f
f
r
’7
fl
7’-
___—E—
.——e
_ W,

7 _
47 , _ ,7
7 7, ..

|
1

h

I
i

H

\ 1‘ \E

__I
N
(D

 

A STUDY OF THE PHYSICAL
PROPERTIES OF MEDIUM CARBON
STEELS

Thesis for the Degree of M. 5.
Kenneth J Trigger
I935

 

\
0

o

w: tum/1 63 Y

 

..",

(lazy

 

h
1",,“ ‘-.%“<

‘xpi-wr‘": “‘5 "

"‘f -' ow? {1:115}; “
‘ 1:14.): .\II :(‘I‘ .’_‘ -‘
.9" “ 'n’fi-M‘V‘.

.. n"? - ".’ '3‘.“-

4...!

"5 t “
.tg'fi-‘TI ‘
div-<1”):

r‘ H.‘ '-

v ‘. ’
$.d“‘.‘..‘z
D . .I l' h"
\ . r. .. » "
I‘,""o)f “fil‘ '(:$c;‘ 3"; ' .f
“r; «.W H141, L:
3.» WWW“.
r. "Haw“! 3W! ’V

. .‘ .- m
“’: 1‘ “.1 . a? "3‘ ‘ "
5" v9“? it". {iii}? 1'}. g‘%¢}$
w 4. as. 3H i a
M?" my gig ‘9

I
. .. ~ '3}.-
.‘é‘i WW ~. ""1".“ ’4
j "- h.‘ '§~‘{- ' ‘

A. I. '“’1"'\’1-’r&'
n . «.239 1"}. - ~
‘, . . .. ‘
1*t‘ ‘K'g ‘. H at: b 9

1“»

 

In”. '-
10"th
EEL'F’

have“

 

 

L}

 

A

i

Nusa=kuv ‘ ”I.

.

_‘

A STUDY OF THE PHYSICAL

PROPERTIES OF HEDIUH CARBON STEELS

A Thesis submitted to
the Faculty of Kichigan State College
in Partial Fulfillment of the Requirements

for the Degree of Easter of Science

Kenneth J; Trigger

1955

 

\leH 11 I ‘

n tying-v.

,uIy m0..v.m.l:.h flu

1‘!

THESIS

ACE’EIOE’EEDGI‘JIEIIT

r“he writer wishes to express his sincere
appreciation for the kindly assistance given
by Professor F. G. Sefing during the progress

of this work.

98?}??20

TABLE OF CONTENTS

Page

1. INTRODUCTION . . . . . . . . . . . . . . . . . l

2. SCOPE 13.23) PUKPOSE o o o o o o o o o o o o o o 5

5. EQUIPMENT USE . . . . . . . . . . . . . . . . 4

4. FUEUACE ATHOSPHERE & IT EFFECT ON THE SCALE

TIT I 01:13.2 8 S o o o o o o o o o o o o o o o o 5
5 . 3.1.4.}; JJLS EVIL; STIGELTED o o o o o o o o o o o o o 9

6. GRAIN SIZE SPECIFICATIONS & PHOTOEICROGRAPHS . 10

8. TESTS ON THE IIPACT SPECIIEE . . . . . . . . 28

9. TESTS ON THE SMALL CYLINDERS . . . . o . . . . 51

10 . COIICLUESIOIIS o o o o o o o o o o o o o o o o 55

11 . BIBLIOGRJ‘LP}IY o O O O o o o o o o o o o o o o o 56

 

{‘1 It],

. 1ygflru nfiiuvvfir.‘

IHTEODUCTION

It has often been observed that while two steels may
meet the same chemical Specifications they may show vastly
different physical prOperties, particularly with reSpect to
toughness, depth of hardness, distortion, forging properties,
and susceptability to heat treatment. This variation of
steel has caused a great deal of trouble to the users of
plain carbon steels. O

In view of such differences in steel, metallurgists have
come to the conclusion that the chemical analysis of steel is
not the only important factor determining its prOperties but
that the grain size must also be considered. With such large
quantities of steel required by present day production methods
it becomes imperative that the steel mills manufacture a
product which can be relied upon to give uniform results. At
present the study of grain size control in carbon steel and
its effect upon the physical prOperties is one of the most
important and most widely discussed phases of steel metallurgy.

T'It is generally believed that it is grain size or perhaps
some still more fundamental condition which is responsible for
the variations in physical prOperties of steels with the same
chemical analysis. With reSpect to the factors controlling
the causes of inherent differences in steel, the following is
quoted from ”Deoxidation of Steel" by C. H. Herty, Jr.(l)

"In the steel entering the mold and the reactions which

take place during solidification it will be noted particularly

 

5111 h

‘II .L

1L. 1. v.

‘b.s'lp £533.. . hi?” 1.3%,!“ Ewifl“

.Io

that, aside from chemical composition, there are non-metallic
elements in solution and in suSpension and that the amounts
of material in solution and the type of suspension are due
largely to the deoxidation methods employed. One of the
prOperties of steel that might be affected by these solutions
and suspensions is that of grain size.

"It is the general concept at this time that fine-grained
steels are the result of the presence of a fine suSpension
which furnishes nuclei for recrystallization during heating
.through the critical range. Steels of the same grade will
vary in grain size as a result of the type of fine suspension
present. For example, if a steel is deoxidized with silicon
alone it will always tend to be coarse grained and will show a
marked coarsening of grain when overheated. Aluminum killed
steels on the other hand, are almost always fine grained and
will not coarsen until a higher overheating termcrature is
applied than was the case in the original coarse-grained steel.”

From the above statements it is evident that the final
quality of steel is largely determined by the deoxidizer and
the method of deoxidation. Steel manufacturers are now making
use of the knowledge gained through recent experiments and are

"grain size controlled steels” by preper manipulation

producing
of the various deoxidizers during the steel making process,

and at present grain size is a very important Specification.

 

 

 

l
C»!
I

SCOPE AND PURPOSE

In view of the present interest in the grain size of
steel, this study was undertaken for the purpose of investi-
gating medium carbon steels with respect to the relationship
existing between grain size and volume changes on hardening
and tempering; grain size and impact strength; and grain size

and cracking tendency of the heat treated steels.

 

T iI.l‘L

 

..l. ‘.s .

. .tr». ,.

. : 4.4.3.5.. .fl...,sn.[,_m..¢ iré‘
-

.

EQUIPL’IE IT SED

s were used to obtain the dimensions

"S

micrometer calipe
of the test cylinders. The same micrometers were used
throughout the work and all measurements were made at 20°C.

An electric muffle furnace equipped with an automatic
termerature controller was used for all the heat treatment
work. A gas burner was installed in the furnace to give
the preper atmosphere for the prevention of scaling. The
chemical analysis of the gas atmosphere was obtained by use
of an Orsat gas analyzer.

A standard metallurgical microscope was used to deter-
mine the grain size of the steel.

A Brinell hardness tester was used to determine the
hardness on the large cylinders, a Rockwell tester being
used to measure the hardness of the small cylinders and
impact bars.

The impact strength of the various steels was determined
by use of an Izod impact testing machine. This machine has

a range of 120 foot pounds and can be accurately read to

my

foot pound.

A small power hammer was used for the forging work.

 

T TT’IV T I

.T ‘ fidVTTHO..fl T ..I‘ I

PURE-L". E lei-IOSPIITIRLS AI-TD ITS EEl-L‘ICT 017 THE SC:‘.LE TIEICI‘QIESS

Since an important phase of this study involved volume
changes it was imperative that the depth of scaling during
the heat treating operations be reduced to as small a value
as possible. This was accomplished by using a controlled gas
atmOSphere within the furnace by the use of a burner placed
just inside the furnace door.

Local city gas was used in the furnace. The composition
of the gas varies and the following analysis is given as
being tyyical:

002 02 CO CZH4 H2 CH4 N2 H20(vapor)

J

n /’ r-‘I’ r: P? r- " f, (1 I O
4.53 . fl 17.0fi 4.5; 05» 2 p 9p remalnder

A series of tests were performed to determine the
scaling effects in the various atmospheres obtainable with
the furnace set-up. The results of the tests are shown in
Table I. From these results it is evident that there was
always some oxygen in the furnace atmosphere, the most favor-
able conditions showing .Sfl oxygen present.

According to Table I, two furnace atmospheres may have
the same oxygen content and yet show quite different degrees
of seali 3. For example, consider tests A and F. Both had
the same oxygen content, but test A had a higher 002 content
and a lower CO content than test F. 002 is known to be an
active scaling agent‘Z) and probably does play an important
part in the scaling, as is brought out in the tests just

mentioned.

TABLE I

REST: TS OF PRELITIITIAT'ZZ' 'OTC OI? FURNACE ATTIOSPHEIES.

*“ *---.~-o . u. u. -. -.... -- -.--~.—-.--.-.m-.--.~--.—.w—‘n “-H ~o.—.—-O-- -—-—‘ —

Edfiace Gas
Test Analysis Ratio of Temp.
002 02 co 002 e co 0

v— = ’-".“.'

 

0

Remarks

~—' g... -'~

 

 

 

 

A 7.4 .4 8.8 .842 595 Woticeable scaling
B 6.2 1.5 6.9 .90 650 Uoticeable scaling
C 6.8 .4 12.5 .548 595 Very slight scaling
D 5.6 .6 15.2 .424 815 Very slight scaling
E 4.9 .5 16.0 .506 595 Exceedingly thin
scale
*r 4.9 .4 12.6 .333 595 Average depth of
scale calcula ed
to be l.5x10“0 in.
**G 4.1 .4 17.5 .257 850 Avg. depth of scale
calculated to be
2.6 x 10’0 inch.

 

 

* Tenpering Temperature

*rQuenching Temperature

 

 

,L figwfql Chiba-J..- TI‘

A.

From the test data it appears that the ratio of 002 to
CO is an important factor. Referring to Table I, it is seen
that if this ratio is about .8 or higher the steel invari-

/

ably shows appreciable sealing. However, if the ratio is
around .5 or lower it is possible to heat the steel with

but slight scaling. A C02 to CO ratio of about .5 was found
to give very satisfactory results, provided the oxygen was
kept as low as possible with the present equipment.

Tests were performed to determine the thiekress of the
scale under the different heat treatments used. In the test
at the tempering heat, (Test F - Table I), 12 bright cylin-
drical sanples were used, 6 of which had been coated with
quenching oil. The samples were weighed on an analytical
balance before heat treatment. They were then given the
tempering treatment, 59500 for 1% hours, in the furnace
atmOSphere, and after furnace cooling, were again weighed.

The increase in the weight of the samples was due to

oxygen combining with Fe to form an iron oxide. Under the

conditions of this test it appears that the particular oxide
f v, 3A 1" (V l 0"- (5) .9 H. 1' Q .(‘j J. r. r- r~e~
olmCa Mao argcly FeO . KnowinQ tae weight of oxygen

going to form FeO, the weight of FeO formed was calculated.
The specific gravity-of FeO is reported in the International
Critical Tables to be 5.99 and since the dimensions of the
cylinders were known, the thickness of the FeO film was
calculated, assuming uniform distribution over the surface
of the cylinders.

The preportion of Fe in the FeO was determined and the

 

’{I

Hum

.. ~ v. flunk 5A .hllp.‘ ‘

thickness of the Fe film entering into the FeO was calculated.

‘0
The difference between the thickness of the FeO film and the

fl
1'

Fe film represents tie increase in dimensions due to scaling

.1.

during heat treatment. The scale thickness produced in ttst

\D

F was found to be 1.5 x 10'5 inch for the cylinders coated
with oil.

It was found that the presence of an oil film on the
samples reduced the thickness of the scale to about half of
that on the samples which were not coated with oil. As a
result of this observation, all of the steel used in this
study was dipped in oil prior to heat treatnent.

A test similar to the one previously described was

i
I
7

performed to determine the scale thicgness under hardening

conditions, 850°C for one hour, followed by water quench.
(Test G - Table I). The thickness of the scale formed under
these conditions was found to be 2.6 x 10—5 inch, slightly
more than was formed under the tempering treatment.

Since the f nest measurement in this study is of the

i
order of l X 10-4 inch, the results of the two scaling tests
were considered as sufficient proof that the scale thickness
formed under the two heat treatments used was so small as to
be entirely neglected.

With the previous results on scale thickness capable of
being duplicated by preper adjustment of the furnace atmos-
phere, it was concluded that the scale produced would cause no

appreciable error in the volume determinations to be made

later.

 

I dial 441*-

L Giuliani ....nlraI: In?!"

AL ANALYSIS OF DIWTERCIT
UAITUFAC TURERS

STEJLS

TABLE II

C‘ WT‘T"
L.) .l._J 4...!

IEVBSTIGATED

C0

A8 HEP

OWTED BY

 

exp. Heat \teel

io._‘f No._fi Co. _.C Kn. Phos. SUl' 811'
51 - D .4?-.44 - - - -
54 90,185 B .42 .70 .016 .055 .21
56 81,112 B .48 .41 .015 .028 .21
57 55,254 C .44-.48 .77-.80 .014 .024 -
58 51,544 C .49-.55 .75-.76 .012 .022 -
59 29,101 C .55-.57 .79-.81 .014 .024 -
40 89,011 B .48-.51 .72-.74 .010 .025 -
41 59,045 B .45-.48 .71-.75 .050 .052 -
42 89,114 B .46-.50 .74 .014 .021 -
45 29,095 B .47-.50 .68-.70 .014 .028 -
44 62,480 C .42-.46 .64-.36 .011 .021 -
45 2J,189 D .45-.47 .65-.65 .028 .017 -
46 9-9215 A .49-.52 .64-.66 .018 .028 -

- 10 _

GRAIN SIZE SPECIFICATIONS ND PHOTOHICYOGRAPHS

The grain size of the steel was determined according to
standards of the A.S.T.H. Designation E19-55 (carburized at
927°C (17000F) for 8 hours and furnace cooled). The grain
size classifications run from No. l with about 1% grains per
square inch up to E0. 8 with 96 or more grains per square
inch.

According to this standard those steels showing grain
size Nos. 1 to 5 are considered as "coarse grained" and the
steels that show grain size Nos. 5 to 8 are considered as
"fine grained". A steel of U0. 5 grain size may be either
"coarse” or "fine grained", d.pending on whether the other
grains present are mostly coarser or finer than No. 5, e.g.,
a steel showing a grain size of mostly Ho. 5 but with some
No. 6 present would be considered a "fine grained" steel.

A steel which the writer refers to as having a "duplex"
grain size is one in which the grain size classification shows
it to contain both coarse and fine grains. For example, a
No. 2 to No. 6 grain size would indicate a "duplex" grain
structure in the steel.

The discussion brought out later in this paper will
frequently refer to coarse and fine grained steels. A series
of photomicrographs are therefore included here in order that
the reader may have a clear picture of the differences in
grain size and microstructure of typical examples of coarse

and fine grained steels.

 

.....fl.’ «J

 

The steels selected as being representative of the two
classes are No. 57 and No. 42. The chemical analysis of
these two steels is essentially the same. Some of the

physical preperties are shown in Table III.

TABLE III

PROPERTIES OF A CORRSE & A :IIE GRAINED STEE

 

 

Steel Grain Brinell Increase ERIKQEEEG Izod Cracking in
No. Size Hardness Hardened” Tempered Impact Large Impact
Cuenched (5950C) value Cyl. bar
_ :23 130.. fans. _ v
57 612% 565 .0756 .044 94.5 No No
42 .2- 514 .1590 -.0504 69.6 Yes Yes

(severely)

*A bar under the grain siz, number indicates predominate size.

The photomicrographs and the notes appearing with them
are, in general, self-explanatory, but there are a few facts
brought out that merit further discussion. It will be noted
that throughout the series of photomicrbgraphs, steel No. 57
shows a consistently finer structure than does steel No. 42
when both have been treated under identical conditions.

An interesting feature is brought out by comparing
Figs. 5 and 11 and 6 and 12. Figs. 5 and 6 show the hyper-

"as

entectoid zone of steels 42 and 57 respectively when the
received" pieces are carburized and rated for grain size.
Figs. 11 and 12 show the same zone of steels 42 and 57

respectively when carburized after being forged. It will be

noted that both Figs. 5 and 11 show He. 2 grain size and both

,

Figs. 6 and 12 show Ho. 7. Several other steels given the
carburizing treatment were examined under the_nicrosc0pe and
in all cases the grain size was the same before and alter
the steel was forged. This would indicate that the grain
size as determined by the A.S.T.I. method may be regarded as

the ”inherent" grain size of the steel as it is not altered
’

by severe plastic deformation at forging temperatures.

A cohparison of Figs. 1 and 5 shows the coarse grained
steel to be of about the same grain size in "as received"
condition and after being carburized. Similarly Figs. 1 and
7 reveal that the coarse steel exhibits much the same grain
size "as received" and after forging. The same thing is
evident to a lesser degree in Figs. 8 and 2. In forging the
hot work was stepped and the piece allowed to air cool from
about 875°C.

Fig. 15 is included to show what may be an inclusion in
the steel. The photomicrograph was taken as representing the
beginning or "root" of a crack in one of the impact specimens
machined from a forged bar of steel No. 42. Several other
cracked impact bars were examined in the same region and Fig.
15 is typical of the microstructure in that area. Different
polishing methods were resorted to in an attempt to ascertain
whether the slag-like appearance was due to polishing compound
lodged in the crack. The appearance was the same regardless
of the polishing method used. The writer has no positive
proof that the "inclusion" was slag, but it had the same
appearance as the inclusion (ferrous silicate) in Fig. 5 of

a recent paper by Urban and Chipman.(4)

 

at)“. z .1 4. w! s . L . . v 4 . D‘ A .7 . K flI‘EI‘,’-i!.oi.lu.j~l|vu4r:§i‘v ..
4 4. v . a

 

r‘
v_ j

0

q , o
I

_~

k}.

a
k)

l

N,

i190]. "13 3‘

100x

7

.’"‘.": '3 7'1

{A
L 12

‘1; n

1 OF

- —\_’I

-- '1 Y’
steel ‘13

1(3Ct:

x? 1.791

'I

" 1,3

‘1‘? J
-n
'I
..L

.g ,4.

"'q

"‘.

(1" (I , . E,
f U
f1 1 ‘ \
A
V‘.
v a
o 91 -
r o < . ,
. ‘ II A
-.f,
,
- :m' . -, x .1,

 

 

 

 

 

 

v3 d"

0.1“

nnfiOn

gl\>_’ , '

a p '5
p '._) _

l

N

(

..
li"

r
(

U

-/

a

d..-
)JJL

10

t

". "-3, I"
C .' fi —

treat

«p
-I-

00»

5.4

I

n10, 1:13;-

1‘

Q Jv
L.)

~-’. 5

"I.

O
H
, I“ . k . f» I
~, ‘.- V I . I
LJ V‘ - \ ' .1 \
--rv‘
‘4 \
I; ._ .
r l V " I 1
O I
‘0 I ‘ V
(J ‘
t. <
".4
a. Q j
'I
. 011‘. r v“ ;p_
'Q r) f- ‘>. f‘ .1
'fi 1 " I1. I .
'J a, ", . ‘.
fi'lr)!’
J. P . J .
3 :1.) .__“:

I‘ ‘
r

r\

.'~ ,k

L

'.
LI

_ -L '
i". It‘wo -
V17 .-.

3 J 9f! -

 

 

 

 

a
\

.- . ‘4
I 'Q “.~ f."

 

 

..,.~.. 2 .1"
L Hut-3 f-‘f'.3l_L

1

(“”111

\__t

1-
.L J. I--

"‘

\_0

1575

L. .
LI

trea

1_ o o
v.1].‘3‘ss‘, "w-fir-
'-- ‘,‘_A- _._l-.__.-

C

q

(‘1
8.. Id.

.‘
I

.1

. c . .10

“(1‘
on

a‘b

100~

J.

 

J L.
fix .Qs
.u a...
o.

r n).
. .

.. r.»
I;

C

'. a.
f r.
o ...
..q ..
0'.

\z

. Fl
“a

'l)
I. A
. _
D
I I
n.
O .
.\ If
.(\

a

b I".
‘-

r
1
,-

“ff

..(

air cooled fro»

10o—

- c 5

00. u?» ~ . r '3 s {31‘11('t':.“.?‘£‘-

for‘el

79.011“:

.... o . A
3')
‘41. ‘V
v br- ~- _
-.O. sued. :. 1‘» 0.“

Air cooled 140w

Coupare sURWCiwi

#p“|fl

K 1 9 }‘_I

f

I
~

 

a
...4‘5

..

b L
\'.
(
r:
«w .
fl
.'*
{\v
3.
IL-
. I
(I1 H .
f .7
..H a;
a v I.
u”\ .
. Ow.
Or
7
..
OI”.
_
.

H
\—

J ‘3‘!
‘C’Li'g: “

, .

it
I-..

.r.

“I

\l:

1

w‘.

‘ff"(‘\

1..

"r fr?!

‘»'0
‘1

 

 

 

 

.7-

4-
IL- 0 ‘J

C 03 r3 ar 3

O

0".
/

’—/

-.L .

" l'

I v
(7
‘1'

.

‘ - _“
-‘ L ‘0""\
n.‘ '1 _ !.\,

~fi.

J6 -. . c
1. or 'ml
...
s ." v.‘ VJ
w ‘ "
r ‘. .‘lj
..-' _' ‘, -"o_
.‘ ‘1 _:-
" a
(-3 "."{,t
_ . \J_ L
“O"
‘0 ]( I I
' _/‘.J .
Li-.. 1.,11L,
.° '4 ‘ ’7 o M
Ll-i: ;‘ l

m n

\

'
l
i A

(
' \.

..L ,3
\l

__ ,. r‘
5"rflt f‘k‘x.‘

L47-

‘:_J

h

v, .1

3
J

.i

'T’r.’
o) 1'

TN .'-\. fl"
‘ —’ ‘~-' on L

m

.fi.

r. r
l . ,
.L ,

7! .‘
o O

" ‘ I ‘ Pi
‘ ...
v '-

b _

W}

- -
rt’ "'5 -.
..L 4 L1 L,-
. _-_ I,
. (W
_ . ‘ . .. ..
o o r
r: - (.r ,
\ . L; -
A
:1. r.
. r ” 9
.0 _ . u_'
'1 o' - r'
u. t . '1
....
,.
.'. ..L
l ’I O .1~'—‘
~’\) (.1 . x.‘
- .n
t (W: )1 ,4
" ('5 I r.
._ \ L.-.L.
(jft'f' (\f,’ ‘0.)

 

 

5.

K

- I
p ‘ 4‘

I
flew

.“

«I ’ *( f . r, ’-
.' -?"K~Y\‘" ‘5 :ég

 

.'
- I i
. -

-..
. ‘ fiv—e
— —L, 0 .Lu

1 .~. - -. ..
"10 ‘-)“LIjs-\‘J.‘~Y_’\ ‘I ’ ”'1 'f
. I ..k (J (3“)'I (‘mfi
eon ‘— 3 .
Ca. MC :AL ‘9'!“— '; , '. (-3.9. (‘V‘ .
_ ‘ a: a) )PCHL -1‘1
7‘"
I )0}:
Q. r J,
~‘Ot(3 U"_’.? a“ r; ‘_1 i7..-_) ...
”“e t ” “PP9arance

of‘ :1 ~
.- J ‘ {Tani}? area

43.
i

14

O
’1" ‘ - - "\ ~
‘-1 (JIM; 0 .llf‘t ln “—1-
) J " () (‘1'1 ' ‘
_ / ,u__ 1‘W1L.)pg
1n. .
x.’ "1 " “'-
M.u'3 dra'cwn‘ “no
- - H, A ,‘1

- o l'
. -3 r 3- .. , rr ._ -. _l_ ... .- _ ..TH
.43: . ‘ 3: 7'1) J‘fi ,‘ 1' r4!“

IJ

(~0(‘\|"'m

. \J utJC')

' F
".-'1‘
J‘L'v’l‘.

 

 

 

TESTS ON THE LARGE CYLIHDERS

Procedure:

 

Bars of medium carbon steel from the heats shown in
Table II were obtained from various steel mills in order to
introduce the variables, due to different heats, met with in
industrial practice. Samples were cut from each of these
bars to be used for metallographic specimens and to be sub-
jected to the various heat treatments used during this study.
Two four-inch lengths were cut from each bar and the
remainder was forged to about a 2 inch round bar. The four-
inch lengths were normalized at 900°C for 1% hours after
which the ends were machined parallel, with a very fine
finish cut, and the diameters were machined to within 1/10000
of an inch of being perfectly round. These pieces will
hereinafter be referred to as the large cylinders.

They were accurately measured by four independent read-
ings taken of both length and diameter at approx'mately 90O
intervals. Two such sets of readings were taken on each
cylinder at different times prior to any heat treatment.
After the observer acquired the technique, the dimensions
could easily be read within l/10000 inch of a previous
reading.

The Brinell hardness was taken of each cylinder. The
large cylinders were heated to 850°C and held for 1% hours
in the controlled atmosphere electric furnace. After being

prOperly heated, the cylinders were quenched at the rate of

 

|.|
.H‘

40..

one a minute, in water kept at nearly constant temperature
(16°C) by a steady stream of fresh water.

The dimensions of the quenched cylinders were obtained
as soon as the steel had reached room temperature, about 20°C,
at which temperature all of the volume measurements were made.
The dimensions were measured to the degree of accuracy
previously described. The Brinell hardness was read on each
cylinder. Within 1 hour after quenching each cylinder had
been measured and heated to about 260°C to relieve the
stresses set up in quenching.

The cylinders were tempered at 595°C for 1% hours in a
controlled atmosphere electric furnace and allowed to furnace
cool in this atmOSphere to about 150°C. The dimensions of
the tempered cylinders were obtained as before, two complete
sets of readings having been taken. The Brinell hardness

was read on each cylinder.

Discussion of Results:

 

The volume changes referred to in this work are eXpressed
in per cent of the original volume, and are all positive
unless otherwise indicated.

Table IV shows the results of the tests on the la ge
cylinders when hardened and tempered, and the results of the
impact tests on the same steels forged and machined to impact

specimens and subjected to the sane heat treatment as the

large cylinders received.

 

..- 0 III 316‘14 .um DRIVE. . - v

LIL.

TAB 3 IV

RESULES 0T TESTS 0] LARGJ CYLINDERS ND IXPACT BARS

 

 

 

Steel Grain Brinell Increase in Volume Izod Cracking_in

No. 813 Hardness Hardefied' Tempered Impact Large ImpaEt
Quenched (595°C) Value Cyl. Bar

__. _ 53 .11 % ft.1bs.
51 2-5 555 .181 .058 71.6 Yes 'Yes
42 '2-4 514 .150 -.0504 69.6 Yes Yes
45 1-5 555 .150 .0518 58.9 Yes Yes
59 2-5 555 .175 .154 75.5 Yes Yes
41 5-5 555 .181 .077 77.1 Yes Yes
46 £15 514 .174 .058 55.5 Yes Yes
40 Efé 413 .0718 .055 75.1 Ves Yes
58 4f§~6 444 .104 .072 75.0 Yes yes
54 4-gr6 477 .152 .067 84.3 To 33
44 5-6 565 .004 .048 85.4 ho NO
45 512 555 .0658 .055 81.5 Ho Ho
56 .2‘7 565 .081 .020 84.9 No No
57 641 err .0756 .044 94.5 Ho No

 

All steels normalizaiat 900°C for 1% hours.
All steels tempered to a hardness of 212 to 229 Brinell.

Original hardness on all steels about 179 Brinell.

Cylinders 2.2 to 2.4 inches in diameter and about 4 inches lone.

" U

 

 

.....PCAV
p . a

.fill
a...

.0

V5! ... . ..

I .3:

I

.1I‘ .

.1; I! I. \I Lvn.

I'lIo‘. .

.0

.ll «
....

Classifying the steels with respect to grain size,
steels Nos. 51, 42, 45, 59, 41 and 46 are considered as
coarse grained. Steels Nos. 40, 58 and 54 all exhiiit a
duplex grain size, but of the th ee mentioned ho. 40 seems to
have a more truly duplex structure than do the other two.

No. 40 had both Ho. 2 and No. 6 grains present, while Nos. 58

and 54 both show mostly No. 5 with some No. 4 and No. 6

present in each. Steels hos. 58 and 54 might be considered

as being on the borderline between corase ahi fine grained

steels. Steels hos. 44, 45, 56 and 57 are all fine grained

U

Considering the relationship existing between grain size

and increase in volume on hardening, it is seen that the

coarser grain size is associated with the greater increase in

L

volume. The per cent expansion is of the same order of

L

..4.

magnitude for all of the coaese grained steels. The f’ne

!

grained steels show a lower percentage increase in volume on
hardening.. Steels having a duplex grain size, as steel I0.

40 with grain size No. 2 to to. 6, may show the eXpansion

1‘

characteristics 01 either a coarse or fine grained steel.
Those steels, Nos. 54 and 58, which lie midway between coarse
and fine grain classification show a smaller increase in
volume than do the trul* c arse grained steels.

The preceding observations are clearly indicative of the
greater tendency of the coars- grained steels to eXpand on

hardening the degree of eXpansion increasing with the increase

in grain size, excepting, of course, those steels diich shov

e characteil.t°cs and.which may act as either

N

duplex grain si
coarse or fine rained steels in some of their preperties.

The tendency of coarse grained steels to mlow a hiCh de cgiee

of eXpansion indicates that they are much more subject to
warpage and distortion upon hardening than are the fine
grained steels.

Likewise, there is a definite relation hip between grain
size and surface Brinell hardness of the steel in the hardened
condition. From Table IV it is evident that a coarse grained
steel is associated with the hi her Brinell hardnesses, (514
to 555) and the fine graized steel with Brinell hardness of
565. Here again the d.uplex grain steels e: {hibit hardne
values betw en those of the coarse and the fine grain steels.

Davenport and Bain(5) and others have shown that coarse
grained steels possess greater hardenability than steels hav-
ing a fine grain. The hardening power s directly measured
by the Brinell reading and also by the eXpansion. Bain and
V.'aring(6) make the following statements concern win the densi-
ties of the various constituents in steel.

"Carbides are the most dense, austenite somewhat less
dense, and martensite the most voluminous. Ordinary ferrite
lies between austenite and martensite as does also the mixture
of ferrite and carbide found in annealed steels."

Steels which have the grea ater ha rdenability would show a
greater preportion of martensite when hardened and since
martensite is the most voluminous constituent produced in

quenching it is logical that the coarse grained steels, which

show the higher Brinell hardnesses, should also show a greater
degree of eXpansion than the fine grained steels.

The per cent volume change in the tempered condition did
not seem to have any relationship to the grain size. The
greatest inconsistencies were found in the coarse grained
steels in the tempered condition, Ho. 42 showing a decrease
in volume of .OSOdfi over the volume in the normalized condi-
tion. It may also be observed that fine grained steels show
greater consistency in volume changes than do coarse grained
steels, the variation being from .029% to .055%

Another interesting relationship may be pointed out
between grain size and the tendency toward cracking when the
steels are hardened and tempered. The various steels were all
heated at 850°C and quenched into water at about 16°C as
previously described. Table IV shows a definite relationship
between the grain size and cracking tendency. All of the large
coarse grained cylinders showed varying degrees of cracking
when quenched. Steel Ho. 40 with a duplex grain showed decided
cracking and in this respect is typical of a coarse grained
steel whereas its volume change and Brinell hardness were more
like that of a fine grained steel. Steels 58 and 54, both
coming under the duplex classification, again exhibited the
preperties of steels on the "borderline”. Ho. 58 cracked
decidedly while Ho. 54 showed no visible cracking whatever.
This emphasizes the point that steels showing a duplex grain
size are inconsistent in their preperties and may act either

as typically coarse or typically fine grained steels.

I
[O
C)

I

None of the fine grained steels showed any tendency
toward cracking. This observation suggests that coarse
grained steels are either more sensitive to heat treatment
or that they have inherent characteristics which give rise
to quenching cracks.

Herty(l) claims that coarse grained steels have a lower
transformation termerature and a relatively lower transform-
ation rate than do fine grained steels. It is therefore
necessary that they be treated in a limited temperature
range. It may be that the hardening temperature used in
this study was not suited to the coarse grained steels.

'It is worthy of note to state that those steels which
cracked most severely were Nos. 51 and 42 in which the
smallest grain size was No. 4. The cracks in the large
cylinders were all circumferential and of about the same
intensity with the two exceptions just mentioned. A typical
crack is shown in Fig. 14, page 19 of this paper. The
tendency of the coarse grained steels toward cracking was
not confined to the large cylinders, but was also observed
in the impact specimens. It is interesting to note that
‘gll of those steels which cracked in the large cylinders also
cracked in the impact bars; and those steels which did.22§
crack as cylinders showed no signs of cracking as impact
Specimens. The agreement in this reSpect was perfect. The
impact bars and cylinders were all quenched from the same
temperature. This seems to be positive proof that the
tendency to crack on quenching is inherently in the steel

Q

for it is not altered oy the forging Operation.

§pmmary of Tests on Large Cylinders:

 

The preceding observations on the large cylinders may
now'be smmnudzed.as follows:

1. When steels of practically the same chemiea
analysis are hardened the coarse grained steels may be
expected to show large volume changes and high surface
Brinell hardness and to exhibit a marked tendency toward
cracking or distortion. The fine grained steels when hard-
ened may be ex acted to show small volume increase, lower
surface hardness, and no cracking or distortion.

2. The grain size of the steel has little effect on
the volume change in the tempered condition. Generally, the
fine grained steels show greater consistency of volume
change in the tempered condition than do coarse grained
steels.

5. A steel having a duplex grain size is inconsistent
in its behavior, and may exh bit the characteristics of
either a coarse or a fine grained steel. Some of the preper-
ties may be consistent with those of fine grained steels
while other properties are characteristic of coa se grained
steels. Steels having a grain size midway between the
coarse and fine classification are quite consistent in their
eXpansions and Brinell hardnesses, but may differ markedly

with respect to cracking tendency and impact value.

 

TESTS ON THE IKPACT SPECIKEH

Procedure:

 

The & inch forged bars were normalized at 900°C for one
hour and two round Izod impact specimens were machined from
each. The impact Specimens were heated at 850°C for one
hour, water quenched and immediately tempered at 5950C for
one hour. Rockwell hardness readings were taken in both
quenched and tempered condition. The impact values, six in

all, were obtained on an Izod impact tester.

Discussion of Results:

 

The Izod impact values are included in Table IV. It is
doubtful if some of these values can be interpreted very well
since a number of the specimens were cracked longitudinally,
and the impact values determined on the cracked sections are
apt to be in error. In steel Ho. 58 one of the impact
Spec mens was cracked and the other was not, yet the impact
values were the same on all of the six breaks obtained as is

shown in Table V.

1

ES*LTS 0T IHPACT TFSTS OH STEEL N0. 58 - TEHPERED 5950C.
Break Ho. 1 2 5 4 5 6 Avg.
Impact value ft.lbs. 76 77 77 76.5 76.5 75.5 76.1
Breaks 1, 2 and 5 were on uncracked bar, while breaks 4, 5

and 6 were obtained on longitudinally cracked bar.

..l.‘ O 1'.’

 

to

This may indicate that longitudinal racks do not
materially affect the impact values, but it is better to
interpret the values on the cracked pieces as being subject
to error. The impact values of such specimens varied from
55 to 77 foot pounds, all of the steels being coarse grained.

The impact values for the fine grained steels are
reliable since no cracks were revealed in any specimen.

Steel No. 57 with he. 7 grain size, showed the highest impact
value, 94.5 foot pounds. The other fine grained steels, with
mostly No. 6 grain, showed impact values of from 81.5 to 86.4
feet pounds. This indicates the higher impact value for the
Smaller grain though the data is rather meager.

Steel No. 40 with a duplex grain structure, showed
cracks in the impact specimens and the impact values were
consistent with those of the coarse grained steels. Steels
hos. 54 and 58 again showed the peculiarities of a duplex
structure. Ho. 54, which did not crack, gave an impact value
of 84.5 foot pounds, while Ho. 58 showed consistent impact
values of 76 foot pounds even though one Specimen cracked and

the other did not.

Summary of Tests on;the Impact Specimens:

 

The results of the impact tests are summarized as follows:

1. Higher impact values are associated with the fine
grained steels.

2. Cracking in the im act specimens is associated with

coarse grain size.

5. Steels showing a duplex grain size may have impact

values typical of either coarse

O

r fine grained steels.
4. In steel No. 58, at least, the impact values are

consistent even though one specimen cracked longitudinally.

m ("m

lens 01: T1113 SInLL cmnmms

1

Procedure:

 

The small test cylinders were machined from the forged
stock remaining after.the impact Specimens were made. The
small cylinders varied from .625 to .750 inch in diameter,
and from 1.70 to 2.20 inches in length. They were measured
in the same manner and heat treated under the same conditions
as were the large cylinders. All hardness determinations
were made with a Rockwell tester in order that the sample

should not be made unfit for subsequent measurement.

Discussion of esul s:

 

Table VI shows the results of tests on the small cylinder.
A survey of this table reveals that the increase in volume on
hardening is much greater in the small cylinders than it was

T

in the larger samples. For example, steel Lo. 59 showed an

increase of 1.545 in the small cylinders and.l75% in the
large cylinders. Similarly steel ho. 41 expanded 1.1265 in
the small cylinders and .181% in the larger pieces.

Both of the steels just mentioned are coarse grained and
as such would possess good hardening preperties. The small

cylinders would harden more completely because of the reduced

cross section, and a greater prOportion of martensite would

’-

.9

Ho

. .L. .
mar t0 [1 S

(D

be produced than in the larger sections. Sine
is the most voluminous constituent(6) in the structure it is
reasonable to eXpect a higher percentage of volume increase

in the smaller cylinders. Genera‘lv, the fine grained steels

showed the lower volume increase, but several inconsistencies

are apparent.

TABLE VI

RESL.TS OF TESTS Oh SHALL CYLINDERS

 

 

 

 

 

 

Test Grain Rockwell Diam. Volume Charges, per cent
Steel Size Hardness inches -T§fiifi§i§i‘ Tempered”
1'10 - “wheels-Casi“... .. 1-.--
*N51 ‘2-5 59 .750 .852 .595
N42 {2-4 59 .687 .800 .561
N52 5 54-57 .688 .555 .581
52 5 55 .662 .700 .425
N45 1-5 58 .750 .608 .162
N59 2-5 61 .750 1.540 .504
N41 [5:5 58 .750 1.126 .527
N40 [216 55-60 .687 .016 .519
N58 475-6 49—58 .688 .996 .505
N54 412-6 55-57 .750 .644 . 55
T45 516 60 .750 .906 .507
N56 6-7 57-60 .625 .667 .236
56 6-7 57-61 .685 .651 .552

4A _‘ A O ‘«_W.

*”N” indicates samples normalized at 900°C for one hour.

Rockwell Hardness as tempered - C22 to 28.

The results in Table VI on volume increase in the hard-
ened condition is not of much value. The diameters of the
cylinders varied from .625 to .750 inch, depending on the
diameter of the forged piece. It is common knowledge among
steel treaters that a sample in this diameter range is on
the borderline between complete and partial hardening throu3h
the section. Such being the case, it is not advisable to
rely upon the volume chan3es .iven in Taole VI, because some.
of tile sections may have hardened CONlpthGl y throu3h the
section, while others were only partially hardened. The
variations in diameter, and this particular diameter intro-
duce what mi3ht be termed a "dimensional problem", vfhich has
not oeen inves i3ate1 in this work.

There was no apparent relationship existing between
grain size and Rockwell "C" hardness, probably because all
of the steels hardened thoroughly at the surface regardlw

of the grain size Likewise, there is no relationship between

grain size and volume change in the temp red steels.
Both cylinders of steel No. 42 showed H1i3“ cracking
when water quenched from 850°C, being the only steel to show

crackin3 under three conditions.

Sunmarv of Results on
.——.—.!L._._.. A .r

*-—.— *H-u-

 

1. The per cent volume increase on hardening is much
greater in the small cylinders than in the larger sections.

2. The lar 3est ex pansions, (1.545 and 1.1265) occur in
coarse grained steels.

5. Grain size bears no relationship to Rockwell "C"

Q

hardness in the Lardened cylinders of the diameters used.
4. The diameter of cylindzrs used in this study was
such as to introduce a ”dimensional problem” and leads to
inconsistencies in the results obtained.
5. There is apparently no relationship between the

grain size and the volume changes in hardened or tempered

conditions (in the chosen diameter ran3e).

CONCLUS IONS

1. Fine grained steels may be expected to show high
impact values, low volume changes on quenching (which are
indicative of a low degree of warping or distortion) and no
cracking on quenching and tempering.

2. Coarse grained steels may be GXpOCted to show lower
impact values, a decided cracking tendency on quenching and
tempering, high surface hardness and large Xpansion on
hardenins indicating that they are subject to warpage or
distortion.

5. Duplex grain steels are inconsistent in their prOp-
erties and may exhibit the characteristics of either coarse
or fine grained steels.

4. The Specifications for a particular part, based on
the results of this study may be stated as follows: If the
part were required to have consistently good toughness, as
is indicated by impact value, a low degree of eXpansion or
distortion, no cracking tendency when quenched, and/or good

5

forging prOperties, a steel of grain size greater than No.
ioult be Specified.
5. All steels showing a duplex structure or grain size
below No. 5 should be rejected.
6. If high surface hardness is most desired, without
regard for distortion or cracking, a steel of a larger grain
than No. 5 should be specified. Duplex structures should

not be accepted.

4.

BIBLIOGRAPHY

Deoxidation of Steel - C. H. Herty, Jr.

Transactions A.S.H., Vol. XXIII, Narch, 1955.

The Effect of Furnace Atmospheres on Steel - R. G. Guthrie

Transactions A.S.S.T., Vol. XV, January, 1929.

U. S. Bureau of Nines Bulletin #296, pp. 7, 55, 81-2, 95-6
Iron Oxide Reducation unlibria
A critique from the standpoint of the phase rule

and thermodynamics.

Inclusions in Steel - Urban & CLipnan

Transactions A.S.N., Vol. XXIII, March, 1955.

General Relations Between Grain Size & Hardenability &
he Normality of Steels - E. S. Davenport & E. C.
Bain

Transactions A.S.N., Vol. XXII, December, 1954.

Austenite Decomposition & Length Changes in Steel -
E. C. Bain and W. S. N. Waring

Transactions A.S.M., Vol. XV, Januawv 1929.

"d’

 

 

m1|milHamlinIgutguljlflluliltltmlilillnu1