the data analysis.

3 Results and Conclusions Figs. 1 & 2 show high correlations between the grade of SiC and both the melt yield and the slag weight as a percentage of charge weight. The higher purity 65% SiC offered a yield increase of 1.7% over 36% SiC. A review of the standard deviation differences showed the increase was statistically significant.

The performance of 65% SiC was superior to that of 36% SiC when gauged by melt yield and slag weight. The mass balance based on calcium and corroborated by the zero degrees of freedom for carbon and other elements (the carbon outputs/inputs ratio was 105.3% before limestone carbonates and 101.3% after limestone carbonates) shows that the greater impurity (gangue) content of 36% SiC relative to 65% SiC makes no contribution to melt yield. If the impurity content of SiC is defined as all of its solid constituents plus moisture that are deducted from the contained SiC and free carbon, then the 1.1% difference in slag weight as a percentage of charge weight between the two grades of SiC closely approximates the difference in impurity contents added between the two grades.



It is important to note the 0.6% discrepancy between the 1.7% greater melt yield and 1.1% lesser slag weight as a percentage of total charge weight for 65% SiC relative to 36% SiC. This distinction between melt yield and slag weight as a percentage of total charge weight is necessary because cupola slag has no solubility/capacity for carbon other than un-dissolved particles of coke or briquetted SiC, while spout iron does have capacity to pick up carbon, including the free carbon contained in both the coke and briquetted SiC.

Comparing carbon recovery with and without the free carbon in SiC included as a carbon input illustrates the relative propensities for carbon pickup from either the coke or SIC (Figs. 3 & 4). Fig. 3 clearly indicates greater carbon recovery for 65% SiC when excluding free carbon in the SiC as an input, while Fig. 4 indicates approximately equal carbon recoveries for the two grades when including free carbon in the SiC as an input. The differences shown by Figs. 3 & 4 taken together then predict and account for the additional 0.6% greater melt yield for 65% SiC that is not accounted for by the difference in slag weight percentage alone.

The excellent melt yields for these trials are a reflection of the minuscule iron oxide contents of the slags, which average less than 2% of the total slag weight. Furthermore, since the slag weights are small percentages of the total charge weights, the resultant iron losses are exceedingly small at approximately 0.1% of the total iron charged for both grades of SIC.

The two grades of briquetted SiC had approximately equal performances when gauged by all performance metrics other than melt yield and slag weight. This was likely due to the mass balance based on calcium standardizing all metrics other than melt yield.

Calcium was used as the basis for the mass balance for the following reasons:

* The calcium input derives in substantial amounts from several different sources. Although the calcium inputs are insubstantial relative to the iron content, they are substantial relative to all elements other than iron.

* The sources of calcium for both grades of the briquetted SiC are the same, so even if the actual measured amounts of calcium entering the cupola are not totally accurate, the measurement errors would apply equally to both grades of SiC. The relative differences in calcium introduced by each grade of SiC would be a valid metric.

* It can be assumed 100% of all calcium entering the cupola exits as slag.

Long running steady-state melt conditions for the two grades of SiC were desired because of the assumption that, relative to shifting iron chemistry, transitioning from (flushing out) the slag composition in the cupola takes much longer when switching from one grade of SiC to another. However, this assumption was challenged by the demarcations between SiC grades in Figs. 1 & 2, as both graphs indicate abrupt changes in short periods of time when changing from one grade of SiC to another.


Breaking down the industry's latest research papers is as easy as 1-2-3:

"Effects of Varying SiC Purity on Cupola Performance"

Adam Buchcuski, Brent Buchcuski and Greg Jarski, Grede-Iron Mountain; James Cree, Grede-New Castle

1 Background--To achieve the maximum benefits of silicon carbide (SiC) as an alloying agent in cupola melting, the differences in the performance of 36% SiC and 65% SiC must be better understood.

2 Procedure--Consecutive cupola tests, each with a unique combination of SiC and coke, were run over a week's time at Grede Holdings LLC's Iron Mountain, Mich., facility.

3 Results and Conclusions--For overall melt yield and reducing slag generation, high purity (65%) SiC outperformed 36% SiC. No statistically significant differences were found for other performance metrics.

The research summarized in this article was overseen by AFS Division 8 Melting Methods and Materials and funded by AFS. The division would like to acknowledge the support of Sy Katz, Katz & Associates, West Bloomfield, Mich., and Bill LaFramboise, proprietor of Auburn Analytical Labs, Midland, Mich.


For an additional article on SiCin melting, visit

COPYRIGHT 2012 American Foundry Society, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:TESTING 1-2-3
Publication:Modern Casting
Date:Jun 1, 2012
Previous Article:Regulation, taxes & reform: what metalcasters need to know: in a year when the economy and control of the White House are up in the air, numerous...
Next Article:What would Custer do?

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters