Full Text

Mechanism of Work Hardening in Hadfield Manganese Steel

Y. N. DASTUR AND W. C. LESLIE

When Hadfield manganese steel in the single-phase austenitic condition was strained in tension, in the temperature range - 2 5 to 300 ~ it exhibited jerky (serrated) flow, a negative (inverse) strain-rate dependence of flow stress and high work hardening, charac- teristic of dynamic strain aging. The strain rate-temperature regime of jerky flow was determined and the apparent activation energies for the appearance and disappearance of serrations were found to be 104 kJ /mol and 146 kJ/mol, respectively. The high work hardening cannot be a result of mechanical twinning because at - 5 0 ~ numerous twins were produced, but the work hardening was low and no twins were formed above 225 ~ even though work hardening was high. The work hardening decreased above 300 ~ because of the cessation of dynamic strain aging and increased again above 400 ~ because of precipitation of carbides. An apparent activation energy of 138 kJ /mol was measured for static strain aging between 300 and 400 ~ corresponding closely to the activation energies for the disapperance of serrations and for the volume diffusion of carbon in Hadfield steel. Evidence from the present study, together with the known effect of manganese on the activity of carbon in austenite and previous internal friction studies of high-manganese steels, lead to the conclusion that dynamic strain aging, brought about by the reorientation of carbon members of C-Mn couples in the cores of dislocations, is the principal cause of rapid work hardening in Hadfield steel.

A U S T E N I T I C manganese steel, called Hadfield man- ganese steel after its inventor, Robert Hadfield, t is a tough, nonmagnetic, Fe-C-Mn alloy, useful for severe service combining abrasion and heavy impact. The ASTM Standard A-128-642 covering this steel allows composition ranges from 1.0 to 1.4 pct C and from 10 to 14 pet Mn. However, commercial alloys with manga- nese contents greater than 12 to 13 pct are seldom used because of cost. Moreover, work hardening in a 1.15 pct C alloy reaches a maximum at 13 pct Mn. 3 Hadfield steel is usually austenitized to dissolve carbides and to produce homogeneous austenite, which is preserved by water quenching from above 1000 ~ Typical me- chanical properties are: 3 a) yield strength (0.2 pct offset), 379 MPa; b) ultimate tensile strength, 965 MPa; c) elongation in 50 mm, 50 pet; d) reduction of area, 40 pct; e) hardness, as quenched, 190 HB; f) hardness, at fracture, 500 HB; g) Charpy V-notch impact, 169 J at 22 ~ 7 J at - 196 ~ The unique feature of this tough, high-strength steel is the rapid work hardening, from a yield strength of 379 MPa to an ultimate tensile strength of 965 MPa. In gouging abrasion tests, the Hadfield steel performs better than wrought alloy steels, cast alloy steels, stainless steels, tool steels or high-chromium white irons. 4 These com- binations of properties make it useful in such diverse applications as crawler treads for tractors, railroad frogs, grinding mill liners, crusher jaws and cones, impact hammers, dipper bucket teeth and nonmagnetic plates for electromagnets.

Y. N. DASTUR and W. C. LESLIE are at the Department of Materials & Metallurgical Engineering, University of Michigan, Ann Arbor, MI 48109.

Manuscript submitted August 11, 1980.

Although this steel has been used for nearly a century since its development in 1882, the mechanism of rapid work hardening remains unclear. The purpose of this study was to determine the mechanism of rapid work hardening in Hadfield steel with the intent of solving one of the classical mysteries of physical metallurgy.

PREVIOUS STUDIES A N D PROPOSED MECHANISMS OF W O R K H A R D E N I N G IN H A D F I E L D STEEL

It is commonly taught that the rapid work hardening in Hadfield steel arises from strain-induced transfor- mation of y to ~ or c martensites: but it has been shown that the austenite of Hadfield's composition is stable during plastic strain, 6-8 even below -196 ~176 Strain- induced transformation occurs only because of decar- burization or local segregation of manganese that leads to unstable austenite compositions. Some workers l~'12 attributed the rapid work hardening to fine mechanical twinning. However, their studies did not include meas- urements of rates of work hardening nor description of microstructures obtained at a variety of strain rates and temperatures. Lambakakhar and Paska113 observed no correlation between frequency of twins and hardness. Instead, they concluded that the hardness of Hadfield steel is more likely a function of the general dislocation structure than of the specific microstructure. Drobnjak and Parr 14 suggested that stacking fault-dislocation interactions were responsible for increasing the strain- hardening rate. However, according to Roberts, 7 stack- ing faults were present only in hammered specimens, not in tensile or explosive shocked specimens which deformed by twinning. The stacking fault energy of 1.1 C, 12 pct Mn steel was determined to be 50 m J / m 2

ISSN 0360-2133/81/0511-0749500.75/0 �9 1981 AMERICAN SOCIETY FOR METALS AND

THE METALLURGICAL SOCIETY OF AIME METALLURGICAL TRANSACTIONS A VOLUME 12A, MAY 1981--749

at room temperature, decreasing with decreasing tem- perature and decreasing carbon content. 15 This value is higher than those of fcc metals such as Ag (16 m j/m2), ]6 Au (32 m j /m2) ~7 and austenitic stainless steels (---20 m J/m2) ~8 whose work-hardening rates are lower than that of Hadfield steel. Serrated flow in this alloy has been observed 9,19 but its limits have not been deter- mined. Several authors 7a~ have proposed that the rapid work hardening is due to the interaction of dislocations with carbon atoms in solid solution in austenite. No one has observed carbide precipitation in this alloy at ambient temperature nor seriously pro- posed that the exceptional work hardening is due to carbide precipitates. From a comparison of the defect structures and hardening in shock-hardened Fe-32 Ni, Fe-23 Ni-0.6 C and Hadfield steel, Leslie 2~ concluded that the most likely cause of the rapid work hardening in Hadfield steel is interactions between dislocations and Mn-C couples in solution in austenite. Sastri's TM

studies of MOssbauer spectroscopy of Hadfield steel indicated a clustering of carbon atoms in austenite during cold working with the process being enhanced by aging.

MATERIALS AND P R O C E D U R E

Tension specimens 6.25 mm in diameter with a gage length of 25 mm were machined from commercial hot-rolled bars of Hadfield manganese steel having the composition 1.13 pct C, 11.4 pct Mn, 0.2 pct Si, 0.17 pct Ni, 0.16 pct Cr and 0.08 pct Mo. These specimens were sealed in stainless steel envelopes, annealed at 1100 ~ x v l v x * ~ - x x o , z l l l ~ t , ~ l , u x ~ , x L ~ - l l . , ~ l x ~ . , L L ~ . , ~ LLL l l , ,~ . ,~ . . t t ) l L I X ~ . , t V

produce a single-phase austenite with grain size ASTM 3-4, without decarburization.

T O3 5O LO c r

I-- C/)

Prestroin

~o"

I / l U n l o o d , AcJe, Relood

STRAIN

Fig. l --Stress-strain curve illustrating static strain aging.

transmission electron microscopy were prepared by cutting sections about 300/~m thick with a diamond saw, grinding them carefully to about 100/~m thickness and then electrothinning in a dual-jet electropolisher/ thinner in a mixture of 80 g anhydrous sodium chro- I I I K ~ . I . ~ K ~ . l l ~ t " I ' K / K / 1 1 1 1 ~ I ~ I I . ~ , I K I . I ( , t~ l . ,~ , l , l v ~ O . ~ l ~ l ~ L ,./KS (.I.IIK.I~ 1 J, J K S l t t , / ~

for 6 to 8 min. Foils were examined in a JEOL 100CX STEM at 100 kV.

Strain aging

In tests of dynamic strain aging, several heat-treated specimens were strained in tension at rates between 10-Ss - ] and 10-2s -l and at temperatures ranging from - 50 to 600 ~ in a 250 kN Instron machine equipped with a radiant-heating chamber and a low-temperature bath. Static strain aging experiments were performed on other specimens as follows: specimens were prestrained 4 pct in tension at - 2 5 to 400 ~ (_+ I~ in con- trolled-temperature liquid baths for short aging periods (less than 30 mins) and in an air-circulating furnace for longer periods. At aging temperatures below 300 ~ the stress increment, Ao, was measured as shown in Fig. 1. At higher aging temperatures Ao was defined as the 0.2 pct offset yield stress after aging less the flow stress before aging. Straining before and after aging was done at - 25 ~ to eliminate the possibility of dynamic strain aging during the test; therefore, no aging tests were performed for periods of less than two minutes.

Metallography

Sections parallel and perpendicular to the tensile axis were cut from dynamically strain-aged specimens, elec- tropolished in a sodium chromate-acetic acid solution and examined by light microscopy. Thin foils for

MOssbauer Spectroscopy

A Ranger M6ssbauer spectrometer was used in conjunction with a Tracor Northern NS-600 multi- channel analyzer. 25/~m thick samples were prepared from specimens strained 20 pct in tension, within and outside the dynamic strain-aging region. All spectra were taken at room temperature, in the transmission mode, using 57Co/Rh as source of y-rays and a-Fe foil, enriched in 57Fe, as a reference absorber. The data were computer curve fitted by the flexible least-square rou- tine for general MOssbauer effect spectra fitting of Wilson and Swartzendruber. 22

RESULTS

In the temperature range between - 2 5 and 300 ~ Hadfield manganese steel, strained in tension, exhibited serrated flow (jerky stress-strain curves), negative strain-rate dependence of flow stress (decrease in flow stress with increase in strain rate) and high work- hardening rates. Stress-strain curves, at room temper- ature, showing these characteristics are presented in Fig. 2. Figure 3 illustrates serrations at various temper- atures and a nominal strain rate of 3 • 10-% -1. Three different types of serrations, A, B and C, were observed, depending on strain rate and temperature. Types A and

750- -VOLUME 12A, MAY 1981 M E T A L L U R G I C A L TRANSACTIONS A

|OOC E = 3 x IO -4 s -1

W~t~" 600 r r I-- O r ) 400

2OO

I I I I I I I I I I I I ~ I 0 4 8 12 16 20 24 28 32 36 40 44 48 65

STRAIN, %

Fig. 2- -Engineer ing stress-strain curves o f Hadfield manganese steel at 24 ~

-10~

YPE A

EG=45%

I 50 sec

24~

TYP 7

05 INN ~Ec=09%

40 s e c

125~ 190oc

TYPE C.

E==6.2% = , %

24 s e c

Fig. 3--Serra ted flow in Hadfield steel as affected by temperature. Nominal strain rate 3 x 10 -4 s - l , ec = "critical s t ra in" .

B serrations were observed at lower temperatures and type C at higher temperatures. The serrations increased in magnitude and frequency as the temperature was raised, then vanished at temperatures above 190 ~ The strain at which serrations first appeared, e~, the so-called critical strain, first decreased then increased with increasing temperature. The effect of strain rate on serrations is shown in Fig. 4. On increasing the strain rate type B serrations disappeared, with a drop in average level of flow stress, that is, the flow stress had a negative strain-rate dependence. The temperature range of this phenomenon is shown in Fig. 5.

Figure 6 shows the strain rate-temperature regime of serrations. Outside this region the flow was smooth. The apparent activation energies for the appearance and disappearance of serrations were 104 kJ /mol and 146 kJ/mol respectively. The estimated error in these values is _ 15 pet. Figure 7 is a comparison of stress-strain curves at -25, 24 and 500 ~ to illustrate rates of work hardening within and outside the dynamic strain aging region. The correlation of serrated flow, negative strain-rate dependence and work hardening is shown in Fig. 8. The measure of work hardening used in this figure is the difference in flow stress at plastic strains of 0.04 and 0.002 (0oo 4 - 0o.oo2). It is obvious that work hardening is high within the negative strain-rate depen- dence range and drops off sharply at both ends of this range. At temperatures above 400 ~ the work harden-

t . - ; . , 2800

/ T at 600 m~ , . . . f I KN on | f f Mogmhed

61- ~ STle 1200

4 I 800 aW~.(Type B

2 [ Serrohons) I 2% I 400

g -

o ELONGATION P

Fig. 4 - - A por t ion o f the load-elongat ion curve o f Hadf ie ld man- ganese steel at 24 ~ showing the effect of strain rate on serrated flow.

El

/ ~t ~164 secl /

/ , I

Fig. 5- -Var ia t ion of strain-rate dependence of flow stress with tem- perature in Hadfleld steel.

_

L~ ~ t6 3 z

r~ I-- r . .

1(~4 -- o o

r 1.5 2.0

TEMPERATURE ,~ 300 200 t00 25

I I I

i Termination * t46 Kd/mole

o -

tions

/ ,S

KJ/mole

Serrotions

J 1 3.0

t O 0 0 / T , K - I

0 -25 I L

oO No

;erra- 'ions

4.0

Fig. 6- -St ra in rate- temperature regime of serrations in Hadfield steel.

M E T A L L U R G I C A L T R A N S A C T I O N S A V O L U M E 12A, MAY 1981--751

= 3 x 10"4s -I 500 -

o 400 (3-

~o~300 I.l.l IX I-- 200 O 9

t 00

I I I I I I I I 0 2 4 0 4 8 12 16 20

S T R A I N , %

Fig. 7- -Compar ison of engineering stress-strain curves of Hadfield manganese steel illustrating different rates of work hardening within and outside the dynamic strain aging region.

t601f Range of 9 J Serrated Flow

r i - i / i / i

',, loo , , / - i5 ~,

bq 80 =tO'"sec-'~, / \ \ / ~2" -10

6C V \q1oZsec' 4 0 I I I I i i I

- 5 0 0 100 2 0 0 3 0 0 4 0 0 .500 6 0 0 T EM PERATURE,~

Fig. 8- -Tempera tu~,dependence of work hardening in Hadfield steel. (*Negative strain-rate dependence)

1100

~000

900

800

70o

o 600

~E .5OO

400

30C

20C

Ioc

c

~ m Z ~ 6o ~ `5o B 4 I.iJ N 0 T

-100

e ~' 10"4sec "I

( RE~,O~ ,1 I" ~A 4 "~

0 2 % Offset

150

I00

Ksl

50

I " E,~. -: E~ r "I g

~ s e c - ' < z

I i J t [ f 0 100 200 300 400 500

TEMPERATURE, *C Fig. 9---~'Temperature dependence of tensile properties of Hadfield steel (El u = uniform elongation, El T = total elongation).

ing rises again. Similar results (not shown here) were obtained when the measure of work hardening was (%.08 - 00.002) or (002 - 00.002).

Figure 9 shows the tensile properties of the Hadfield steel. A peak in ultimate tensile strength and a slow, continuous decline in yield stress were observed in the dynamic strain-aging region. Ductility, measured as pct elongation in 25 ram, is high and also reaches a maximum within the dynamic strain-aging region. In the temperaure region from -25 to 300 ~ in which the uniform elongation equaled the total elongation, specimens failed predominantly by intergranular frac- ture with void formation at grain boundaries, as ob- served by scanning electron microscopy. At 500 ~ the fracture was accompanied by a large reduction of area. The fracture surface contained dimples.

Figures 10 to 18 show the microstructure of Hadfietd steel after 20 pct strain in tension at several temper- atures. Mechanical twins are present in profusion at the lowest temperature studied ( - 50 ~ Comparisons of microstructures at several temperatures within the dy- namic strain aging region (Figs. 10 to 12) indicate that the twin density decreases with increasing temperature, approaching zero at 225 ~ At this temperature, the

(u)

(b)

Fig. 10--Deformation twins in Hadfield steel after 20 pct strain in tension at (a) -50 ~ and (b) 24 ~ Sodium chromate, acetic acid electrolytic etch.

752- -VOLUME 12A, MAY 1981 METALLURGICAL TRANSACTIONS A

Ca)

Fig. 13--Carbide precipitation on austenite grain boundaries in Hadfield steel after 20 pct strain in tension at 400 ~ Sodium chromate, acetic acid electrolytic etch.

(b) Fig. 11- -Deformat ion twins in Hadfield steel after 20 pct strain in tension at 190 ~ (a) ~ -- 10 -4 s - l , (b) e = 10 -2 s - l . Note the reduced twin density compared to Fig. i0. Sodium chromate, acetic acid electrolytic etch.

Fig. 12--Microstructure of Hadfield steel after 20 pct strain in ten- sion at 225 ~ Austenite matrix with few mechanical twins. Sodium chromate, acetic acid electrolytic etch.

(b)

Fig. 14--Carbides in Hadfield steel after 20 pct strain in tension at (a) 500 ~ and (b) 600 ~ Sodium chromate, acetic acid electro- lytic etch. Magnif icat ion 185 times.

M E T A L L U R G I C A L T RANS ACT IONS A V O L U M E 12A, MAY 1981--753

r

(a)

(b) Fig. 15--Transmission electron micrographs of deformation twins in Hadfield steel strained 20 pct in tension at 24 ~ (a) Bright field image. (b) Dark field image from (131) reflection.

work hardening is the same as at temperatures where twinning was profuse (Fig. 8). A comparison of Figs. 8 and 10 to 12 reveals no correlation between twinning and work hardening. Figures 13 and 14 show carbide (M3C) precipitates, which correlates well with increased work hardening above 400 ~ as shown in Fig. 8. Compar ison of Figs. 1 l(a) and (b) shows that strain rate has little effect on twin density, within the range of strain rates studied (10 -5 to 10-2s-~).

Figures 15 to 18 are transmission electron micrographs and electron diffraction patterns of Hadfield steel after 20 pct strain in tension at 24 and 500 ~ These corroborate~d observations by light microscopy and confirmed the presence of twins at low temperatures and precipitation of (Fe, Mn)3C at higher temperatures.

Results of static strain aging at temperatures below 300 ~ are shown in Fig. 19. The increase in flow stress, Ao, shown in Fig. 1, rapidly increased to a constant value of about 18 MPa at all temperatures studied except for -25 ~ No further rise in zao was observed even after long periods at these temperatures. Samples taken from this group of specimens showed a single- phase austenitic structure by light microscopy. Results of static strain aging at temperatures of 300 ~ and

rb) Fig. 16--(a) Electron diffraction pattern from twinned austenite for the area in Fig. 15. [111] matrix and [114] twin normal to the foil. (b) Indexed pattern. �9 matrix reflections, O twin reflections.

higher are shown in Fig. 20 with data from Fig. 19 included for comparison. In contrast to low-tempera- ture aging, At, increased sharply after relatively long aging, then approached a plateau. Samples from this group of specimens showed carbide precipitates under light microscopy. These specimens had very limited total elongation. The activation energy for the aging process, between 300 and 400 ~ was 138 kJ/mol , as shown in Fig. 21. The kinetics of aging were determined from plots of (Ao -- Aoo) vs time at various tempera- tures, as shown in Fig. 22. The slopes of the lines are close to one, i .e. , (Aa - Aoo) is proport ional to aging time, which is not in accord with Cottrell 's t 2/3 law, 23 perhaps because of the very high carbon concentration and precipitation of carbides.

Figure 23(a) is a MOssbauer spectrum of Hadfield steel in the as-quenched condition. Figures 23(b) through (d) show spectra of specimens strained 20 pct at 24, 190 and 350 ~ respectively. Each spectrum can be resolved into a central peak, with a negQ.tive isomer

754--VOLUME 12A, MAY 1981 METALLURGICAL TRANSACTIONS A

(a)

(a)

(b) Fig. 17--Transmission electron micrographs of a carbide plate in Hadfield steel strained 20 pct in tension at 500 ~ (a) Bright field image. (b) Dark field image from (021) carbide reflection.

shift, and a quadrupole-split pair with values indicated on the figure. These peaks agreee with those published for iron-carbon austenite 24-27 and iron-carbon-manganese austenite. 21 The central peak can be attributed to iron atoms free of carbon neighbors and the quadrupole pair to iron atoms in a noncubic environment because of carbon neighbors. No strain-induced martensite was observed, in agreement with the results of Sastri and Ray 21 and Katz, Mathias and Nadiv. 28 There was no difference, within experimental error, between the rel- ative areas under the central and the quadrupole peaks for different spectra, therefore, no effects of strain aging could be detected.

DISCUSSION

A. Summary of Observations

The results of tension tests of Hadfield steel at various temperatures and strain rates, supplemented by data from static strain aging experiments, show that a strain-aging mechanism operates between - 25 and 300 ~ and within this temperature range the alloy work hardens rapidly. As in other alloys undergoing dynamic strain aging (DSA), such as low-carbon steel, 29

(b) Fig. 18--(a) Electron diffraction pattern from carbide plate in austenite for the area in Fig. 17. (b) Indexed pattern. �9 Austenite spots, O carbide spots.

2m l h Id 7d I I I I

2s

b <3

10

(

5

0

[] o g 4~ n

0 o 0 ~ - 2 5 C

I I I 101 10 2 10 3 10 4

TIME, s e e

t 0 5

Fig. 19--Static strain aging of Hadfield steel at low temperatures.

METALLURGICAL TRANSACTIONS A VOLUME 12A, MAY 1981--755

150 t j L ~ 125 1oo 4 o o * c f / / ~ -

':' r~ / / / 35~176 r o / / / oo c/-

so I 7 / E .... , / j / /2z~c a 25 . 200 ~C

102 103 104 ]05 106 TIME, sec Fig. 20--Static strain aging of Hadfield steel at high temperatures.

5 min. 1 hour 24 hrs 7d 15 d a y s

2O ]5 I0 5 0

t00 ~ S L O P E : 118

~50F ~ ~~ rlOj~_L / / 5 7 ~ ~ /SLOPE=098/

t]" I " IO s t04 tO 5 t06

TIME, sec Fig. 22--Stress increment after aging, corrected for initial time- independent stress (Ao - A a o ) , plotted vs time, from data in Fig. 20.

400 350 500,0C t V a l I I I [

Q= t38 Kd/mole

1(9"*

t0 M P a M P o

--~ M P o

t.25 1.50 1.75 2.00 IO00/T, K -I

Fig. 2 t - -Act iva t ion energy for static strain aging of Hadfield steel between 300 and 400 ~ (data taken from Fig. 20).

austenitic stainless steel3 ~ a T i , 31 V - N b , 32 Inconel 60033 and maraging steel, 34 such pronounced work hardening is associated with serrated load-elongation curves and a negative (inverse) strain-rate dependence of the flow stress. Serrated flow curves during DSA have been explained in terms of the initiation, propagation and pinning of Luders bands. 3s,36,37 Numerous studies 3s-45 have shown that the increased work hardening accom- panying serrations can be attributed to an increased rate of dislocation multiplication, which increases the dislocation density for a given strain.

The activation energy for the onset of serrations, 104 _+ an estimated 15 k J/tool, (Fig. 6) is considerably less than the activation energy for bulk diffusion of carbon in austenitic Hadfield steel - 1 4 0 to 150 k J / m o l (Fig. 24) 46.47

The static strain aging experiments showed that an increase in the lower yield stress was produced in aging periods of less than two minutes at 22, 100 or 200 ~

and that the magnitude of the increase was independent of temperature and aging time, up to one week (Fig. 19). The much smaller increase in yield stress noted after aging at - 25 ~ may be attributable to a Haasen- Kelley effect 48 but this cannot be stated with certainty.

The much larger increases in the yield stress observed after lengthy aging at 300 to 400 ~ (Fig. 20) can be attributed to formation of Cottrell atmospheres and to precipitation of (Fe, Mn)3C. The activation energy for these static strain aging effects is in excellent agreement with the activation energy for bulk diffusion of carbon in Hadfield steel (Fig. 21).

Work hardening was high and remained nearly constant throughout the dynamic strain aging range and decreased at temperatures above and below this range (Fig. 8). At temperatures above 400 ~ the initial rate of work hardening was high (Fig. 7) because of the presence of carbides.

In the present study, no strain-induced transforma- tion of austenite of Hadfield composition to a or c martensite was observed by light or electron microscopy or by MOssbauer spectroscopy, in agreement with results obtained by a number of previous workers. 6-;~

B. Mechanism of Work Hardening

As a result of the rapid pinning of dislocations in Hadfield steel during plastic deformation within the temperature range of normal usage, work hardening is high. This pinning is a diffusion-controlled process with an activation energy much below that for bulk diffusion of carbon in this steel. The only process that would appear to satisfy these requirements is short-range diffusion of carbon within the cores of dislocations. In his review of core diffusion in fcc metals, Balluffi 49 concluded that the activation energy for diffusion within the core was 0.4 to 0.7 of the activation energy for bulk diffusion. The activation energy for the onset of serrations is about 0.6 to 0.7 of the activation energy for bulk diffusion of carbon in Hadfield steel. Esti- mations of the relative mobilities of substitutional solutes in dislocation cores and in the undisturbed lattice of bcc iron have shown that serrations appeared at temperatures where the solute was mobile in the core but immobile in the lattice. 36 Similar calculations from our data indicate that a few jumps of carbon atoms can occur in a dislocation core in the time between serra- tions, even at - 10 ~

756--VOLUME 12A, MAY 1981 METALLURGICAL TRANSACTIONS A

1.2-- ' ~-

Fig. 23--(a) Room temperature M0ssbauer spectra of Hadfield steel. (a) as quenched; after 20 pct strain at 24 ~ (b), 190 ~ (c), and 350 ~ (d). �9 Data, --f i t . IS = isomer shift of central peak with respect to t~-iron. QS = quadrupole splitting. Estimated error in IS = • mm/s . Estimated error In QS = _+0.021 mm/s .

Z s

Co 03

Z <

If) I f) M E (13 Z < IE I--

1 . C - -

0.8-

0.G

0.4

0.2

0O I �9 - I O -5 0 5 I0

VELOCITY, mm/s (a)

1.2

1.0

0 .8

0.6-

0.4

0.2

O. - t 0

I I -5

VELOCITY, mm/s (c)

I S = - O . 06 Omm/s

OS= O. 65 mm/s

0 5 IO

P.

1 . 0 _

0.8

0.6

0.4

0.2

0.0 I -tO

I S = - 0 . 064 mm/s

OS= O. 63mm/s

I I I -5 0 5 10

VELOCITY, mm/s (b)

1.2

1.0 _~

0.

0.6

0.4

0.2

O. _110 -5

I S = - 0 . 068 mm/s

QS= 0 .64min i s

I I 0 5 tO

VELOCITY, mm/s (d)

- ~

-5

o

-15

"25

1200 6 0 0 5 0 0 100 0 - 2 5 1 [ 1 1 I I

mocrodiffusion)

Ke end Wong. internel friction

\ \ \ \ \ \

\ \ \

a2 Temp, T = 12D

200"C I sec. 50"C 68 see.

100eC 7.6 hrs. 24~ 109 yrs.

-35 I I I I I 0 I 2 5 4 5

IO00 /T , K "1

Fig. 24--Temperature dependence of the diffusivity of carbon in Hadfield steel (r is the time for one jump of a carbon atom).

On the unrealistic assumption of a uniform dispersion of manganese and carbon atoms in a steel containing 12 at. pct manganese and 5 at pct carbon, the mean distance between manganese atoms on a volume basis would be 0.46 nm and between carbon atoms 0.62 nm. Thus, within one Burgers vector of the center of a dislocation there is an ample supply of solutes to serve as pinning agents.

In fcc alloys, internal friction peaks are attributed to point defect pairs. Ke and Wang 46 and Kandarpa and Spretnak 5~ concluded that the internal friction peak in austenitic manganese steel arose from stress-induced ordering of carbon atoms arising from the distortion of octahedral sites by neighboring manganese atoms. Such ordering can occur more readily within dislocation cores than in the elastically strained lattice.

Thermodynamic data 5~ for Fe-Mn-C alloys show that manganese decreases moderately the activity of carbon in austenite, indicating an attractive force between manganese and carbon which is necessary to produce the point defect pairs which pin dislocations. The manganese-carbon combinat ion is unique in that their interaction energy is appreciable but not so high that carbides are quickly precipitated. Thus, large concen- trations of both can be held in supersaturated solid solution in austenite. These large concentrations have an important practical effect. It is generally noted that DSA reduces ductility, as measured by uniform elon-

METALLURGICAL TRANSACTIONS A VOLUME 12A, MAY 1981--757

gation. 52 Baird and Jamieson 53 proposed that the early onset of necking during extension in the DSA range is caused by local exhaustion of interstitial solutes so that the supply becomes insufficient to lock mobile dislo- cations, thus providing a site for localized plastic flow and necking. Because of the high concentrations of carbon and manganese in Hadfield steel, this exhaus- tion cannot occur and the result is a high work- hardening rate, high uniform elongation and a tough alloy.

Work hardening of Hadfield steel decreases at tem- peratures above and below the DSA range. At low temperature, ~ - 25~ or at high strain rates at higher temperatures, the carbon atoms are immobile both in the dislocation cores and in the lattice so pinning does not occur. It is tempting to speculate that at temper- atures and strain rates at which serrations disappear the carbon atoms can diffuse with the gliding dislocations, as might be indicated by the similarity of the activation energies for bulk diffusion of carbon and for the termination of serrations, Fig. 6. Unfortunately, this encounters the same difficulty noted previously36--an unreasonably high density of mobile dislocations is required. The mechanism of the termination of serra- tions remains obscure.

Both twinning and DSA were observed in the tem- perature range -25 ~ to 225 ~ in which work harden- ing was high and constant, but the density of twins decreased monotonically as the temperature was raised. High work hardening persisted at temperatures above 225 ~ in the absence of twinning, but in the presence of DSA. Work hardening decreased when flow became smooth at temperatures below - 25 ~ even though twinning was profuse. We conclude that twinning cannot be the principal cause of rapid work hardening in this alloy.

C. Recommendation for Increasing the Wear Resistance of Hadfield Steel

Hadfield steel performs best under conditions of gouging or battering wear, the conditions existing in jaw or cone crushers. Diesburg and Borik 4 showed that the gouging wear ratio of Hadfield steel--the weight loss of the test specimen divided by the weight loss of a reference specimen of "T- l Type A" steel, 260 HB --decreased with increasing carbon content. Although it is known that the gouging wear resistance of Hadfield steel can be improved by increasing the carbon content, there is a practical upper limit to the carbon content, set by the need to minimize carbide precipitation during quenching from the austenitizing temperature. Such precipitation decreases the toughness of the alloy. To increase the concentration of carbon in solution in Fe-Mn austenite, it may be necessary to reduce the activity of carbon by the addition of another substitu- tional solute.

CONCLUSIONS

1) Dynamic strain aging is the principal cause of work hardening in Hadfield steel. The carbon members of C-Mn couples reorient in the cores of dislocations,

thereby locking the dislocations. This then leads to a high dislocation density for a given strain.

2) Twinning is not a major factor contributing to rapid work hardening of Hadfield steel.

3) No strain-induced transformations were observed in Fe-Mn-C alloys of Hadfield composition.

4) The service performance of Hadfield manganese steel may be improved by increasing the carbon content in solution in austenite, which may be accomplished by adding a second substitutional solute that further de- creases the activity of carbon in austenite.

A C K N O W L E D G M E N T

We are grateful to Professor J. Datsko for advice and assistance in machining Hadfield steel, to Professor D. Vincent for his continued interest and valuable advice in Mossbauer spectroscopy, to Mr. Y. Hong for assistance in light microscopy, to Mr. L. F. Allard for assistance in electron microscopy, to Mr. R. J. Sober of the U.S. Steel Research Laboratory for his advice on high-temperature strain measurements and to Mr. R. Schoone, also of the U. S. Steel Research Laboratory, for his advice on preparation of thin foils for electron microscopy. The support of the National Science Foundation, under Grant No. DMR 77-02000, is greatly appreciated.

REFERENCES

1 R.A. Hadfield: Metallurgy and Its Influence on Modern Progress, p. 91, Chapman and Hall Ltd., London, 1925.

2. ASTM A 128-75a, Annual Book of A S T M Standards, Part 2, p. 95, ASTM,, Philadelphia, PA, 1979.

3. E. R. Hall: A S M Technical Report No. C6-17.1, p. 15, 1966. 4. D. E. Diesburg and F. Borik: Symposium on Materials for the

Mining Industry, p. 15, Climax Molybdenum Company, Green- wich, CN, 1974.

5. R. W. Cahn: The Encyclopedia o f Ignorance, p. 140, Pergamon Press, NY, 1977.

6. H. M. Otte: Acta Met., 1957, vol. 5, p. 614. 7. W. N. Roberts: Trans. TMS-AIME, 1964, vol. 230, p. 373. 8. S. A. Sastri: Proc. 3rdlCSMA, Cambridge, England, Aug. 1973,

vol. 1, p. 596. Institute 6f Metals Monograph and Report Series, no. 36.

9. H. C. Doepken: Trans. AIME, 1952, vol. 194, p. 166. 10. C. H. White and R. W. K. Honeycombe: J. Iron Steellnst., 1962,

vol. 200, p. 457. 11. Z. Nishiyama, M. Oka, and H. Nakagowa: Mem. Inst. Sci. Ind.

Res., Osaka Umv., 1964, vol. 21, p. 57. 12. K. S. Raghavan, A. S. Sastri, and M. J. Marcinkowski: Trans.

TMS-AIME, 1969, vol. 245, p. 1596. 13. O. O. Lambakakhar and Yu. I. Paskal: lzv. Vyssh. Vchebn. Zaved.

Fiz., 1973, vol. 16, p. 26. 14. Dj. Drobnjak and J. G. Parr: Met. Trans., 1975, vol. 40, p. 554. 15. P. Yu. Volosevich, V. N. Gridnev, and Yu. N. Petrov: Fiz. Met.

Metalloved., 1975, vol. 40, p. 554. 16. D. J. H. Cockayne, M. L. Jenkins, and I. L. F. Ray: Philos. Mug.,

1971, vol. 24, p. 1383. 17. M. L. Jenkins: Philos. Mug., 1972, vol. 26, p. 747. 18. P. J. Brofman and G. S. AnseU: Met. Trans.A., 1978, vol. 9A, p.

879. 19. G. Collette, C. Crussard, A. Kohn, J. Plateau, G. Pomey, and M.

Wexsz: Rev. de Metal., 1957, vol. 54, p. 433. 20. W. C. Leslie: Metallurgical Effects at High Strain Rates, p. 571,

Plenum Press, NY, 1973. 21. S. Sastn and R. Ray: Met. Trans., 1974, vol. 5, p. 1501. 22. W. Wilson and L. J. Swartzendruber: Comput. Phys. Commun.

1974, vol. 7, p. 151. 23. A. H. Cottrell and B. A. Bilby: Proc. Phys. Soc., 1949, vol. A62, p. 49.

758--VOLUME 12A, MAY 1981 METALLURGICAL TRANSACTIONS A

24. P'. tJlelen and It. Kaplow: Acta Met., 190/, vol. 13, p. 49. 25. B. W. Christ and P. M. Giles: Trans. TMS-AIME, 1968, vol. 242,

p. 1915. 26. J. M. R. Genre and P. A. Flinn: Trans. TMS-AIME, 1968, vol.

242, p. 1419. 27. M. Leosoille and P. Gielen: Met. Trans., 1972, vol. 3, p. 2681. 28. Y. Katz, H. Mathias, and S. Nadiv: Proc. First J IM Int. Symp. on

New Aspects of Martemitic Transformation, Suppl. to Trans. JIM, 1976, vol. 17, p. 381.

29. A. S. Keh, Y. Nakada, and W. C. Leslie: Dislocation Dynamics, p. 381, McGraw-Hill Book Co., NY, 1968.

30. J. T. Barnby: J. Iron Steel Inst., 1965, vol. 203, p. 392. 31. A. Santhanam and R. E. Reed-Hill: Scr. Metall., 1970, vol. 4, p.

529. 32. T. Shimomura, T. Kainuma, and R. Watanabe: J. Less-Common

Met., 1978, vol. 57, p. 147. 33. R. A. Mulford and U. F. Kocks: Acta Met., 1979, vol. 27, p. 1125. 34. I. S. Kim and M. C. Chaturvedi: Met. Sci., 1979, vol. 13, p. 691. 35. A. H. Windle and G. C. Smith: Met. Sci. J., 1970, vol. 4, p. 136. 36. L. J. Cuddy and W. C. Leslie: Acta Met., 1972, vol. 20, p. 1157. 37. P. R. Cetlin, A. S. Gulec, and R. E. Reed-Hill: Met. Tram., 1973,

vol. 4, p. 513. 38. D. J. Dingley and D. McLean: Acta Met., 1967, vol. 15, p. 885.

39..I.W. Eamgton and R. E. :~maiiman: Acta Met., 1964, vol. 12, p. 1313.

40. B. A. Wilcox and G. C. Smith: Acta Met., 1964, vol. 12, p. 371. 41. J. S. Blakemore: Met. Trans., 1970, vol. 1, p. 151. 42. B.J. Brindley and J. T. Barnby: Acta Met., 1966, vol. 14, p. 1765. 43. J. D. Baird and C. R. Mackenzie: J. 1ton Steel Inst., 1964, vol.

202, p. 247. 44. T. Takeyama and H. Takahashi: Trans. Iron Steel Inst. Jpn., 1973,

vol. 13, p. 293. 45. W. C. Leslie and A. S. Keh: Mechanical Working of Steel, 2, p.

337, Gordon and Breach Science Publishers, NY, 1965. 46. T. S. Ke and C. M. Wang: Sci. Sin., 1955, vol. 4, p. 501. 47. M. E. Blanter: Zh. Tekh. Fiz., 1951, vol. 21, p. 818. 48. P. Haasen and A. Kelly: Acta Met., 1957, vol. 5, p. 192. 49. R. W. Balluffi: Phys. Status Solidi, 1970, vol. 41, p. I1. 50. V. Kandarpa and J. W. Spretnak: Tram. TMS-AIME, 1969, vol.

245, p. 1439. 51. J. Chipman and E. Brush: Tram. TMS-AIME,, 1968, vol. 242, p. 35. 52. J. D. Baird: The lnhomogeneity of Plastic Deformation, p. 191,

ASM, Metals Park, OH, 1973. 53. J. D. Baird and A. Jamieson: J. lron Steel lnst., 1966, vol. 204, p.

793.

METALLURGICAL TRANSACTIONS A VOLUME 12A, MAY 1981--759