Read Microsoft Word - 05-145_05_.doc text version

Paper 05-145(05).pdf, Page 1 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

Relationship Between Tensile Properties and Matrix Microstructure in Austempered Gray Iron

K.B. Rundman, J.R. Parolini, D.J. Moore

Michigan Technological University, Houghton, Michigan Copyright 2005 American Foundry Society ABSTRACT Gray cast iron cylinder liners (nominally 4.0 CE) were used to produce material for the production of tensile bars austenitized at two temperatures, 871°C (1600°F) and 927°C (1700°F), and austempered for a variety of temperatures, 260-375°C (500700°F), and times (1 or 4 hours). Tensile tests were performed according to ASTM-E8 and the stress and strain data were recorded. Tensile results showed significant increases in fracture stress and ductility to fracture over normal grades of gray cast iron. These beneficial property increases were shown to correlate well with the fraction of austenite present and the particle size associated with the ferrite in the ausferrite microstructure measured previously by X-ray diffraction (XRD) on similarly treated coupons. INTRODUCTION MATERIAL SELECTION FOR CYLINDER LINER APPLICATIONS A diesel cylinder liner must be able to withstand a challenging environment, one in which the components are constantly subjected to high temperatures and a corrosive environment under a cyclic load. These conditions necessitate that the liner material must have good fatigue properties, excellent dimensional stability and thermal conductivity, as well as be resistant to corrosive liquids and gases. Additionally, the high production volume of cylinder liners dictate that the material be produced relatively inexpensively. Gray cast iron, an extensively used material in cylinder liner applications, has excellent thermal conductivity, good castability and machinability. With the development of more powerful engines it is natural that the liner material requirements would also need to increase. The tensile strength of gray iron can be increased with additions of chromium, copper, molybdenum, and nickel and the dimensional stability at elevated temperatures of gray iron improves with additions of molybdenum, copper, chromium, and manganese (Goodrich, 2003). Another way to obtain increases in tensile strength of gray iron is to subject the material to an austempering heat treatment, in which the ferritic matrix of normal cast iron is replaced by an austenitic matrix (Kovacs, 1994 )(Van Maldegiam, 1987). BASICS OF THE AUSTEMPERING PROCESS Figure 1 is a schematic phase diagram for a gray cast iron which can be used to illustrate the austempering process. Step 1. Heating to the austenitizing temperature, T, selects the carbon content of the matrix austenite, Co, which will be attained on holding at T for 1 ­ 4 hours. Step 2. Cooling rapidly into the metastable ( + ) field to an austempering temperature, TA, allows the matrix austenite of composition Co to transform to the two phase ausferrite product, (0 C) + (C). This transformation is normally complete in 1 ­ 4 hours. The structural change key to this process is the production of high carbon austenite, C, together with a fine-scale dispersion of ferrite within the austenite, commonly known as the microconstituent ausferrite (Rundman, 1988). It can be seen in Fig. 1 that the austenite matrix, C ~ 2 wt. Pct, has an MS (martensite start) temperature much below room temperature, TR. Thus the austenite formed during austempering is stabilized so that the formation of martensite does not occur on cooling to room temperature. The final structure in service is then an fcc matrix with a fine-scale dispersion of bcc ferrite (or ferrite and carbide at low TA). The fine-scale dispersion of ferrite in the austenite is thought to be responsible for the higher strengths in both austempered ductile iron (ADI) and in austempered gray iron (AGI), and the austenitic matrix is responsible for the improved ductility, especially in ADI (Rundman, 1988)(Van Maldegiam, 1987).

Paper 05-145(05).pdf, Page 2 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

Fig. 1

Schematic Fe-C-Si phase diagram.

FRACTURE MODE IN TENSION FOR GRAY CAST IRONS The fracture mode of gray iron is fundamentally different than that in ductile iron due to the near continuous nature of the graphite, and to the sharp tips of the graphite flakes which can act like preexisting cracks. Thus failure occurs by crack initiation at the tips of graphite flakes, propagating first along the graphite flake and then across the intercellular matrix space. Clearly it is this matrix between flakes which gives the strength and ductility to gray cast irons. It follows that austempering will change the matrix and would therefore be expected to affect the tensile properties of the iron. It has been observed previously that both strength and ductility will increase in gray irons as a result of austempering (Kovacs, 1994) but there has been no effort to relate the magnitude of the changes to measurable microstructural parameters. It is the purpose of this work to provide a survey study of the tensile properties of cylinder liner iron, both strength and ductility, and to relate these to the measured microstructural parameters of austenite volume fraction, X, and ferrite particle size, d, both of which had already been documented in this liner material. EXPERIMENTAL PROCEDURE The material provided for this study was in the form of commercially produced gray cast iron cylinder liners. The gray iron was poured at 1349°C (2460°F). The castings were static cast in green sand and shakeout occurred at approximately 538°C (1000°F). The nominal chemical composition (in wt.%) of the iron is as follows: 3.26 C, 2.25 Si, 0.46 Mn, 0.10 Ni, 0.57 Cu, 0.22 Cr, 0.20 Mo, 0.02 Ti, 0.03 P, 0.04 S

Paper 05-145(05).pdf, Page 3 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

Large sections of the liners (5"x5"x ¾") were commercially austempered to simulate as closely as possible what would happen to a real liner. The austempering schedule is indicated in Tables 1 and 2. The selections of these austempering treatments were based upon kinetic and microstructural studies conducted previously at Michigan Tech (Moore, 2003), the results of which are used in this paper. A total of three tensile blanks (each measuring 5"x1"x¾") were taken from each of the 12 austempering treatments. Sub-size standard tensile bars (¼" dia. and 1" gage length) were machined from the blanks according to ASTM-E8. The surface of the gage length was then polished with 600 grit SiC paper so that stress concentrators such as cracks or flaws on the outside surface were minimized. A schematic depicting a finished tensile bar is shown in Fig. 2.

0.250" dia. ± 0.005

0.437" 1.25" ± 0.02 ~3.50"

Fig. 2 Schematic of machined tensile bar (ASTM-E8).

Prior to testing, the average diameter of each gage length was determined by using a micrometer and recording four measurements of the diameter along the length of the bar. The bars were uniaxially loaded with an Instron 4507 Tensile Tester equipped with a 45,000 lb load cell and elongations were measured using a ½" extensometer. The crosshead speed was 0.05 in/min. The load-elongation data were recorded until fracture and then converted into engineering stressengineering strain data. Examples of stress-strain data are shown (later) in Fig. 6. Transverse sections of the gage lengths less than 1/4" from the fracture surfaces were metallographically prepared and high magnification digital micrographs were obtained. Also, five Vickers Macrohardness measurements (10 Kg load, 10 s dwell time) were recorded on each section. These data were then converted into Brinell hardness numbers (BHN) to allow easy comparison with the hardness data reported in the literature. RESULTS AND DISCUSSION All of the tensile and hardness data are given in Table 1. Averages of the tensile data are given in Table 2 together with the results of the X-ray study, measurements of X and d. METALLOGRAPHY Figure 3 shows high magnification digital micrographs (scale bar = 20 µm) that are representative of the structures obtained after a four hour austemper for the six different treatments. From the perspective of the optical microscope, the matrices for all 6 of the austempering treatments appear to be completely transformed into the Stage I product. In actuality, the matrix is only partially transformed at the 260°C austempering temperature, as shown by austenite carbon content data in the next section. The ausferrite structure coarsens as the austempering temperature, and to a lesser extent the austenitization temperature, increases.

Paper 05-145(05).pdf, Page 4 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

(a) T = 871o C, TA = 260o C

(b) T = 871o C, TA = 316o C

(c) T = 871o C, TA = 375o C

(d) T = 927o C, TA = 260o C

Fig. 3 Kinetics of the Ausferrite Transformation

(e) T = 927o C, TA = 316o C

AGI micrographs after a 4-hour austemper.

(f) T = 927o C, TA = 375o C

The following kinetic data is based on the previous foundational research conducted by the authors to determine the austempering times required to produce fully transformed gray iron structures in this liner material (Moore, 2003). Further elaboration is required here to allow an explanation of the tensile results, especially at 260°C. A fully transformed iron would have an austenite content at the metastable boundary (see Fig. 1), a boundary which can be determined by a measurement of the lattice parameter with XRD. The projected (+)/ phase boundary determined by the X-ray measurements made on this austempered gray iron is shown in Fig. 4. The carbon content of iron which was austenitized and water quenched is also shown in Fig. 4. This was obtained from the peak positions of the retained austenite in the quenched patterns. Metastable boundary data points at 316 and 375°C are averages of the data taken including and after the data at 64 minutes. The data point at 300°C is the average of the data taken in the stability study at 300°C (Parolini, 2003). These three data points were used to project a boundary at 260°C of ~2.06 wt. Pct. carbon. The actual data points at 260°C for 1 (64 min) and 4 (256 min) hours show the carbon content in the austenite after quenching from 871 (open diamonds) and from 927°C (closed diamonds) did not reach the projected phase boundary meaning that the transformation had not been completed at 260°C. Dashed curved lines are sketched in to indicate the very slow kinetics of carbon diffusion below 300°C, and especially at 260°C. It can be seen from Fig. 4 that the extent of diffusion after 1 hour at 260°C is such that the carbon content is only about 50% of the way to the projected phase boundary. It is reasonable to assume from this figure that the intercellular regions in these specimens should contain some martensite due to insufficient carbon enrichment of the austenite in these last to transform regions. This is in stark contrast to conclusions made solely by observations with the optical microscope, where after 1 hour at 260°C the matrix appeared to be completely transformed into the very fine ausferrite product. A more graphic picture of the very large difference in the rate of carbon transfer in the austenite is shown in Fig. 5 from specimens austenitized at 871°C. However the most important detail for this tensile work is that the ausferrite transformation is not complete at 260°C for the 1 and 4 hour cycles, a result which has a pronounced impact on mechanical properties.

Paper 05-145(05).pdf, Page 5 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

Fig. 4

Phase boundary for (+)/ determined by XRD.

Fig. 5

Rate of carbon transfer in austenite.

TENSILE AND HARDNESS RESULTS The tensile and hardness data for all bars are reported in Table 1 for the 6 austempering treatments. The ductility is reported as the total strain (elastic plus plastic) multiplied by 100. It has been reported (Akers, 2002) that the nominal fracture stress is ~45 Ksi and the total elongation ranges between 0.5 and 0.75 percent for this as-cast inoculated pearlitic gray iron. It can be seen from TABLE 1 that most of the austempered iron had a larger ductility than the as-cast values, and all of the austempered irons had much greater fracture stress than the as-cast values.

Paper 05-145(05).pdf, Page 6 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

Austempering Treatment (°C)

Test No.

TABLE 1. Ductility (%)

Average Tensile And Hardness Data Ductility Fracture BHN (%) Stress (Ksi) (Kg/mm2) 1 hour Austemper

Fracture Stress (Ksi) 4 hour Austemper

BHN (Kg/mm2)







1 2 3 Average 1 2 3 Average 1 2 3 Average 1 2 3 Average 1 2 3 Average 1 2 3 Average

1.03 1.15 1.12 1.10 1.32 1.28 1.21 1.27 1.55 1.39 1.25 1.40 0.98 1.03 0.77 0.92 1.42 1.16 1.42 1.34 1.00 0.76 1.67 1.14

83.8 89.7 85.4 86.3 78.9 69.5 69.8 72.7 61.8 57.6 54.6 58.0 74.9 66.7 58.7 66.8 85.6 77.5 86.2 83.1 55.1 51.9 61.5 56.2

380 408 388 392 304 319 318 314 253 247 239 246 388 379 375 381 314 309 324 316 261 239 251 250

1.47 1.20 1.16 1.28 1.15 1.45 1.39 1.33 1.45 1.62 1.62 1.56 1.02 1.05 0.99 1.02 1.33 1.17 1.28 1.26 1.82 1.43 1.85 1.70

87.6 84.9 87.7 86.7 76.3 73.7 70.9 73.6 57.9 59.3 59.5 58.9 76.3 82.2 79.5 79.4 76.8 79.7 81.4 79.3 65.7 62.2 65.3 64.4

396 378 390 388 335 322 300 319 252 242 244 246 387 392 368 382 303 -336 319 256 244 246 249

Figure 6 shows stress-strain curves for the specimens that gave the minimum and maximum elongations at each austempering temperature for the 871°C austenitizing temperature. From these data, it appears that this gray iron becomes stronger and less ductile as the austempering temperature is decreased over the range from 375 to 260°C.

Paper 05-145(05).pdf, Page 7 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

100 90 80 70 60 E (Ksi)

375 max 260C max 260C min 316C min 316C max

50 40 30 20 10 0 0.000

375C min



0.009 E (in/in)




Fig. 6

Selected stress-strain curves for T = 871°C.


Hardness and Fracture Stress Vs. Austempering Temperature, TA

The hardness decreases linearly with increasing austempering temperature for both the 871 and 927°C austenitizing temperatures, Fig. 7, in accordance with the coarsening of the ausferrite (see micrographs in Fig. 3 and particle size data in Table 2). This is similar to what is observed in ADI. Figure 7 also shows that, similar to the hardness dependence on TA, the average fracture stress decreases with increasing austempering temperature for the lower (871°C) austenitizing temperature. However, the higher (927°C) austenitizing temperature does not exhibit this same behavior. The maximum fracture stress for the 927°C treatments occurs at 316°C for the 1 and 4 hour austempering cycles. This maximum stress at the 316°C austempering temperature is apparently a result of the ausferrite kinetics being so sluggish at the 260°C austempering temperature (see Figs. 4 and 5), resulting in the ausferrite transformation being incomplete and the intercellular regions containing some martensite. In fact, austempering for a longer time at 260°C would in all probability increase the fracture stress above the value for 316°C, as shown by the substantial increase in the fracture stress (67 to 79 Ksi) as the austempering time is increased from one to four hours.

Ductility Vs. Austempering Temperature, TA

Figure 8 clearly shows that gray cast iron, even in the austempered condition, displays only a marginal amount of ductility (<1.8% total elongation). However, austempering does appear to significantly improve the ductility (as much as a factor of 2-3) when compared to the as-cast inoculated pearlitic gray iron, which ranges from 0.5 to 0.75% (a strain of 0.005 ­ 0.0075). Examination of Fig. 8 also shows that the ductility improves as the austempering temperature is increased. These results are similar to that observed in ductile iron and points to the presence of the fcc austenite phase as the matrix rather than ferrite, and increasing amounts of austenite as TA increases.

Paper 05-145(05).pdf, Page 8 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA


FS 871 C FS 927 C Hardness 871 C

570 520 470



Fracture Stress (Ksi) 70.0

Hardness 927 C



420 370

50.0 320 40.0 270 220 270 290 310 330 350 370 390 Austempering Temperature, deg. C

Fig. 7 Effect of austempering temperature on the fracture stress and hardness.

30.0 250

1.8 1.6 1.4 1.2 % Ductility 1.0 0.8 0.6 0.4 0.2 0.0 250

Ductility in this As-Cast Pearlitic Iron

871/260 871/375 927/316 270 290 310 330 350 370 871/316 927/260 927/375 390

Austempering Temperature, deg. C

Fig. 8 Effect of austempering temperature on ductility.

BHN (Kg/mm2)

Paper 05-145(05).pdf, Page 9 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

Fracture Stress Vs. Hardness

Stress versus hardness behavior can be useful because it relates a mechanical property which requires a significant time commitment to obtain (strength) to a relatively easy measurement to record (hardness). An estimate of the fracture stress can then be interpolated from the hardness. Figure 9 compares the stress versus hardness behavior for this AGI versus data reported in the literature for AGI's of similar treatments and alloy compositions (Kovacs, 1994)(Van Maldegiam, 1987) as well as inoculated pearlitic gray irons (Goodrich, 2003). The AGI studied here compared with the AGI's studied by previous researchers display a much higher fracture stress (~20%) while exhibiting a similar hardness for all comparable austempering treatments. This graph also shows that austempering over the range from 260 ­ 375°C results in much higher fracture stress and hardness values than are possible with even the finest pearlitic gray irons.

100 90 80 Fracture Stress (Ksi) 70 60 50 40 30 20 10 0 150

871/260 871/375 927/375 927/260/4 Kovacs 871/316/2 Van Mald 843/343/2 Van Mald 843/316/2 Inoculated Pearlitic 871/316 927/316 927/260/1 Kovacs 871/371/1 Kovacs 871/260/3 Van Mald 899/316/2 Van Mald 954/316/2

260o C 316o C

375o C







Brinell Hardness, HB

Fig. 9 Fracture stress versus hardness.

EFFECT OF MATRIX MICROSTRUCTURE ON MECHANICAL PROPERTIES TABLE 2 lists the averages of the austenite volume fractions and the effective "particle size", d, of the ferrite in the ausferrite product measured by X-ray diffraction (XRD) experiments performed previously on this AGI (Parolini, 2003). The austenite volume fraction was determined using the direct comparison method. The effective "particle size" was estimated using the Scherrer equation. These procedures and calculations have been used extensively in the past by researchers at MTU, as well as other institutions, and additional information pertaining to the methodology can be found elsewhere (Cullity, 2001)(Rouns, 1991)(Van Maldegiam, 1987).

Paper 05-145(05).pdf, Page 10 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

Austempering Treatment (°C) 871/260 871/316 871/375 927/260 927/316 927/375

TABLE 2. Averages OF X and d in the Fully Austempered Product Austempering Fracture Stress BHN Ductility X Time (Hrs) (Ksi) (Kg/mm2) (%) 1 4 1 4 1 4 1 4 1 4 1 4 86.3 86.7 72.7 73.6 58.0 58.9 66.8 79.4 83.1 79.3 56.2 64.4 392 388 313 319 246 246 381 382 315 319 250 248 1.10 1.28 1.27 1.33 1.40 1.56 0.92 1.02 1.34 1.26 1.14 1.70 0.18 0.25 0.23 0.30 0.39 0.44 0.22 0.23 0.33 0.30 0.41 0.44

d (Å) 101 101 142 144 195 199 102 103 150 145 201 199

Tensile Properties Vs. Ferrite Particle Size, d, and Austenite Fraction, X

Figures 10 ­ 13 illustrate plots of fracture stress, and ductility vs. X and d. It can be seen that, in general, the fracture stress decreases with increasing X and decreases with increasing d while the ductility does the opposite, increasing with increasing X and increases with increasing d. The austenite volume fraction present in the ausferrite has a significant effect on the mechanical properties of AGI. From Fig. 10, the fracture stress (and hardness) decrease with an increasing volume fraction of austenite in the ausferrite product. This leads to the fact that austempering at lower temperatures will result in stronger components due in part to the ausferrite constituent containing less austenite.

100 90 80 Fracture Stress (Ksi) 70 60 50 40 30 20 10 0 0.0 0.1 0.2 0.3 0.4 0.5 Austenite Fraction, X

Fig. 10 Effect of austenite fraction on the fracture stress and hardness.

700 600 500 400 300 200 Fracture Stress Hardness 100 0 Brinell Hardness (Kg/mm )


Paper 05-145(05).pdf, Page 11 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

The amount of austenite present in the ausferrite product should also have a significant impact on the ductility of this AGI. The data in Fig. 11 clearly show that the ductility improves with increasing amounts of austenite from ~1% for 260°C to as high as 1.7% total elongation for 375°C. Again, keep in mind that the 927/260 specimens have not been fully austempered, and therefore, the intercellular regions contain some martensite. As previously mentioned, this as-cast pearlitic iron reportedly exhibits elongations between 0.5 and 0.75%. Extrapolating the trendline in Fig. 11 to X = 0 gives an elongation of 0.72% which is in agreement with these reports, however, many other factors influence the observed ductility, as indicated by the relatively low R2 value of 0.56. The graphitic structure, not the matrix, will in large part determine the amount of elongation in gray irons, austempered or otherwise. Porosity or other casting related defects will also dramatically affect the elongation. No quantitative measurements were performed on the graphitic structure or the possible presence of defects during this study.

1.8 1.6 1.4 1.2 Ductility (%) 1.0 0.8 0.6 0.4 0.2 0.0 0.0 0.1 0.2 0.3 0.4 0.5 Austenite Fraction, X

Fig. 11 Effect of austenite fraction on ductility.

As-Cast Pearlitic Iron

%Ductility = 1.79X + 0.72 R = 0.56


871/260 927/260

871/316 927/316

871/375 927/375

In addition to the amount of austenite in the ausferrite constituent, the size or scale of the ausferrite will also have a significant effect on the fracture stress, ductility, and hardness of this AGI. As might be expected, the fracture stress and hardness increase (see Fig. 12) and ductility decreases (Fig. 13) with increasing fineness of the ausferrite. However, it should be noted that measuring the ferrite "particle size" based on the amount of XRD peak broadening has many inherent problems. In this case, the broadening which occurs has been attributed solely to a decreasing ferrite "particle size." In fact, there are other contributors to XRD peak broadening most notably any strain present in the lattice from the casting process. Therefore, this analysis will predict a smaller particle size than actual but remains useful because the relative magnitudes should still follow in the same manner (lower austempering temperature gives a finer particle size).

Paper 05-145(05).pdf, Page 12 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

100 90 80 Fracture Stress (Ksi) 70 60 50 40 30 20 10 0 75 100 125 150 d (Angstroms)

Fig. 12 Fracture stress and hardness versus ferrite particle size.

700 600 500 400 300 200 Fracture Stress Hardness 175 200 100 0 225 Brinell Hardness (Kg/mm2)

1.8 1.6 1.4 1.2 Ductility (%) 1.0 0.8 0.6 0.4 0.2 0.0 75 100 125 150 d (Angstroms)

Fig. 13 Ductility versus ferrite particle size.




Paper 05-145(05).pdf, Page 13 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

DEVELOPMENT OF EXPERIMENTAL MODELS Multiple linear regression (MLR) analyses (least-squares fits) were performed on the data in TABLE 2 to relate mechanical property data to the heat treatment parameters, and separately, to the matrix microstructure, which is in large part dependent upon the specific heat treatment. In the end, the objectives of these analyses are to determine the relative significance that each of the individual heat treatment parameters and structural features of the matrix have on mechanical properties. It has already been established that the ausferrite transformation in the 927/260 treatments is far from being complete (see Figs. 4 and 5) and the resulting structures contain some martensite, and consequently, these data are ignored in the following regression analyses. MODELING THE EFFECTS OF AUSTENITIZATION TEMPERATURE, AND AUSTEMPERING TEMPERATURE AND TIME ON MECHANICAL PROPERTIES The sensitivity that a mechanical property has on the heat treatment parameters (e.g. T, TA, and tA) is a practical concern to a heat treating operation. From this knowledge, the heat treatment can be tailored to meet the design requirements of the iron. With this in mind, MLR analyses were performed on the data in TABLE 2 (with the exception of 927/260 data) to model the dependence that hardness (BHN), fracture stress (FS, in Ksi), and ductility (total strain at fracture multiplied by 100) have on the heat treatment parameters, the results of which are listed as Eqs. 1-3. The temperatures are in degrees Celsius and the time is in hours.

% Ductility = 0.912 - 0.00058(T ) + 0.002(TA ) + 0.059(t A ) (R = 0.61)


FS = 68.2 + 0.101(T ) - 0.267(T A ) + 0.44(t A )

BHN = 693 + 0.016(T ) - 1.23(T A ) + 0.186(t A ) (R2 = 0.99)

(R = 0.94)


Equation 1 Equation 2 Equation 3

MODELING THE EFFECTS OF AUSTENITE VOLUME FRACTION AND FERRITE PARTICLE SIZE ON MECHANICAL PROPERTIES Following the construct of the previous method, MLR analyses were performed on the data in TABLE 2 (again with the exception of the 927/260 data) to develop mathematical models exploring the dependence that hardness, fracture stress, and ductility have on the austenite volume fraction, X, and the ferrite "particle size", d, in the Stage I product (which should display a Hall-Petch type relationship). The experimental models obtained from the MLR analyses are listed as Eqs. 4-6 with d in Angstroms and X being unitless.

1 (R2 = 0.99) BHN = -73 - 33.5(X ) + 4755 d 1 (R2 = 0.81) FS = -7.37 + 0.42(X ) + 971 d 1 (R2 = 0.49) % Ductility = 0.25 + 1.97(X ) + 5.45 d

Equation 4

Equation 5

Equation 6

Standardized regression coefficients were calculated in order to remove the dimensionality and scale between the variables and allow for determining the relative importance between the particle size and austenite volume fraction. Obtaining standardized coefficients first requires that data in all sets (both predictors and response) are normalized by Eq. 7 after which follow the traditional MLR analyses.

xs =

xi - x


Equation 7

In Eq. 7, the xi are the individual data points (FS, %Ductility, BHN, X, and d), the xs are the corresponding standardized values, x-bar is the average for each data set and s is the standard deviation for each data set. The models developed from the standardized data [ the (*) next to each mechanical property indicating normalized results] are listed as Eqs. 8-10.

Paper 05-145(05).pdf, Page 14 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

1 (R2 = 0.99) BHN * = -0.055(X ) + 0.944 d 1 (R2 = 0.81) FS * = 0.003(X ) + 0.91 d 1 (R2 = 0.49) % Ductility* = 0.99(X ) + 0.33 d

Equation 8

Equation 9

Equation 10

It is of particular interest to examine the coefficients in Eqs. 8 - 10 in view of the particular mechanical properties of hardness, tensile strength and ductility. Equation 8 shows that the relative importance of the austenite volume fraction and the ferrite particle size in determining the hardness is such that the particle size is a factor of 17 more important. This is not surprising in view of the fact that a hardness test is essentially a compression test where the near continuity of the graphite is much less important than it is in tension. The presence of the soft graphite, by the rule of mixtures, only serves to diminish the hardness equally in all specimens because the graphite volume fraction remains relatively unchanged from one specimen to another. An increase in volume fraction of austenite in the amount observed in the experiment as the austempering temperature increases would, of itself, be expected to have only a small effect because austenite is only marginally softer than ferrite. In fact, Eq. 8 shows that increases in X will serve to slightly decrease the BHN. Thus the major contributor to the observed hardness change with heat treatment stems from the microstructural scale observed, it being much finer at low TA than at higher TA. As fine pearlite is much harder than coarse pearlite, so fine ausferrite is expected to be much harder than coarse ausferrite. Equation 9, representing the tensile fracture stress, is quite similar to the BHN results (Eq. 8) in that the scale of the microstructure is the most important factor in determining the fracture stress. On the other hand, Eq. 10, representing tensile ductility, shows relative values of the coefficients describing the effect of X and d are much closer to one another, and that the effect of the austenite volume fraction is more important than particle size of the ferrite. This is intuitively correct because austenite is nominally more ductile than ferrite (Smith, 1981). On the other hand the decreasing ductility with decreasing particle size is a natural consequence of increasing strength observed in all engineering alloys. Of course the relatively low correlation in this data set (R2 = 0.49) likely reflects the scatter normally found in ductility measurements in gray cast iron. CONCLUSIONS 1. 2. 3. 4. Austempering results in substantial improvements in the tensile properties of this gray iron with significant increases in both the strength and ductility (by as much as factors of 2 and 3, respectively) depending upon the particular austempering cycle compared to as-cast nominal values. An optimum heat treatment cycle to produce a quality liner material from this particular gray iron would be T=871°C for two hours, TA=316°C for one hour. Hardness and Fracture Stress are well-behaved with heat treatment conditions as described by Eqs. 1 and 2 with the austempering temperature appearing to play the dominant role in the resulting values. The relative importance that the microstructural scale of the matrix, measured by d, and the amount of austenite in the structure, measured by X, are illustrated to a degree in Eqs. 8-10. It appears that the particle size is much more influential on the hardness and fracture strength than the austenite volume fraction, as evidenced by the large disparity in the standardized regression coefficients (Eqs. 8 and 9). On the other hand the ductility appears to be largely dependent on the matrix composition, increasing with higher amounts of austenite (Eq. 10).

ACKNOWLDEGMENTS The authors would like to express our appreciation to the Department of Education for its financial support and Caterpillar for supplying the cast material. A special thanks is reserved for Dr. K. Hayrynen from Applied Process Inc. for her assistance with the commercial heat treatments and also the numerous helpful comments and suggestions she provided during the course of this work.

Paper 05-145(05).pdf, Page 15 of 15 AFS Transactions 2005 © American Foundry Society, Schaumburg, IL USA

REFERENCES Akers, D., Caterpillar Engineer, Private Communication, (2002). Cullity, B.D., Stock, S.R., Elements of X-Ray Diffraction, 3rd Ed., Prentice Hall (2001). Goodrich, G.M. (editor), Iron Castings Engineering Handbook, American Foundry Society (2003). Kovacs, B., Keough, J.R., "Physical Properties and Application of Austempered Gray Iron," AFS Transactions, . pp. 283-291 (1994). Moore, D.J., Parolini, J.R., Rundman, K.B., "On the Kinetics of Austempered Gray Cast Iron," AFS Transactions, . vol. 111, paper no. 03-110 (2003). Parolini, J.R., MS Thesis-Michigan Technological University (2003). Rouns, T.N., PhD Thesis-Michigan Technological University (1991). Rundman, K.B., Moore, D.J., Hayrynen, K.L., Dubensky, W.J., Rouns, T.N., "On the Structure and Properties of Austempered Ductile Cast Irons," Journal of Heat Treating, vol. 5, no. 2, pp. 79-85 (1988). Smith, W.F., Structure and Properties of Engineering Alloys, McGraw-Hill (1981). Van Maldegiam, M.D., MS Thesis-Michigan Technological University (1987).


Microsoft Word - 05-145_05_.doc

15 pages

Find more like this

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate


You might also be interested in

Australia A4
Microsoft Word - A-0323.doc