Engineering Fractun Mechanics Vol. 13, pp. BSI-864 Pergamon Press Ltd., 1980. Printed in Great Britain
EFFECT OF ELECTROSLAG REFINING ON THE FRACTURE TOUGHNESS AND FATIGUE CRACK PROPAGATION RATES IN HEAT TREATED AISI 4340 STEEL MOHAN G. HEBSUR, K. P. ABRAHAM and Y. V. R. K. PRASAD Department of Metallurgy, Indian Institute of Science, Bangalore 560012, India Abstract-The AISI 4340 steel has been electroslag refined and the improvement in mechanical properties has been assessed. Electroslag refining (ESR) has improved tensile ductility, plane strain fracture toughness, Charpy fracture energy, and has decreased fatigue crack growth rates. The KI, values for the ESR steel are nearly twice those estimated in the unrefined steel and higher than those obtained in the vacuum arc remelted steel. Fatigue crack growth rates in region I and in region III are found to be decreased considerably in the ESR steel, while they are unaffected in region II. Measurements on heat treated samples have shown that the ESR steel has a better response to heat treatment. Both the suggested heat treatments namely austenitizing at 1470-l I40 K as well as the conventional heat treatment of austenitizing at 1140K have been followed. The improvement in the mechanical properties of ESR steel has been explained on the basis of removal of nonmetallic inclusions and reduction in sulfur content in the steel.
INTRODUCTION for sophisticated applications are produced by secondary remelting processes like vacuum arc remelting WAR), vacuum induction melting (VIM) or electroslag refining (ESR). The ESR technique is comparatively a low cost and simple process of producing steels with improved cleanliness, better chemical homogeneity and reduced sulfur content[l]. In this process, the steel in the form of consumable electrode is remelted under a cover of molten synthetic slag of high reactivity. In addition to the removal of non-metallic inclusions, efficient desulfurization occurs because of high basicity of the slag and high temperature prevailing at the tip of the melting electrode. The aim of the present investigation is to electroslag refine the AISI 4340 steel and evaluate the improvement in its mechanical properties particularly the fracture toughness and fatigue crack propagation rates. Both these parameters are known to be sensitive to the presence of non-metallic inclusions in the steel. AISI 4340 steel is chosen for this investigation because of its wide industrial applications, e.g. as in aircrafts and automobiles. In this investigation, the approach has been to compare the data on the ESR steel with unrefined steel and wherever possible also with VAR steel. The studies have been made on the heat treated material so that the data will have practical utility. HIGH QUALITY steels
EXPERIMENTAL Forged rods of commercial AISI 4340 steel (referred to as “unrefined” steel hereafter) were electroslag refined using a synthetic slag mixture containing CaF2 : CaO : A&O3(60: 20 : 20) in a laboratory scale ESR unit. In brief, a consumable electrode of the steel to be refined was immersed in a molten slag of the above composition and melted with the help of the heat supplied by the current passing through the molten slag of high resistivity. The molten droplets forming at the electrode tip fall through the refining slag and form a metal pool which solidifies in a water cooled mould. A current of 8OOA and a voltage of 20 V were found to give satisfactory ingots. The chemical analysis of the material before and after refining is given in Table 1. The microstructural details such as grain size and inclusion content in the refined and unrefined ingots were recorded using a Quantimet. Mechanical testing was conducted on forged and heat treated specimens. On the basis of the available literature[2-41, two different types of heat treatment were given to the specimens: (i) Austenitizing at 1140K for 1 hr and oil quenching followed by tempering in the temperature range 37O-g70K for 2 hr. (ii) Austenitizing at 1470K for 1 hr, step cooling to 1140K holding for f hr and oil quenching followed by tempering in the range 370-870 K for 2 hr. 851 EFM Vol. 13, No. 4-J
MOHAN G. HEBSUR et ai.
Table I. Chemical analysis (in wt.%) of AK1 4340 steel before and after ESR
Unrefined ESR Ingot Top ESR Ingot Bottom
Ni __~..__. _ 1.34
The tensile and Charpy impact testing at room temperature were done using standard ASTM specimens. The plane strain fracture toughness (K,) was measured on a 25.4mm thick compact tension specimen following the procedure recommended~51 in ASTM 399. Fatigue precracking was carried out on a 50-ton MTS machine using haversine load cycle (RlOOflOO kg) at 1OCPS. The precracked specimens were pulled in tension at a crosshead speed of 1.0 mmlmin and the displacement was measured using clip-on gage. The load vs displacement curves were recorded on X-Y recorder. Fatigue crack growth rates were measured on a 25.4 mm thick compact tension specimen following the recommendation ASTM E647-78T. The specimens were first precracked using haversine load cycle (lOOO/lOO kg) at a frequency of 40CPS. Further crack growth in the precracked specimen was measured by fatigue loading with load control under sinusoidal tension at a frequency of 10 CPS. Two different values of stress intensity ratio (R = Kmin/Kmax where K,,, and Krni,,are the maximum and minimum stress intensities during each cycle) equal to 0.1 and 0.5 were used. The fatigue crack length was continuously measured using a traveIling microscope. The printing of moire grid lines on the specimen surface and illuminating it with a stroboscope facilitated measurement of crack length. The average crack extension between two consecutive measurements was approximately 0.025 mm. All the fracture surfaces were examined using scanning electron microscope with a view to determining the mode of fracture. RESULTS The chemical analysis given in Table 1 shows that the sulfur content of the steel is reduced appreciably by electroslag refining while the content of other alloying elements remained within limits. The distribution of non-metallic inclusions in the refined and unrefined material is shown in Fig. 1 and the inclusion rating carried out according to ASTM E45 comparison [email protected]
] is shown in Table 2. The ESR steel is free from large size inclusions ( > 8 pm) of all types. The microstructure of the as-quenched specimens consisted of martensite and the prior austenite grain size in specimens oil quenched from 1470 to 1140K treatment is larger (250 pm) than in those quenched from 1140 K (25-35 pm). The specimens tempered at different temperatures had tempered martensite structure. All these structures are essentially the same in both unrefined and ESR materials. The room temperature tensile properties of the refined and unrefined steel in different heat treatment conditions are given in Table 3. While the yield strength (YS) and ultimate tensile strength (UTS) are unaffected by ESR, the ductility is enhanced to some extent. The plane strain fracture toughness Krc values obtained on the unrefined and the ESR steel are shown as a function of tempering temperature in Fig. 2. The data of Zackay (quoted by Parker  and Schwalbe [S]) on VAR steel are also included in Fig. 2. The data in Fig. 2 exhibits the following features: (i) In comparison with unrefined steel, the ESR steel has nearly twice the fracture toughness in all conditions of heat treatment. (ii) In as-quenched condition, the ESR steel has r<;, comparable with VAR steel, while in tempered condition the ESR steel is better than VAR steel. (iii) The 1470-1140 K quench has given improved K;, than the 1140K conventional treatment. (iv) The temper embrittlement occurring in the range 570-670 K is more pronounced in specimens austenitized at 1470-l 140 K than those austenitized at 1140K.
Effect of electroslag refining I I AISI 4340
, I STEEL
ELECTROSLAG REFINED -
L-1 , 12
: , ?__I 20 24
INCLUSION 51+X, pm
Fig. 1. Plot showing the number of inclusions per mm* vs inclusion size (pm) for AISI 4340 steel before and after ESR.
’ AM 4340 STEEL 180 160
AUST. TEMP. 1140 K
m ’ 1 ZACKA”MI75) i Al
Fig. 2. Fracture toughness of AISI 4340 steel as a function of tempering temperatures at two different austenitizing treatments. temperature in unrefined and ESR AISI 4340 steel are shown in Fig. 3. The data on specimens austenitized at 1140K and 147Q-1140 K are included. The data obtained by Banerji et al.  on VAR steel are included in the figure for the purpose of comparison. Figure 3 exhibits the following features: (i) The ESR steel has superior impact properties than the unrefined steel and is comparable to the VAR steel. (ii) A CVN fracture energy trough with its minimum at about 620 K occurs confirming the earlier observation of “one step temper embrittlement”.
M. G. HEBSUR at al. Table 2. Inclusion rating of AISI 4340 steel according to ASTM E45 comparison method Unrefined Thin Thick
Inclusion type Sulphides Alumina Silicates Oxides Volume fraction, % Average inclusion spacing, km
1.5 2.0 I.5 1.5
1.0 I5 I.0
0.5 0.5 0.5
Table 3. Tensile data on AISI 4340 steel before and after ESR. UR: Unrefined; ESR: Electraslag-refined; T: Temper; El: Elongation; RA: Reduction in Area; OQ: Oil Quench
O.Z%YS MPa ESR UR
114OK OQ 370K T 470 K T 570K T 670K T 870 K T
1592 1600 1612 1498 1470 1455
1595 1605 1610 1500 1480 1450
2215 2155 2087 1755 1700 1600
2220 2150 2090 1760 1700 1610
9 II I4 I3 18 25
I2 I5 19 I9 25 35
I7 26 SO 29 34 48
22 34 58 36 50 60
1470-ll4OK OQ 370K T 470 K T 570K T 670K T 870K T
1590 I590 1598 I500 1440 1390
159.5 1595 1600 1540 1445 139s
2193 2150 2085 1760 174s 15%
2195 2152 2090 1765 1750 1600
7 9 I3 I2 15 22
5 I5 18 20 28 30
8 I? 20 I8 15 20
i2 2.5 30 14 30 ix
AISI 1340 STEEL
UTS MPa UR ESR
Fig. 3. Effect of tempering and austenitizing temperature on the room temperature Charpy fracture energy of AISI 4340 steel.
Effect of electroslag refining
(iii) The 1470-l 140 K quench gives inferior CVN fracture energy in comparison with the conventional 1140K quench confirming the observations of Ritchie et al. . The fatigue crack growth rates da/dN are obtained from the crack length (a) vs number of cycles (N) data using a seven-point polynomial method[lO]. Since the cyclic load limits were kept constant, the applied AK (K,,, minus K,iJ increases as the crack length increases. It is customary to represent fatigue crack growth data in the form of da/dN vs AK as suggested by Paris [ 111.The data obtained on unrefined and refined steel in heat treated condition are shown in Figs. 4-6. In general, the curves show three regions[l2]. At low AK (region I), the curve asymptotically approaches the so-called threshold AK value. The curve then changes to a linear region (region II) before turning asymptotic again as AK approaches the KIc of the material (region III). It is generally observed that regions I and III are microstructure and load-ratio dependent while the slope in stage II is insensitive to these variables. The threshold stress intensity has increased and the crack growth rates in region III have decreased in the ESR steel, whereas the slope in region II is unaffected. Also the two different austenitizing treatments and the tempering temperature do not have any influence on region II. The effect of the austenitizing treatments and the tempering temperature is essentially on the threshold stress intensity.
DISCUSSION In the Electroslag Refining process, the unrefined material is melted under a molten slag cover and the molten metal droplets are allowed to fall through the slag to form a molten metal pool. The droplets get refined as they fall through the slag and refining reactions also take place at the interface of the electrode tip and the slag. The following changes occur during refining: (i) The large sized non-metallic inclusions are removed by physical floatation and chemical extraction within the slag and the finer inclusions are redistributed uniformly. (ii) Because of the high basicity of the slag and high temperature prevailing in the interface, desulfurization occurs effectively. (iii) As the refining occurs drop by drop, high surface area is exposed for refining which ensures chemical homogeneity. In contrast to ESR, the VAR process does not provide effective desulfurization and the
5 10-4 -
P 2: A’O. A
A A0 0
AA AA f
R50.1 1140 K OIL QUENCH
TEMPERING TEMP. 470 K UNREFINED A A
I 60 100 120 150
ALTERNATING STRESS INTENSITY, AK, MPa tm
Fig. 4. Fatigue crack propagation rates for AISI 4340 steel quenched and tempered at 470 and 870K and at R = 0.1.
M.G.HEBSUR eful. I
AISI 4340 STEEL
0 AAm’d OA A 0 A.'.. 0 Al 0,:. f
t UNREFINED ESR
y 2 -_
2 x 1o-7
steel quenched from two different austenitizing the influence of grain size at R = 0.1.
I III 40 60
Fig. 5.Fatigue crack propagationrates for AISI4340 temperatures and tempered at 470 K showing
01s ‘A OA
1140 K 470
1 4 i
I 10 ALTERNATING
Fig. 6. Fatigue
60 60 00 AK,
rates results for AISI4340 steel, quenched R = 0. I and R = 03.
inclusion distribution tends to be non-uniform [ 131.Thus in these two aspects, the ESR process is superior to VAR process. The electroslag refining has completely eliminated inclusions with sizes larger than 8 pm and reduced the number of those with sizes between 4 and 8 pm. Further, the lowering of sulfur content to very small values (0.004%) reduces the number and size of detrimental sulphide inclusions. The effect of non-metallic inclusions on the mechanical properties particularly on
ductility and fracture toughness is discussed by Rogers [141and Schwabe 181,in terms of the well known void formation mechanism. Due to the difference in the deformability of the matrix and the inclusion, decohesion occurs at the interface leading to void formation. The void formation can also occur due to the fracture of inclusions ahead of the crack due to large plastic strains occurring in the plastic zone. The ductile fracture proceeds by a process of void coalescence. Thus the ductility depends on the volume fraction, kind, size and distribution of non-metallic inclusions. The influence of volume fraction of inclusions on the tensile ductility in 4340 steel has been studied by Wells and Hauser[lS]. The transverse tensile ductility decreased with increasing volume fraction while the KI, remained unaffected. However, the type of inclusions studied were mainly oxides and it is not clear whether the size distribution for all the volume fractions is the same. Due to their high deformability index, the sulphide inclusions are most deliterious. Birkle et al. [ 161have shown that the KIc increased by about 40% when the sulphur content in the steel is reduced from 0.049 to 0.008%. Further, the Mn/S ratio decides [ 171the shape of the MnS inclusions and at higher ratios the inclusions prefer a globular shape to a plate-like shape. Gladman [ 181has shown that plate shaped inclusions are more dangerous than equiaxed ones as far as the fracture toughness is concerned. The ductility is also affected by the size of the inclusions since the plastic strain necessary to fracture an inclusion increases with decreasing inclusion size. Schwalbe has shown that the particle size reduction causes increase in fracture toughness. Whereas the particle size determines the onset of voids, the distance between the particles controls the void growth and crack propagation. The electroslag refining has eliminated the large size inclusions ( > 8 pm) at which the void formation would readily occur. Also the volume fraction of inclusions is reduced and the spacing between them is increased. Moreover, the reduction in the sulfur content has helped in increasing the MnlS ratio which helps in the formation of globular MnS inclusions. All these factors contribute towards improving the tensile ductility and fracture toughness in the ESR steels. The scanning electron micrographs obtained on the unrefined and refined steel are shown in Figs. 7(a-b). In the refined steel, the dimples caused by finer inclusions (Fig. 7b) are seen, in contrast to those due to large inclusion in the unrefined steel (Fig. 7a). The critical defect size can be calculated from the KIc values obtained on the refined and unrefined steel following the procedure suggested by Kiessling and Nordberg[l9]. This gives the maximum defect size that the material can tolerate with the stress intensity factor not exceeding the KIc value. For a disc shaped crack, the critical defect size is given by
(9 and for an ellipsoidal surface crack, the critical defect size (depth) is
(ii) where k = factor of safety generally taken as 1. The calculated values of diameter of inner spherical defect A and depth of ellipsoidal surface crack B at u = cy (gy = yield stress) are given in Table 4. Electroslag refining has increased the critical defect size in all conditions of heat treatments. From Fig. 2, it is observed that the ESR steel is somewhat superior to the VAR steel as far as the KIc is concerned. This is related to the following differences between ESR and VAR steels: (i) The sulfur content in the ESR steel is lower (0.004%) in comparison with that in the VAR steel (0.01%) used by Zackay . (ii) Although no mention of inclusion distribution in VAR steel has been made by Zackay, it is generally noted[l3] that the inclusion removal is more efficient in the ESR process than in the VAR process. With a view to comparing the fracture toughness of the unrefined, ESR and VAR steel, the
M. G. HEBSUR et ul. Table 4. Critical defect size at g = n, for AISI 4340 steel. A = Dia. of inner spherical defect; B = Depth of ellipsoid surface crack
Treatment lI4OK OQ 370K T 470 K T 670K T 147&1140K OQ 470 K T 870K T
Unrefined YS: K,, A MPa MPadm (mm) 1592 1600 1610 1470 I590 1595 1390
22 30 40 S6 35 45 55
2.98 5.5 9.6 24 7.5 12.3 24.5
105 I X3
7 60 16.4 25.5 60.6 22.2 44.3 45.20
1.45 1.13 4.87
1610 1475 1600 1600 1400
4.58 1.44 2.30 4.70
35 52 65 95 60 8S 75
II.5 4.24 X.46 X.63
AlSl 4340 S AUST.
Fig. 8. Plot showing the fracture toughness K,, vs ultimate tensile strength for AN 4340 steel.
KI, values are plotted against their individual UTS values in Fig. 8. For a given heat treatment the KI, values in the ESR steel are the highest. It is generally recognized that regions I and III of the fatigue crack growth rate vs AK curve, are sensitive to the microstructure and load-ratio effects. For similarly heat treated steels, the increase in the threshold stress intensity factor and the decrease in the fatigue crack growth rate in region III that occurred in the ESR steel can be attributed to the removal of non-metallic inclusions. In region 1, the crack growth rates are slow and the threshold values could not be estimated accurately. In region III where K,,, approaches KI, the growth rate is rapid and the failure is microstructure controlled. In the 1140K-oil quenched and tempered at 870 K specimen, the failure occurs by ductile mode and is controlled by inclusions (Fig. 9). On the other hand, in the 1470-1140 K oil quenched and tempered at 870 K specimen, an intergranular failure is observed (Fig. IO). The fatigue crack growth rate in region II is reported to be insensitive to microstructure and load ratio effects. Paris[ 1l] has shown that the slope in this region follows the equation: g
= C (AK)“’
where C is a material constant and m is the slope of the growth rate curve. The value of m is unchanged in the uprefined and refined steel and is also not affected by heat treatment. The
Effect of electroslag refining
Fig. 7. SEM Fracto~aph of fracture toughness specimen of AISI 4340 steel (1140K OQ, 870K Tempered) (a) Unrefined steel; (b) Refined steel. Note the presence of large sized inclusion (fractured) in unrefined steel (1000X).
Fig. 9. SEM Fractograph indicating the ductile fracture in region III of fatigue crack growth rate specimen of electroslag refined 4340steel treated with 1140K OQ, 870 K Temper. Note the presence of fine inclusions uniformly distributed within the dimples (550x).
M. G. HEBSUR et al.
Fig. 10. SEM Fractograph showing intergrannular fracture in the region III of fatigue crack growth rate specimen of the electroslag refined 4340 steel treated with 1470-l 140K OQ. 470 Temper (500x ).
Fig. 11. SEM Fractograph exhibiting the ductile striation growth in region II of fatigue crack growth rate specimen of electroslag refined AISI 4340 steel (1140K OQ, 870 K Temperj(550x )
Fig. 12. SEM Fractograph showing the intergrannular failure in fracture toughness specimen of refined steel, in 1470-I140K oil quenched condition (50x ).
Effect of electroslag refining
Fig. 13. SEM Fractograph showing the intergrannular failure in fracture toughness specimen of refined steel in 1470-l I40 K OQ, 470 K tempered condition (360x ).
Fig. 14. SEM Fractograph of as quenched (IMOOQ) fracture toughness specimen of electroslag refined AISI 4340 steel, showing the quasi-cleavage linked by areas of fibrous rupture (1500x).
Effect of electroslag refining
value of m for the AISI 4340 steel lies in the range 2%3.0 (mm/cycle)(MPa~m)-‘. The crack growth in this region gives rise to the ductile striations (Fig. 11) and is controlled by the amount of crack opening in each cycle. Ritchie et a1.[3,20] have noted a discrepancy between the Charpy fracture energy and fracture toughness &) values obtained in 4340 steel by 1470-l 140 K and the conventional 1140K quench. The Charpy fracture energy values for the 1470-l 140 K treated steel are lower then 1140K treated steel while the KI, values showed a reverse trend. The present results (Figs. 2 and 3) also confirm these observations. Ritchie et al.[3,20] explained this discrepancy on the basis of the effect of notch root radius on the toughness and the difference was shown to be independent of strain rate and shear lip energy differences. In the case of sharp cracks (as in K:,, evaluation), the effective root radius decides the characte~stic distance (L) ahead of the crack over which the critical stress must exist to cause failure. This characteristic distance is related to the microstructural feature which controls the fracture (e.g. grain size or inclusion spacing) and represents the minimum distance from the notch where the critical fracture event can occur. On the other hand ahead of round notches (as in Charpy impact) the distance at which the critical fracture event occurs (plastic-elastic interface) is much larger than the characteristic distance (L). Higher austenitizing treatment (1470-l 140K) was shown to cause grain boundary embrittlement due to S or P segregation and/or dissolution of MnS inclusions followed by reprecipitation at the prior austenite grain boundaries on cooling to 1140K. This _ reduces the critical fracture stress thereby decreasing the Charpy fracture energy. On the other hand higher austenitizing temperature increases the characteristic distance (L) or effective root radius due to larger prior austenitic grain size or dissolution of void initiating particles. This increases the sharp crack KIc values. Youngblood et aZ.have also explained the increased Kfc values of 300 M steel with higher austenitizing temperature in terms of the dissolution of second phase particles. The scanning electron micrographs (Figs. 12 and 13) recorded on the specimens heat treated at 1470-l 140 K and oil quenched without or with temper show intercrystalline failure in support of the grain boundary embrittlement mechanism. On the other hand the heat treatment at I140 K followed by oil quench gives quasi-cleavage type of failure (Fig. 14).
CONCLUSIONS (1) Electroslag refining of AISI 4340 steel improves the tensile ductility, fracture toughness (KIc), Charpy fracture energy and decreases fatigue crack growth rates in comparison with the unre~ned steel. (2) The improvement in the above mechanical properties is attributed to the removal of non-metallic inclusions and the reduction of sulfur content. (3) The K,, of the ESR steel is superior to VAR steel for a given value of tensile strength. (4) The fatigue crack growth rates in regions I and III have decreased in ESR steel. (5) The 1470-1140K austenitizing treatment has given better KI, values than the conventional 1140K treatment but the Charpy fracture energy values have shown a reverse trend. Higher austenitizing treatment has exhibited decreased fatigue crack growth rates. Ac~no~~edgemenr~-The authors gratefully acknowledge the help received from Dr. R. V. Krishnan of National Aeronautical Laboratory, Bangalore in doing the SEM work and Prof. A. K. Rao and Dr. B. Dattaguru of the Aeronautical Enginee~ng apartment of Indian Institute of Science, Bangalore, for extending their MTS facility.
REFERENCES 111W. Holzgruber, Possibilities and limitations to influence the structure of ESR ingots and properties of ESR products. Proc. 5th Int. Symp. Ekctroslug and other special melting Techn. (Ed. Bhat G. K.), pp. 70-90.Pittsburgh (1975).  G. Y. Lai, W. E. Wood, R. A. Clark, V. F. Zackay and E. R. Parker, The effect of austenitization temperature on the microstructure and mechanical properties of as quenched 4340 steel. Met. Trans. 5, 1663-1669(1974). I31R. 0. Ritchie,B. Francis and W. L. Server, Evaluation of toughness of AISI 4340 alloy steel austenitized at low and high temperatures. Met. Trans. 7A, 831-838(1976). I41 V. F. Zackay, Fundamental considerations in the design of ferrous alloys. Lawrence Berkefey Loborufory Repoti 3595, Berkeley, California (1975). [S] Standard test method for plane strain fracture toughness of metallic materials. Annual Book ofASTM sfandards. Part LO,Designation E399-74471-490ASTM (1976).
M. G. HEBSUR et al.
 Recommended practice for determining the inclusion content of steel. Annual Book of ASTM Stondords. Part 1I. Designation E45-76 119-130ASTM (1976).  E. R. Parker, Interrelations of composition, transformation kinetics, morphology and mechanical properties of allov steels. Mel. Trans. 84, 1025-1042(1977). [B] K. H. Schwalbe, On the influence of microstructure on crack propagation mechanisms and fracture toughness of metallic materials. Engng Fracture Mech. 9, 795-832 (1977).  S. K. Banerji, C. J. McMahon and H. C. Feng, Intergrannular fracture in 4340-type steels; effects of impurities and hydrogen. Met. Trans. 9A, 237-247(1978). [IO] Tentative test method for constant-load-amplitude fatigue crack growth rates above 10-s m/cycle, Annual Book of ASTM standards. Part 10, Designation E647-78T,ASTM (1978). [I l] P. C. Paris, The fracture mechanics approach to fatigue. Fatigue, an interdisciplinary approach. Proc. IOrhSag~morc~ Army Materials Research Conf., pp. 107-132,Syracuse University Press (1%4). [l2] E. J. Imhof and J. M. Barsom, Fatigue and corrosion-fatigue crack growth of 4340 steel at various yield strength\. ASTM STP 536, pp. 183-205(1973). [I31 K. C. Barraclough, Refining by remelting. Iron Making and Steel Making 4, 92-101 (1977).  H. C. Rogers, Ductility. Americnn Society of Mel& 31-56 (1%7). [IS] M. G. H. Wells and J. Hauser. Fracture prevention and control. American Society of Metals, 199-224 (1972). [I61 A. J. Birkle, R. P. Wei and G. E. Pellissier. Analysis of plane strain fracture in 0.4SC-Ni-Cr-Mo steel. Trans. ASM 59,981 (1%6). I171 . W. J. McG Tegart and A. Gittins. The role of sulfides in the hot workability of steels. Sulfide Inclusions in Steels, ASM 798-211(1974). [IS] T. Gladman, B. Holmes and I. D. Mclvor, Effects of second phase particles on strength, toughness and ductility. Effect of Second Phase Particles on fhe Mechanical Prooerties of Sfeel. The Iron and Steel Institute, London (I971I. [I91 Roland i(iessling and H. Nordberg. Influence of inclusions on mechanical properties of steel. Production and Application of Clean Steels, Int. Conf. held at Balatonfured, Hungary, 179-185(1971).  R. 0. Ritchie and R. M. Horn, Further consideration on the inconsistency in toughness evaluation of AISI 4340 steel austenitized at increasing temperatures. Met. Trans. 9A, 331-341 (1978). 1211 J. L. Youngblood and M. Raghavan, Correlation of microstructure with mechanical properties of 300M steel. Mel. Trans. EA, 143e1448 (1977). (Received 18 December 1979; received for publication 29 February 1980)