Carpenters Stainless Steel Blue Book 26wc

Factors Affecting Scale Adhesion on Steel Forgings J. A. Zitterman, R. P. Bacco, and W. E. Boggs

SUMMARY Occasionally, undesirable "sticky" adherent scale forms on low-carbon steel during reheating for hot forging. The mechanical abrading or chemical pickling required to remove this scale adds appreciably to the fabrication cost. Characterization of the steel-scale system by metallographic examination, x-ray diffraction, and electron-probe microanalysis revealed that nickel, silicon, and/or sulfur might be involved in the mechanism of sticky-scale formation. Laboratory reheating tests were conducted on steels with varied concentrations of nickel and silicon in atmospheres simulating those resulting from burning natural gas or sulfur-bearing fuels. Subsequent characterization of the scale formed during the tests tends to confirm that the composition of the steel, especially increased nickel and silicon contents, and the presence of the sulfur in the furnace atmosphere cause the formation of this undesirable scale.

INTRODUCTION When desirable "sticky" adherent scale occasionally forms on low-carbon steel during reheating for hot-forging applications, mechanical abrading or chemical pickling is required to remove the scale. This adds appreciably to the cost of processing. Characterization of the scale and the metaloxide interface of pieces considered to have undesirable scale adhesion (e.g., Figure 1) was attempted by metallographic examination, x-ray diffraction analysis, and electronprobe microanalysis. Nickel was enriched in the metal at the metal-oxide interface and formed a filigree in the scale in the immediate vicinity of the interface. Silicon and iron were combined to form fayalite (Fe2Si02) in the scale adjacent to the interface. A literature search revealed a number of hypotheses regarding which factors are most important in the formation of desirable, loose, non-adherent scale. Palin l suggested that furnace temperature and heating time determine the adhesion of the scale. He found that adhesion was greater on killed steel than on rimmed steel; he felt that the oxygen content of the atmosphere of a gas-fired furnace had no influence on the adhesion of the scale. Sachs and Tuck2 attempted to for all the parameters that influenced scale growth on commercial steels reheated in industrial furnaces. They suggested that large differences in scale formation were due to a variety of conditions created in industrial furnaces. They further stated that, in addition to these conditions, the adhesion of the scale was influenced by topographic configuration of the scale-metal interface and the composition of the steel, They found that under normal heating conditions with a stoichiometric balance between fuel and oxygen, three layers of 22

Figure 1. Photograph of forging showing undesirable adherent scale.

oxide formed on pure iron heated to high temperatures. These are: wiistite (FeO), comprising about 95% of the total oxide volume; magnetite (Fe304), about 4%; and hematite (Fe203), about 1%. However, they stated that in an atmosphere of low oxygen potential, such as would result from combustion with an excess of fuel (or a deficiency of air), the higher oxide, Fe203, might not form. In most industrial furnaces, the fuel is burned with some excess of air so that the combustion products usually contain enough excess oxygen to permit hematite to form on the outer surface of the scale. Sachs and Tuck also dealt briefly with the effects of steel composition on the formation of scale. They suggest that a concentration of CO or C02 in cavities in the scale might cause cracking and an increase in the scaling rate or prevent collapse of the cavities, healing, and thus slow the rate of oxidation. On the other hand, they felt that the gases might fill the gaps and void!'!, thus preventing healing of the defects and slowing the rate of oxidation. Sachs and Tuck found that in contrast to the scale modifiers carbon and phosphorus, other elements form separate phases. One of the most important of these is silicon, which is present at --{).25 wt.% in killed steels. They state that JOURNAL OF METALS· April 1982

this amount is sufficient to form Fe-Mn-Si pools and stringers in the scale. They further state that with higher silicon concentrations, a film of silica may form at the metal-oxide interface. This film would slow the rate of scaling. Because the oxygen potential of silica is much lower than that of wustite, oxygen diffuses from the iron oxide into the iron to react with the dissolved silicon to form an internal precipitate of Si02. These particles of Si02 are caught up in the advancing metal-oxide interface to form a layer of fayalite (Fe2Si04) under the scale. Nickel and copper are less easily oxidized than iron. As the iron at the metal-oxide interface enters the wustite lattice, the nickel and/or copper are rejected by the scaling reaction and are concentrated at the metal-oxide interface. Boggs6 found that, because nickel does not diffuse rapidly back into the steel, a discontinuous nickel-rich layer is built up at the metal-oxide interface. According to Wagner,7 interfacial instability develops as iron is internally oxidized in preference to the nickel. Eventually a filigree of nickelrich alloy develops, extending for an appreciable distance into the scale. This filigree forms a strong mechanical bond leading to an adherent scale that is difficult to remove in steel finishing. Melford8 studied the effects of residual elements on hot shortness of mild steel. He considered nickel, copper, and tin especially, but he also considered antimony and arsenic. Although his concern was primarily hot shortness and not scale adhesion, the enrichment of residual elements, especially nickel and copper, resulting from the oxidation of commercial steel at temperatures between 1100 and 1120°C is of special interest. By electron-probe micronanalysis, he found that nickel in a steel having a bulk nickel concentration of 0.14% had segregated in the subscale to a concentration as high as 5-6% after oxidation. Copper sulfide (CU2S) was occluded in the scale.

Gesmundo9 reported that the formation of sulfides mixed with oxides is often observed in high-temperature reactions of a pure metal in a complex atmosphere containing both sulfur and oxygen. This effect, he stated, leads to accelerated attack of the pure metal or alloy. This effect appears to be related to the high diffusion rate of cations in the sulfides as compared with that of corresponding oxides. PRELIMINARY STUDIES An experimental program was undertaken to test these hypotheses and to determine what factors are responsible for the descaling problems encountered in hot forging of low-carbon steels. In the preliminary study, specimens A through D of steels with compositions listed in Table I were reheated according to the schedule shown in Table II. These specimens, approximately 89 mm square by 25 mm thick, were reheated in a laboratory muffie furnace with Globar electric heating but with the combustion products from burning natural gas at fixed fuel-air ratios fed into the furnace to produce a furnace atmosphere similar to that of a reheating furnace fired with natural gas. Air-fuel ratios were set to provide either a 5% excess of oxygen or a 5% deficit of oxygen with respect to combustion stoichiometry. Total heating and soaking times of 200 and 400 min were used. The final temperature was 1230°C, approximating the reheating temperature of an industrial forging furnace. After completion of the reheating schedule, the samples were air cooled. Electron-probe microanalysis showed enrichment of nickel in the metal at the metal-oxide interface to about 1% and enrichment of silicon to concentrations of 1-6% in isolated areas of the scale in the vicinity of the metal-oxide inter-

Table I: Composition of Steels Used In Reheat Experiments, wt.% Heat

----r B

C D E

F G H I

J K

Mn 0.61 0.90 0.78 0.75 0.049 0.62 0.62 0.72 0.75 0.70 0.71

C 0.32 0.35 0.29 0.29 0.066 0.05 0.042 0.32 0.29 0.34 0.34

P 0.009 0.Dl5 0.022 0.009 0.001 < 0.001 0.003 0.011 0.012 0.010 0.009

S 0.013 0.23 0.018 0.021 0.007 0.007 0.005 0.025 0.015 0.010 0.015

Si 0.19 0.21 0.20 0.22 0.011 0.08 0.012 0.18 0.16 0.18 0.25

Cu 0.03 0.01 0.04 0.02 0.026 0.02 0.026 0.06 0.02 0.03 0.03

Ni 0.10 0.02 0.05 0.01 < 0.002 0.10 0.10 0.02 0.02 0.13 0.02

Cr 0.22 0.06 0.05 0.02 0.006 0.09 < 0.003 0.003 0.04 0.16 0.09

Mo 0.05 0.01 0.02 0.02 < 0.003 < 0.004 0.007 0.01 0.01 0.02 0.02

AI

N.D.· N.D.· 0.024 0.031 0.021 < 0.002 0.013 N.D.· .D.· N.D.· N.D.·

• Not dotermined

Table II: Schedule of Research Laboratory Reheating Tests at Furnace Temperature of 123O"C

Test Number 1 2 3 4 5 6 7 8

Time in Furnace, min 200 200 200 200 400 400 400 400

Heat A B C D C D C

D

·With respect to stoichiometric combUltion Hi'i,

95~

JOURNAL OF METALS· April 1982

% Deficiency or Excess of 02 in Furnace Atmosphere· +5 +5 +5 +5 +5 +5 -5

-5

Major FeO FeO FeO FeO FeO FeO FeO FeO

Scale Constituents Minor Trace FesO. Fea04 FeaO. FeaO. FeaO. FeaO. FeaO. FeaO.

Fe20 a F820a Fe20 a Fe20 a Fe20 a F820a Fe20S Fe20 S

of the amount of 0, required and +5 i. 105"< ofthe 0, required for theorolically complete combustion).

23

face. Results of x-ray diffraction analysis are shown in Table II. Each of the scales contained a major concentration of wllstite, a minor concentration of magnetite, and a trace concentration of hematite. No fayalite was detected. The silicon detection by electron-probe microanalysis may have been present as fayalite or crystalline silica at less than the 5% detectability limit for x-ray diffraction analysis or may have been present as amorphous silica.

STUDIES IN SULFUR-CONTAINING ATMOSPHERES Since steel used to forge the piece shown in Figure 1 was heated in an industrial furnace using No.4 fuel oil containing

Figure 2. Optical photomicrograph showing area used for characteristic x-ray imaging, test no. 19, Heat F, 100x.

1.1 % sulfur, a second series of reheat tests was run in the manner of the first series, except that 668 ppm S02 was introduced into the furnace atmosphere during heating to simulate the furnace atmosphere resulting from burning No. 4 fuel oil. Many of the samples the water quenched after heating, to stop air oxidation of the specimen during cooling. Of the 20 samples reheated in the second series shown in Table III fayalite was detected by x-ray diffraction analysis as a trace constituent in the adherent scale of 13 samples. As shown in Table IV, the heats on which detectable quantities of fayalite did not form were sample G, reheated in oxidizing and reducing atmospheres, and samples H, I, and K with the indicated silicon and nickel concentrations in the bulk steel, heated in reducing atmospheres. In addition to qualitative and quantitative electron-probe microanalysis, characteristic x-ray imaging was performed on preselected areas to show the distribution of specific elements. Optical photomicrographs of the approximate areas used for x~ray imaging are shown with photographs of the images. Figure 2 shows a optical photomicrograph of the approximate area used for characteristic x-ray imaging for one of the steels studied. Photographs of the distribution of iron, silicon, nickel, and sulfur at a grain boundary in the metal at the metal-oxide interface are shown in Figure 3. Figure 4 shows the elemental distribution in an area of the scale near the metal-oxide interface of the same sample. The characteristic x-ray images for virtually all the samples studied in Table III shows: 1. Enrichment of nickel in the steel at the metal side of the metal-oxide interface, in grain boundaries of the metal matrix, and in the filigree structure in the scale. 2. The association of iron and silicon as fayalite as a distinct phase (determined by x-ray diffraction analysis) in grain boundaries of the metal matrix, at the metal-oxide interface, and in the grain boundaries of the scale. 3. The association of iron and sulfur as FeS (determined by x-ray diffraction analysis) in the grain boundaries of the metal matrix, at the metal-oxide interface, and in the grain boundaries of the scale.

Table III: Schedule and Results of Samples Heated 400 Min at 2250°F in Atmosphere Containing 668 ppm S02

Test Number Heat 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

D

C D

C F G

F G I H

H J K J K D

C I H

% Deficiency or Excess of O2 in Furnace Atmosphere

+5 +5 -5 -5 +5 +5 -5 -5 +5 +5 -5 -5 +5 +5 -5 -5 -5 -5 -5 -5

Cooling Method* W.Q. W.Q. W.Q. W.Q. W.Q. W.Q. W.Q. W.Q. W.Q. W.Q. W.Q. W.Q. W.Q. W.Q. W.Q. W.Q.

A.C. A.C. A.C. A.C.

Major FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO

Loose Scale Minor

Trace

Major

Fea04 Fe304 Fe304

Fe304 Fe304 Fe304 Fe304 Fe30 4 Fe304 Fe304

Fe304 Fe30 4 Fe30 4 Fe304 Fe304

Fe304 Fe 304-U** U U U U U Fe 3 04-U U

Fe304 U U U Fe304 + U

FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO FeO

Adherent Scale Trace FezSi04 FezSi04 FezSi04 + FezOa FezSi04 + Fe Z03 FezSi04 Fe2Si04 + Fe203 + Fe304 Fe304 + U FezSi04 + Fe20a Fe2Si04 + Fe304 Fe203+ Fe304 Fe203 + FezSi04 Fe2Si04 FezSi04 FezSi04 + Fe304 FezSi04+ Fez03 + Fe304 Fea04 Fe304

*W.Q. = Water quenched, A.C. = Air cooled **U =: Unidentified x·ray diffraction pattern

24

JOURNAL OF METALS· April 1982

4. The association of iron, silicon, and sulfur in what appears as a complex compound of the elements, but is probably a mixture of iron sulfide (FeS) and iron silicate (Fe2Si04). Even though fayalite was not detected by x-ray diffraction analysis of the steels shown in Table IV, there is reason to suspect that these iron sulfide and fayalite compounds were present in quantities below the 5% detectability limit of this technique, Small amounts of iron sulfide of

fayalite are evident in characteristic x-ray images of cross section of these specimens. CONCLUSIONS

It can be concluded that the presence of nickel contributes to the formation of an undesirable adherent scale. The presence of sulfur in the furnace atmosphere enhances the oxidation of iron and silicon. Grain-boundary attack by

Table IV: Reheating Tests in Sulfur-Bearing Atmosphere Scale Showed No Detectable Fayalite % Deficiency or

Test Number 18 20 23 24 28 31 32

Heat

------cr G

H

K H

%Si 0.012 0.012 0.16 0.18 0.25 0.16 0.18

%Ni 0.10 0.10 0.02 0.02 0.02 0.02 0.02

Excess of 02 in Furnace Atmosphere +5 -5 -5 -5 -5 -5 -5

Sulfur Dioxide in Furnace Atmosphere, ppm 668 668 668 668 668 668 668

Figure 3. Characteristic x-ray Images showing elemental distribution at grain boundary in test no. 19, Heat F, 500 x.

JOURNAL OF METALS· April 1982

25

sulfur exposes more silicon to oxidation and to reaction with iron oxide to form fayalite. It therefore appears that the concentrations of nickel and silicon in the steel and the concentration of S02 in the reheat furnace atmosphere must be controlled to permit reheating of low-carbon steels for hot-forging applications without formation of sticky adherent scales.

References 1. G. M. Palin, STAL 8,119651, pp. 677-670. 2. K. Sachs and C. Tuck, Iron and Steel Institute Publication III, 119681, pp. 1-17. 3. K. Sachs and G. T. F. Jay. Journal of Iron and Steel Institute, 196, 119591. pp. 34-44. 4. H. P. Haast;,r, W. Meischsner, W. Luckerath, and J. Williams, Thyssenforschung, 4,11969), pp. 127-135. 5. N. G. Vannerberg and I. Svedung, Corrosion Science, 11, 119711, pp. 915-927. 6. W. E. Boggs, "The Role of Structural and Compositional Factors in the Oxydati on of Iron and Iron-Based Alloys," in High Temperature Gas-Metal Reactions in Mixed Environment, edited by S. A. Janson and Z. A. Foroulis, Metallurgical Society of AIME, (1973), pp. 84-128. 7. C. Wagoer, J. Electrochem, Soc" 103, 119561, pp. 571-580. 8. D. A. Melford, J. Iron and Steel Institute, (19621, pp. 290-299. 9. F. Gesmundo, Oxidation of Metals, 13, (19791. pp. 237-244.

ABOUT THE AUTHORS

J. A. Zitterman, Research Engineer, Analytical Chemistry Division, U.S. Steel Research Laboratory, 125 Jamison Lane, Monroeville, Pennsylvania 15146. Mr. Zitterman received his BS in chemical engineering from the Pennsylvania State University in 1947. He has been involved in studies of the effects of oxidation, corrosion, and diffusion in steel and titanium alloys.

R. P. Bacco, Research Engineer, Analytical Chemistry Division, U.S. Steel Research Laboratory, 125 Jamison Lane, Monroeville, Penn7 sylvania 15146. Mr. Bacco received his BS in materials science from the University of Pittsburgh in 1973. He has been involved in the materials analysis techniq!Jes of transmission and scannihg electron microscopy, electron microprobe, x-ray spectrometry, and x-ray diffraction analysis.

W. E. Boggs, Research Consultant, Basic Research, U.S. Steel Research Laboratory, 125 JamiSon Lane, Monroeville, Pennsylvania 15146. Mr. Boggs received his BS in chemistry in 1949 and his MS in 1962 from Carnegie-Mellon University. He is the author or co-author of experimental and/or theoretical papers and research reports in various aspects of gas-metal reactions, particularly the oxidation and scaling of iron and iron-based alloys.

Coming in May

Journal of Metals • Aluminum plant implementation of electromagnetic casting • Analytical methods for calcination control and alumina properties • Titanium technology • Steelmaking technology Plus:

Figure 4. Characteristic x-ray Images showing elemental distribution In the scale, test no. 19, Heat F, 500x.

26

Class of 1982-Seniors in Metallurgy/Materials Science

JOURNAL OF METALS· April 1982