TEMPER-BRITTLENESS: PART I -UNALLOYED STEELS

B. R. NIJHAWAN
National Metallurgical Laboratory, Jamshedpur

Abstract (1) plain carbon steels are susceptible to

temper-brittleness, (2) temper-brittleness
It was until recently accepted that unalloyed
steels were not susceptible to temper-brittleness develops so rapidly that even drastic quench-
and that alloying elements introduced temper ing from a high tempering temperature is
brittleness . A directly opposite view has been set insufficient to suppress it,
forth in this paper , namely that unalloyed steels The comnton]y accepted lower toughness
are not only susceptible to temper -brittleness
of plain carbon steels compared with alloy
1}ut. also temper-embrittle much more .rapidly
th.rn alloyed steels based on the observations of
steels may be partly due to their being more
brittle transition temperature ranges. Temper- temper-brittle, as ordinarily heat-treated.
brittleness and strain - age-emhrittlement are dis- Addition of suitable quantities of the proper
cussed in relation to minor or residual chemical alloying elements, to be determined by fur-
constituents like nitrogen , oxygen , etc. The role
ther research, may provide a method of re-
of stabilizing elements like aluminium , titanium,
tarding temper-brittleness to such an extent
vanadium and of nitrogen in minimizing and often
wholly suppressing temper - brittleness and strain- that it will be of practical importance.
^geing is discussed at some length. The subject of It may be noted that the transition tem-
blue-brittleness has also been touched upon. The ob- peratures of alloy steels water-quenched
served phenomena are considered in terms of disloca-
from high tempering. temperatures do not
tions and certain gaps in the theory are pointed out.
necessarily represent material entirely free
from temper-brittleness; alloy steels may
embrittle to some extent on quenching from

J
T has until recently been universally the temper, though not as much as plain
accepted that plain carbon steels are riot carbon steels. All steel parts and specimens
susceptible to temper - brittleness and tempered at moderate or high temperatures
that alloying elements only introduce tem- may be somewhat tern per-em brittled. Tem-
per-brittleness . Zaffe and Buffumz have set per-brittleness may thus be an almost uni-
forth a directly opposite view . They sup- versal phenomenon in tempered steels,
port the possibility that plain carbon steels superimposed on or competitive with the
are not only susceptible to temper-brittle- tempering, but in a sense slower.
ness but are much more susceptible than Temper-brittleness in plain carbon steels
alloy steels , in the sense that they em- is markedly influenced by aluminium do-
brittle much more rapidly. The criterion oxidation which decreases strain sensitivity
referred to earlier involved notched-bar and lowers the transition temperature. These
impact tests only at room temperature, effects are associated with the composition
whereas the criterion now generally used of nitrides present in steel. If there are iron
retained the same heat treatments , but the nitrides, the steel is sensitive to cold-work
impact tests were made over a range of tem- and possesses a higher transition tempera-
peratures covering the transition from ductile ture than aluminium deoxidized steel. Al-
to brittle fracture. Based on their limited though considerable thermodynamic and ex-
tests in accordance with the latter criterion, perimental data support the fact that alu-
Zaffe and Buffumt arrived at two conclusions: minium nitride is present in aluminium-killed
187
185 SYMPOSIU_MI ON PRODUCTION, PROPERTIES & APPLICATIONS OF STEELS

VEE NOICH CHAIZP( IMPACI TESTS.

cast and wrought steels and that the presence
1 1HICK Pi JOE.
of this compound is directly associated with
goo
much lower sensitivity to strain and tempera- SPIC,AL CARBO14-
MAYGAPK SL SNECL
ture , the mechanism by which aluminium 80 ;RNIw CONTgo Ace
AND NORMALISED
nitride brings this about remains unclarifted d
and forms an improtant hiatus in our knowl- 60
edge of de.oxidation of steel and its low tern- w dp
perature properties in relation to temper-
brittleness. The mysterious behaviour of 20
aluminium nitride is further complicated
when it is realized that in suitably deoxidized
steels containing excess of 0.5 per cent carbon,
aluminium deoxidation has no significant
influence on low temperature properties and
in general on strain -ageing and blue heat
characteristics. The ultimate solution in
case of low-carbon steels may revolve round 60
the role of aluminium in fixing up the nitrogen -

as aluminium nitrides and throwing these

out of action so far as their activity on 40
temper- emhrittlemcnt and hardening is con-
cerned. For carbon in excess of 0.5 per cent
in steels, the effects of Al-1` may be over- a 2
shadowed by the influence exerted by carbon
on temper-brittleness. Figs. 1 and 2 illus-
trate the effect of aluminium deoxidation
in relation to transition temperature ranges. - -410 0 40
Kruger" has recently investigated the sub- TEMPL A!)QE `t,
iect of temper- brittleness of plain carbon
Fro. I ( After W . BARR, West of Scotland, Iron &
steels. The duality of basic-Bessemer steels Steel Institute , Presidential Address 1950)
has improved considerably in recent years by
blowing with oxygen-enriched air, or steam, so
oxygen and CO., or a mixture of these. Rim-
ming and killed steels from the basic conver-
Go
ter blown with oxygen-enriched air and a mix-
ture of air, oxygen and steam were compared
by Kruger2 with ordinary basic-Bessemer,
0
0.11. and electric furnace steels with respect
to composition and mechanical strength
before and after cold-working. The nitrogen ac
and phosphorus contents were approximately
the sane in all the steels, but the oxygen con-
tent differed considerably. Tensile strength,
40 to o to 40
necking, elongation and notch-impact data TEM}ELIklT &E .C
increased less after cold-working than those
2 - IMPACT VALUE ON TESTING TEMPERATURES
of ordinary basic-Bessemer steel. Ageing FIG.
(BARR AND HONEYMAN, Institution of Engineers
was also studied and the dominating influence & Ship Builders, 1948, 72 )

lm•^wnw . ^•l^n.^l ^ 0w0"1 W F p 11^ IMLl^ll^lgrl o m r ^wgn ^ ^ I , p q 1iglT IpqIIIII WWWR ➢W I

 NIJHAWAN - TEMPER-BRITTLENESS; PART I 189

of nitrogen and phosphorus was established. lity and ability to withstand shock; and the
With high nitrogen and phosphorus there yield point will be more sharply defined.
was a marked increase in tensile strength Ageing indicates a tendency in the direction
after cold-work. The effect of phosphorus of equilibrium after the disturbance caused
alone appeared to be stronger than that of by quick cooling as in quench-ageing or cold-
nitrogen alone. Steels blown with steam working as in strain-ageing. Its cause is
plus oxygen added to the blast were superior attributed to a breaking up of the super-
and very similar in all respects to O.H. saturated solid solution. Strain-ageing, in
steels. Thus Kruger2 strikes a rather dif- addition to rendering the steel harder and less
fering note concerning the role of nitrogen ductile, eliminates the influence of temper
and phosphorus in strain -ageing of these rolling cycles designed to eliminate stretcher-
steels. strain markings. Hardness increase pro-
duced by strain-ageing is not by any means
equal to the loss in ductility, which frequently
Strain- ageing of Steel in Relation to
Temper- brittleness far exceeds the relatively small gain in tensile
strength and hardness.
The second edition of Webster's New Many aspects of the phenomena of strain-
International Dictionary defines the verb ageing are those to be expected of the age-
age ' as follows : ` AGE ( verb intrans.) : hardening process. It has to be assumed
1. To become old; to grow older; to show that the solid solubility of a constituent or
marks of age; to undergo changes with age constituents is decreased by cold-working,
or the lapse of time. 2. Specif.: (a) To providing a precipitation potentiality or that
suffer with lapse of time a diminution of straining generally accelerated such precipita-
essential qualities or faces; as, an incan- tion. The nature of the ageing agents has
descent lamp or a transformer ages; (b) To been well argued for the last two decades
become mellow or ripe; to acquire a de- and is still on the anvil. Strain-ageing has
sirable quality by standing undisturbed for been observed in steels very low in carbon
some time , as wine or varnish ages; (c) Metal. and it has been reasoned that the ageing
To remain or stand undisturbed after or agent may be oxygen; this point of view
during heat treatment, so that molecular or received ostensible support from the develop-
crystalline adjustments may occur; as, an ment of deoxidation techniques for steels
alloy ages.' with low strain-ageing sensitivity as in the
There are two important points brought case of Izett and stabilized steels. However,
out by a study of this definition. The first oxygen is not now believed to be directly
is that time is the essential element in ageing; responsible, although it may exercise sec-
and the second, that the effects may be ondary effects on the solubility of carbon
either deleterious or beneficial. and nitrogen in a-iron and their quantum
The phenomenon of strain-ageing in steels needed to set up ageing. Strongly de-
is closely linked with yield point phenomenon oxidized stabilized steels do not exhibit
and it would be worthwhile outlining salient strain-ageing - this may be due to the
features thereof, in relation to chemical cons- ' fixing ' up of nitrogen by the deoxidants
tituents present in steel of mild and struc- and its resultant inactivity thereafter, Thus
tural varieties. while precipitation of ageing agent proceeds
Various effects may be produced by ageing. slowly at room temperatures and much more
For example, the material may become accelerated at higher temperatures, say at
harder and have a higher tensile strength; 300°C., this precipitation causes the crystal
it will correspondingly suffer a loss of ducti- lattice of iron to be distorted and the metal
190 SYMPOSIUM ON PROD UCTION, PROPERTIES & APPLICATIONS OF STEELS

offers increased resistance to movements 0.O2 C Sree( 0, Cj/rd+0d1%Ti'

along the slip planes. At temperatures

above 300°C., the precipitated phases coalesce
which, on being totally disentangled from the
ferrite lattice, no longer exert any hardening
effects thereby causing softening and over-
ageing effects. The exact causes have yet
to be finally established, and although of
recent years stabilized non-ageing steels have
been developed, their cost and yield factors
are often prohibitive, so that the consumer
0 t0 20 30 10 z0 30
tends to eliminate a proportion of his diffi- Pertene. E/ongolion Pereenf,Ebyvhon

culties by, as far as possible, finishing the

FIG, 3 - EFFECT Of ADDITION OF Ti UPON YIELD
deep-drawing and pressings before the steel POINT AND UPON STRAIN-AGEING OF LOW-CARBON

has appreciably aged or the influence of STEELS. Full line: Normal stress elongation curve.
Broken line: Stressed to 6 per cent elongation, aged
temper rolling has worn off. at 250"C. for 1 hr. and re-stressed
Notable work has been done by Thomas
and Leak-, Jones and Owen-Barnett'^ , and 60
I1undy"'6 on these subjects.

Strain- ageing and Yield Point Mild Stee l

O
M N
Strain-ageing and yield point phenomena
do not appear so much to be directly re-
lated as to arise from the same cause, namely
due to small additions of carbon, nitrogen 20 Stabi//ze d
,$j Mater/o/
and oxygen. Suppression of strain-ageing m co a.
0
by removal of these or converting them to
.S'frgin, Per eel) t,' 2h . = 9:
inert compounds also suppresses the yield
point as illustrated by the 0.02 per cent C FIG. 4 - TII ^'. ('IIANGG of T1lE STRESS -STRAIN DIA-
GRAM OF TWO STEELS ON AGEING AT ROOM
steel with Ti added ( pies. 3 and 4 ). Fur- TEMPERATURE FOR DIFFERENT PERIODS OF TIME
ther, as in strain-ageing, oxygen alone does SUBSEQUENT TO COLD-WORKING ( KENYON AND
BURNS
not seem to produce a yield point, but acts
indirectly by enhancing the effect of carbon
and nitrogen. In unstrained irons, the yield
point can he absent, even though strain- of these conditions and the mechanism by
ageing properties are still detectable, e.g. for which the precipitate acts are quite unknown.
0.003 per cent C, but in this case a yield At 300°C., which is above the optimum tem-
point is observed after strain-ageing. perature for strain-ageing, the yield point
Neither strain-ageing nor yield point is a disappears, presumably because the preci-
property of pure iron but is due to carbon pitate grows too coarse to play its effective
and nitrogen. Strain-ageing arises when role.
these carbon and nitrogen are in super- Early studies on the effect of hot-working
saturated solid solution, while the yield point and of tensile testing at these temperatures
appears to be caused, under certain condi- disclosed a characteristic brittleness in steel
tions, when these have been precipitated in a in the vicinity of 250°C. - the ' blue-brittle-
very finely divided state. The exact nature ness phenomenon '. For tensile tests at

77-
 NIJIIAWAN-TENPER-BRI'r1'LENESS. PART I 191

elevated temperatures, the ' blue-brittleness ' 5-0$% C BessernerSteel

range is lower the slower the speed of 180 (Strain- Ageing).
of elongation - similar results were obtained
on impact testing, where in ' static ' notch- *0.13% C Steel, Al killed
bar tests a brittle range was observed around (Non -Strain -Aq eit q ).
250°C., but in impact tests was observed to
he shifted to 55°C. Strain-ageing and blue-
brittleness are only different aspects of the
cb
strain-ageing process itself, i.e. at 250°C. the
10 0
ageing process is rapid enough to occur con-
0 10 20
currently during the test; at lower tempera- Percent, Cold-working CBy Compression'
tures the time is not enough, whereas at
FIG. S - A ` NON- AGEING ' STEEL ,
higher temperatures it is sufficient for strain- PRODUCED BY
A DEOXIDATION WITH ALURIINIUM , WORK- HARDENS
ageing to occur which manifests itself through MUCH LESS RAPIDLY THAN ORDINARY LOW-CARBON
STEELS
characteristic jaggedness in the tensile curve
believed to be due to rapid strain-ageing on
successively active slip planes. Non-ageing
steels work-harden much less rapidly than
ordinary steels at room temperatures and
S Ordinary
possess a much less marked blue-brittleness hli ld
region, whilst highly purified iron free from Sree1
strain-ageing is also free from ` blue-brittle-
ness Optimum additions of aluminium h
( 2-3 lb. per ton ) to steels completely sup-
pressed blue-heat hardening, In steels de- "Non-Stroin-Ageing"SI-eel
oxidized with aluminium, titanium or zirco-
nium, the nitrogen occurs as nitrides of
30L
these elements, and since the solution be- 0 100 200 300
haviour of these nitrides is radically different Temperature °C
from that of iron nitrides, the ageing and
FIG. 6 --- ` NON-A GEING ' STEELS SHOW A MUCH
blue-heat characteristics of such steels under- LESS MARKED BLUE-BRITTLENESS REGION THAN
go material changes by the deoxidation ORDINARY LOW -CARBON STEELS

practice. Although control of ageing is of

great practical significance, it is not ipso
facto an indication of unsatisfactory steel centrations contributed to the formation of
quality. Fig. 5 depicts the work-hardening precipitate nuclei, thus increasing the rate of
characteristics of a Bessemer steel and a precipitation,
killed low-carbon steel. Fig. 6 depicts the A radically different explanation of the
' blue-brittleness ' characteristics of a mild mechanism of strain-ageing was advanced by
low-carbon steel and stabilized non-strain- Andrew and Lee". They regarded that the
ageing steel. strain induced localized allotropic trans-
The precipitation theory discussed above formation of alpha-iron to the gamma-phase
appears to offer a satisfactory explanation so that some carbon was dissolved during
for the phenomenon of age-hardening. Mehl the work-hardening process. The films of
and Jetter7 claimed that plastic deformation austenite thus formed at the slip planes will,
created loci of high energy content in the under the quenching effect of the surrounding
crystal lattice, and that these energy con- mass of metal, transform to minute crystals
192 SZ`1iPOSiiJ^I O` PRODUCTION, PROPERTIES & APPLICATIONS OF STEELS

of alpha-phase and this reaction would result is applied to a metal, the deformation grows
in an increased resistance to further slip along steadily greater as the force increases until
the gliding planes . The net effect is pro- the yield point is reached. Then the metal
gressive strengthening of the ferrite in the suddenly gives way, and the deformation
course of time following cold-work. The continues to increase even if the force is re-
age-hardening theory advanced by Andrew duced. The yield point marks the transi-
and Lee, although apparently supported by tion from elastic to plastic behaviour. The
some clilatometric and X- ray evidence, did theory is that in the elastic range the forces
not receive wide acceptance . It was pointed are not large enough to pull the dislocations
out that this theory did not satisfactorily loose from the Cottrell atmospheres. At a
explain the work -hardening and ageing be- certain critical value the dislocations are torn
haviour of austenitic steels and was not in from their anchor, and may then be kept in
agreement with the evidence that strain de- motion by a smaller force. Metals without
creased the stability of austenite and the abrupt yield points are now being made
solubility of carbon in ferrite ; further, it simply by purifying them enough to eliminate
shed no light on the role played by nitrogen the Cottrell atmosphere from all but a few
in the strain - ageing of steel and in any case it of their dislocations. During ageing, the
could not explain the yield-point phenomenon carbon and nitrogen atmospheres have time
observed in certain face - centered non-ferrous to move and anchor the dislocations and
alloys. thereby introduce strain-ageing and strain-
Nadaia and several German investigators age embrittlement. Similar phenomenon is
have advanced a hypothesis that the yield- supposed to take place in quench-ageing.
point phenomenon in steel is attributable to The simpler picture explaining the age-
the existence of an ultramicroscopic grain- hardening after straining appears to be that
boundary film of carbide which serves as a when iron which contains no carbon or nitro-
skeleton for the ferrite grains. gen is strained the metal undergoes plastic
A new theory , which explains strain- deformation like other pure metals showing
ageing by the anchoring and dislocations of no yield point or discontinuity in view of the
solute atoms such as carbon and nitrogen, position that there is nothing to prevent
has been recently formulated . Evidence in movement along the 110 slip planes. Inter-
support of this theory is given by Cottrell' stitial atoms like carbon and nitrogen
Nabarro1l , Cottrell and Churchman '-, Bilby13 located at these slip planes increased the re-
and Harper14 . Because of their distortion of sistance to slip. Thus stress to initiate the
the lattice , dislocations tend to attract the slip is raised on these slip planes. Once,
foreign atoms in a crystal , such as the carbon however, it starts, other planes follow like-
atoms in steel. A foreign atom that is wise. On ageing after straining, carbon and
larger than the lattice atoms will tend to nitrogen atoms diffuse back to their original
move to the tension side of a dislocation, relative position and then the phenomenon
where there is more room between neigh- of yield point reappears.
bours. :Similarly, a smaller foreign atom
will tend to migrate to the compression side Nitrogen in Steel
of a dislocation . This effect was first pointed
out by Cottrell and the concentration of The iron-nitrogen system is of particular
foreign atoms is called a ` Cottrell ' atmo- interest, since it provides several examples
sphere. It is supposed to explain for the of a highly metastable equilibrium of con-
first time the phenomenon in metals known siderable industrial importance. The solu-
as the yield point. When a distorting force bility data for nitrogen in solid iron suggest

i
P7 FWRM
 NI JHAWAN - TEMPER-BRITTLENESS: PART I 193

that it obeys Sievert's law, i.e. the solubility 80o

varies as the square root of the pressure. A
knowledge of the solubility relationships of
nitrogen in both crystal forms of iron is of
importance. An idealized version of the re-
coo
sults of several investigators is shown in
frig. 7; two points are notable, firstly the
markedly higher solubility of nitrogen in
gamma compared to alpha-iron, and sec-
ondly, the solubility increases with tempera-
400
ture in alpha-iron.

32

200
24

0-02 0-06 0•J0

ti
NITROGEN WEIGHT °Jo
16

O FIG. 8 - SOLUBILITY LIMITS OF NITROGEN AS A

O FUNCTION OF TEMPERATURE

absorption by austenite, until reaching a

600 1000 0200 1404D 1600 critical aluminium content which depended
TEMPERATURE. *C.
upon temperature. Above this critical con-
centration, the nitrogen content increased
I+IG. 7 - IDEALIZED SOLUBILITY CURVES FOR
NITROGEN IN IRON FROM THE WORK OP VARIOUS sharply, presumably due to the formation of
INVESTIGATORS. NOTE (a) THE HIGHER SOLUBILITY
a new phase, aluminium nitride. This
OF NITROGEN IN GAMMA-IRON AND (b) THE
SOLUBILITY INCREASE WITH TEMPERATURE IN critical aluminium content bears further rela-
ALPHA-IRON
tion to its role for austenitic grain re-
finement.
Concerning the effects of carbon and nitro-
More recent data by Paranjpe15 and his gen on damping capacity of ordinary steels,
colleagues express the solubility relationships only very small quantities (0.007 per cent of
of nitrogen in ferrite as shown in Fig. S. A carbon and 0.001 per cent of nitrogen) of
notable point from this curve is that there is these elements can dissolve in iron at room
a sharp reversal of solubility above about temperatures; but the two different atoms
600°C. occupy preferred positions in the iron lattice
Darken16 has investigated the solubility interstices. There is an alternative position,
of nitrogen in austenite as influenced by with actually more room, available within
various alloying elements and concluded that the lattice, but the other is the preferred
carbon , manganese , sulphur and phosphorus position. On the average some kind of dyna-
in amounts normal to low- carbon mild steels mic equilibrium is maintained and along any
have negligible effects . Manganese contents x, y or z-axis, taken at random, there will be
up to 1.5 per cent were claimed to exercise in either position a statistically constant
no effect . Darken16 further indicated that number of solute atoms at any given instant.
aluminium showed no influence upon nitrogen If the iron lattice is extended along one
 194 SYMPOSIUM'I ON PRODUCT[ON, PROPERTIES & APPLICATIONS OF STEELS

axis, more room will be provided for the pique. They worked with pure iron con-
carbon and nitrogen atoms in this direction, taining: C, 0.002-0.004 per cent ; Si, 0002-0.003
and these will desert their positions along per cent ; Mn, 0004 per cent; S, 0.0004-0.006
the y and z-axes in preference fort he positions per cent; P, <0001 per cent; Al, 0001 or less;
on the x-axis. When the iron lattice is al- Oz by wt., 0.001-0.002 per cent; N, 0001-0.002
lowed to retrun to its original length, the per cent; H-trace to <0.000005 per cent.
solute atoms will be crowded back to their This excellent work has thus been clone with
former positions. Energy required for atomic a purity of material that has rarely been
movements of this kind must be obtained exceeded"-21.
from the vibrational energy and will result Pure polycrystalline iron shows a brittle
in a reduced amplitude of vibration, i.e. by behaviour in a notched-bar test at about
damping. Two opposing factors operate -15°C., but does not become brittle in plain
with rise in temperature; the number of tension until much lower temperatures of
solute atoms may increase, but there also about -150°C. are reached, while a single
occurs an expansion of the iron lattice. crystal of iron at -196°C. is either ductile
These two opposing factors combine to give or brittle according to the direction of the
it maximum in the damping capacity when stress.
plotted against temperature. I t is shown Phosphorus exercises a detrimental effect
that the peak in the damping curve is exactly on the properties of iron because it raises the
proportional to the content of carbon or temperature of transition from a tough to a
nitrogen present. Both elements have a brittle fracture. The behaviour of poly-
deleterious effect on the toughness of steel in crystalline material is further complicated
certain circumstances. by the fact that some elements, e.g. oxygen,
In the case of rimming steels Epstein" has nitrogen or phosphorus in iron, reduce the
claimed that the addition of 0.03-0.15 per brittle fracture stress by weakening the grain
cent of v-uiadiuln and/or 0.3 per cent of boundaries to such an extent that brittle
chromium to a rimming steel confers a high fractures may take place partly or wholly
resistance to strain-ageing without seriously along the grain boundaries. The sensitivity
affecting the rimming properties of the steel. of this effect to heat treatment is illustrated
Epstein, Cutler and Frame" have supported by the remarkable grain-boundary weakness
this claim with vanadium-treated rimming that can be developed in iron-nitrogen alloys.
steels. Steel differs from pure iron principally in
In Great Britain, the subject has been containing carbon and manganese. The
studied on an extremely scientific and syste- effect of carbon on iron depends upon the
matic basis, viz. to start with as pure iron as heat treatment and on the quantity present,
can be experimentally obtained and study and when slowly cooled from the austenitic
the effect of trace impurities ( with which condition, steels of all carbon contents are
iron is commonly alloyed in practice to yield brittle at -196°C. while steels of very low
steel ) on it. This excellent work undertaken carbon content ( below 0.05 per cent carbon )
at the National Physical Laboratory, Ted- are much improved both in strength and
dington, has yielded valuable additions to toughness at low temperatures by being
our knowledge of this subject. cooled rapidly. Steels containing enough
Perfectly pure iron has never been pre- carbon to be hardened are embrittled by this
pared. Realizing this the National Phsyical treatment. Excellent properties can be ob-
Laboratory workers have adopted a standard tained at ---196°C. for steels containing 1-2
of purity which could be regarded as suffi- per cent manganese and 0.05 per cent carbon
cient with the help of a special vacuum tech- suitably heat-treated.

i a ^ ^ r i u 'I i m F ^ ^ R q mi qF "Ign „ 14.1 .1 n i 'a'! -,- -I^11A , PPP PIPIP OF

`FMAI/ i lI.?.H III
 N1JHAW,\N - TEMPE R-ImI I T1.L NJ S S: PART I 145

It is considered worthwhile to further the brittle state, it is pointed out that the
elaborate the excellent work of National great toughness, with increased yield stress,
Physical Laboratory investigators. Under that can he conferred on iron by retaining as
the conditions employed in their work, the little as 0-04 to 0-05 per cent of carbon in the
addition of more than about 0.003 per cent ferrite by quenching, is diminished by ageing
of oxygen to pure iron causes failure in the at room temperature. Ageing is accom-
brittle condition to occur increasingly along panied by the precipitation from the super-
the grain boundaries, with a consequent pro- saturated ferrite of carbon-rich areas and can
nounced decrease in brittle fracture strength be substantially prevented by the addition
in plain tension at -196°C. and a progres- of about 0-75 per cent manganese. These
sive increase in the transition temperature results obtained at Teddington render in-
in impact. The effect of oxygen on impact telligible the effects of manganese content
transition temperature was most marked and heat treatment upon the susceptibility
between about 0-003 and 0.011 per cent of mild-steel plates to brittle fracture and
oxygen. The rise in impact transition tem- indicate how the susceptibility may be con-
perature with increasing oxygen content was trolled in practice.
largely due to a decrease in brittle fracture Nitrogen is often considered to be similar
strength. At the higher oxygen levels, how- to carbon in its effects on iron , but in an iron
ever, because of the small temperature de- containing 0-01 per cent of nitrogen a number
pendence of the yield stress and possibly also of effects quite dissimilar to those of carbon
of the brittle fracture strength above, say, have been found. In particular, the grain
room temperature, a rise in impact transition boundaries appear to be greatly weakened
temperature could be accounted for by a by the presence of nitrogen in solution in
small change in either. ferrite, and the precipitation of iron nitride
No intergranular phase has been found to appears to be less effective in decreasing the
explain the grain-boundary weakness con- ductility of iron than is the precipitation of
ferred on iron by oxygen, so that the precise iron carbide. The degree to which the grain-
mechanism of the embrittlement is not boundary embrittlement occurs is strongly
known. Definite evidence of iron oxide in- influenced by the rate of cooling through
clusions was first obtained with 0-007 per certain ranges of temperature.
cent of oxygen ; raising the oxygen content Phosphorus has long been known to re-
above this level increased the number of duce the ductility of iron. Its effects prove
oxide inclusions, which were randomly dis- to be in many ways similar to those of nitro-
tributed. gen. Grain-boundary embrittlement is pro-
The addition of manganese was found to duced to a degree that depends upon the
counteract very largely the euibrittling action quantity of phosphorus and the rate of
of oxygen, but some grain-boundary failure cooling. Some of these results obtained by
remained. Brittle fracture in a normalized National Physical Laboratory workers have
mild steel occurred partly along grain been depicted in Figs. 9-13.
boundaries, and its transition temperature The problem of brittle fracture of mild
range of in the Charpy impact test steels is such a bite noire in the design of large
was higher than it would be otherwise be- welded structures such as ships, bridges, etc.
cause of this intergranular weakness. General features of the problem are now so
Concerning the important effects of small well established that further elaboration of
quantities of carbon on the mechanical prop- the subject-matter would be largely re-
erties of high purity iron, with particular dundant. One characteristic which should
reference to the transition from the tough to be recalled is that brittle fracture has a
196 SYMPOSIUM ON PRODUCTION, PROPERTIES & APPLICATIONS OF STEELS

5o too
1RtOUCT 1OfI OF AREA
80

40 60
"u ELONGATIO

40
/
50 20

a TEH51LE STRGN G'r H.

YIELD OR PROOF
10 TAES .

O'
-200 -100 -50 0 20

TEMPERATURE OC

1`IG. 9 - TENSILE PROPERTIES OF PURE IRON OVER A RANGE OF TEMPERATURE

32 160

FUSN CE COOLED 95OC.

120

80
AIR COOLED 95 e .

40

WAT ER QUENCHED
40
sso^

V 0
-60 -40 - 20 0 20 4D 60.
80
0 0-02 0-04 0.06 0.12
TEMPERATURE 0C. CARBON %

FIG. 10 - THE IMPACT TRANSITION TEMPERATURE FIG, 11 - CHARPY IMPACT TRANSITION TEMPERA-
IN PURL: IRON TURES OF IRU -CARBOY ALLOYS
 N1JII TWA - `l'1;h1F'L R-Bkl`'TLl Mass: PART I 197

?ENSILE PROPER T%ES

ROOM TEMPERA3uRE -196 C

5 0- 90
REDUCrIONOF AREA

G1
40 70

05°/0
L1 POOF
STRESS TEN SILE SIREN-
fl, 30 G+TH 5o
16 • ELONGA TION
0
z
a Z0 T ENSILE STRENGT M
30

10
EL0NGAT%OM
LOWERJ 1( IELb O R REOuCT ^ o4 Q1~
PROOF III STRESS AR A

0 0.004 o 1 oo6 01012 0.004 0006 o•o12..

OX GEw %
FIG. 12 -EFFECT OF OXYGEN IN IRON ON TENSILE PROPERTIES

distinct crystallographic connotation, in the dependent on temperature or strain than the

sense that the fracture path follows the ( 100) yield stress or flow stress. In general, the
planes or the cube faces of the iron lattice, cleavage strength is greater than the shear
while normal ductile failure appears to fracture strength of the material, and obeys
possess no crystallographic preferences. The Sohncke's normal stress law. When a notch
term brittle fracture can be a misnomer since is present, larger sections on either side of
it may be preceded by considerable plastic the root constrain the material thereby
flow before cleavage of the grains follows. generating a hydrostatic tension, The ratio
In ordinary structural steels cleavage fracture of normal stress to shear stress is, therefore,
appears at about 0°C. in notched-bar impact raised relative to the plain bar. By this
tests, and about -160°C. in tension. In means the cleavage strength of the material
studies of brittle fracture, therefore, it is not may be attained before the onset of ductile
surprising that an acceptable theory of notch failure, Recent evidence23 rather suggests
brittleness will command attention. The that in the special case of single crystals
hypothesis which Orowen22 employs to Sohncke's normal stress law is not obeyed,
account for this property postulates the although the unique crystallographic iden-
existence of a cleavage strength which is less tity of the cleavage plane is retained.
198 S"lPOS1uM ON i- ROi) CTiON, PROPERTIES & APPLICATIONS OF STEELS

CHARPY IMPACT PROPERTIES the equations. The piling of dislocations

behind an obstacle, function of grain bound-
aries acting as barriers to moving disloca-
tions, explanation of the sharp yield point on
` avalanche ' basis under the stimulus of the
clustering atmospheres of solute atoms
anchored around dislocations, similar to the
ionic atmospheres of the Debye-I-Iuckel
theory of electrolytes, though reasonably
so argued and well presented, all tend to be
highly speculative. The metallurgist is al-
lowed to escape with the impression that
40 what is being attempted in terms of physical
concepts is the tendency to create master
equations from which all metallurgical obser-
0 vations could be deduced or manoeuvred.
These physical theories including that of
-20
` dislocation ' move fast - Moat's recent at-
tempt to explain the fatigue phenomenon on
0 0.004 X0/.040 0-012
the basis of the creation of extra vacant
OXYGEN /e
lattice sites under alternating stresses is an
lie. 13 - LFFEC'r O! OXYGEN IN IRON ON IMPACT illustration of the mobility.
PROPLRIlES
There are often other characteristics intro-
duced by micro-additions of stabilizing ele-
ments like aluminium, vanadium, titanium,
A characteristic cleavage strength seems to
etc., to steel in conferring high impact notch
exist, however, for certain polycrystalline
toughness, reducing strain-ageing effects in
metals.
steel; micro-additions of boron in improving
harden ability, etc., that have yet to be fully
Physical Concepts accounted for in modern physical theories.
The writer may perhaps illustrate this with
It is not the purpose of this paper to dis- a specific case, viz. the influence of carbon in
cuss the phenomena of temper-brittleness or relation to ' blue-brittleness ' in steel. In
strain-ageing in terms of dislocations. How- the regions of 250°-300°C. the mobility of
ever, some gaps tend to define themselves carbon solute atoms is assumed with the
with a view to present the metallurgical result that a moving dislocation is surrounded
standpoint. The physical approach very by an incomplete carbon atmosphere which
often in proceeding to explain the observed causes a viscous drag on it. Thus the drag
behaviour of metals tends inevitably to is smaller faster the flight of the dislocation,
include a number of pure speculations, some but at the same time higher the temperature,
of which are not unoften short-lived. Some greater is the mobility of carbon atoms and
may feel concerning the role of the theory thus larger is the stress required before the
of dislocation in explaining well-established dislocation breaks away catastrophically.
phenomena, like strain-ageing, yield point, However, at temperatures higher than the
work-hardening, etc., that mathematical dev- ` blue-brittle regions ' the whole phenomenon
clopments of the theory have run far ahead changes since the blue-brittleness ' itself
of the knowledge of the values denoted in disappears. Here it is assumed that each

1{1-MR I R1 { nl l l' !Ii11^^^.^6Tf wn1^ A^i I'' '?

 ^IJH_S1t.S\ - l'I:.l]I'l:R 131t1'1"IL1_-NESS: I'.AR1' 1 199

dislocation is accompanied by its atmosphere, 10. COTrRELL, A. H., " Effect of Solute Atoms on
the Behaviour of Dislocations", in Report of
but the carbon atoms are then so mobile as
a Cf'nferencc on Strength of Solids ( held at the
not to appreciably retard the flight of the H. 11. Wills Physical Lab oratory, University
dislocation. One looks in vain for the effects of Bristol, on 7-9 J my 1947 ) ( The Physical
of stabilizing elements like aluminium or Society, London ), 1948, 30-38.
titanium in eliminating ` blue-brittleness ' 11. NABARRO, F. R. N., " Alechanical Effects of
Carbon in Iron ", in Report of a Conference on
to be explained in terns of the flight of dis-
Slrenglh of Solids ( held at the 11. 11. Wills
location vis-a-vis mobility of carbon solute Physical Laboratory, University of Bristol,
atones. Similar remarks can equally (told on 7-9 July 1947 ) ( The Physical Society,
good for other phenomena, such as non- London), 1948, 38-45.
strain-ageing of stabilized steel to be 12. COTTRELL, A. If. & CHURCHMAN. A. T ., J. Iron
k7, Steel Inst., 162 ( 1949 ), 271-276.
accounted for in terms of dislocation
13. ISILUY, B. .5., Sheet Mclccl Iced., 27 ( 1950),
theories.
707-718.
14. HARPER, S., " Precipitation of Carl>un and Nitro-
gen in Cold-Worked Alpha-iron ", PAY. Rev.,
References 83(1951 ),709-712.
15. PARANJPL•', V. G., 185 ( 1949 ),
1. I-AFFE, L. D. & Burrtr*I, D. C., .11elals Tech., 252.
( Dec. 1948). 16. DARKEN, L. S., Smini, H. 1'. & FILER, L.
2. KRCOER, A., Stahl it. Eiden, 74 ( Dec. 1954), 191 ( 1951 ), 1174.
1757-1766. 17. EPSTL^'IN. S., U.S. Patent 2,356,450 ( 22 Aug_
3. TI10M:1S, \1'. R. & LE.u, G. It., J.I.S.I., 180 1944).
19. KEL•'S, \\'.I'. & l1UI'KI s, B. E., J.I . S.I., 172
( 1950), 155.
4. JONES , B. & OWEN - BARNL•' TT, It . A., J. I.S.I., 1952 ), 403.
180 ( 1950) ; 177, 209. 20. H0PKINS, 13. L. & TIPLER, H. It., f.IS.I., 177
5. HuNDY, B. B., 178 ( 1954 ), 34. ( 1954 ), 110.
6. HUNDY, B. B.. J.LS.L, 178 ( 1954 ), 127. 21. RERS, W. I'., HOPKINS, B. L. & TIPLER, H. It.,
7. MEHL, It. F. & JETTER , L. K., Age-hardening of J.l. S.I., 169 ( 1951 ), 157.
Metals ( A.S.M.), 1940, 342-438. 22. OROWEN, E., Reports on Progress in Physics,
8. ANDREW , J. If. & LEE, I1., J.I. S.I., 145 ( 1942 ), 12 ( 1948-49), 185.
23. DERUYTTERE & GREENOUGH, N ature (July
153.
9. 2%rADAI, A., Plasticity, 349. 2501, 1953 ), 170.

41

iF