na
A
~
ROSH
ai
iM
OF
a
CARBON
ae
NS OF TH
SU sty
eS COA NTS Clie atl) Oh ae
OF
1919
=)
be
<<
=
1
T
3
E DE
May
PE
BY
ar:
Nyaa 8
amas 3 red a ‘ 7
ek teins Prerdeinata crium, but if the variation of one of the conditions
causes a change, then the system is said to be in unstable equil-
ibrium. Often a system will fail to underso a transformation when
tne change is due, in this case it is said to be in a metasteble
condition of equilibdriun.
Third-
By degrees of freedom, we mean the number of tne tnree
variables, temnerature,pressure and concentration, which ay be
cnanged witnout disturbing the equilibriun. If the systena has
no degrees of freedom it is said to be "umvariant", and if it
has one desree of freedom it is said to be Bunivariant", while if
it has two desrecs of freedom it is said to be"bivariant?
Fourth
By phase is meant the homogenous physically disting-
uisheble and :nechanically separable constituents of a systen.
Fifth-
By components, we mean thoseconstituents, the concentre-
tion of which can undergo indevendent variation in the different
phases. Howe gives a clearer definition,"those free elements and
compounds which in the nature of the case are undecomposable under
the conditions, and thus play the part of elements. It is some-
times difficult to determine the number of components in a system
» unless the following rule is aoplied.
As the components of a system, there are to be taken the smallest
nuaber of independently variable constituents, by ineans of wnich
the composition of each phase participating in the sutee of equil-
ibriun can be expressed in the form of a chemical ecuation, In °
the caseof alloys, such difficulty does not arise, as the co:mon-=
ents are always the constituent metals, or some simple compound.
Wnen cealing with alloys, the influence of pressure is gener-
ally neglected, because of the fact that they are so feebly volat-
tle, and are thus subject to atmospheric pressure. If the press-
ure i186 eliminated, we necessarily reduce by one the degre:. 0:
freedom, so that tne rule becones-
PeOt+t-P
This limits tne number of vossible pheses, for an alloy in equil-
ibrium, made uo of X metals cannot have more than X pheses. If it
uic “41 pheses, the rule becomes-
Fe X¥+T1- (X¥-1) = 0
and tne alloy would not be in stable eyuilibrium, on the other ham
if the alloy had X-1 vhases' then-
F2e=X+1- (X-1) = 2
and the alloy would have the maximum number of degrees of freedom.
PHASE RULE APPLIED TO ALLOYS
In general tne inetallic elements way exist in either a liquid
or solid state, and the study of what occurs as they pass from one
state to another, brings in the application of the phase rule. We
know that if a pure metal in the liquid state is gradually cooled
to a solid state, at the temperature at wnich the change takes
plece there is an evolution of neat. If we are ::easuring the dropa
in temperature by :neans of a suitable pyrometer and plotting a
curve, usin;; time and temperature as the coordinates, the curve
would look lixe that +iven in Fiyure 1, Plate 1.
The portion BC of the curve shows where the change of stcte occup
there being an evolution of heat. As there is only one neteal,
there can be only one co-:zponent, but three phases are nossible,
(1) liquid metal, (2) solid metal, (3) alotronic forms of the se
solid metal. In Figure 1, Plate 1, above the point T, we have
all licuid, therefor one phcse and the rule becomes—
Fe2-t%1#s1
That is the system is univariant, therefore we can alter its tem
teuperature within wide limits without disturbing ites equilib-
rium, At the teuperature T, we find crystals foraing in the
liquid, thus we nave two phases present and the rule reads~
Fe2-220
and the system is unvariant. If we either raise or lower tne
te:verature, we will cause the disanpnearance of one phese of the
other. Pelow T, the system is entirely solid, that is the number
of phases is again one, and we have-
Feo2-1# 1
tne systen being univariant.
Now let us turn to binary alloys. That is, those alloys
which are made un of two metals. In this case we have two
components and the phase rule reads-
Fe2+t1-+P orjZ-P
First let us exanine the cooling curves of a binary elloy.
Here we have three cases-
(1) = Alloys soluble in each other when both
hot and cold.
!
} :
'
tt?
tit
. . it
i geek tt
; ear
ae
: aces it
Sea {
elie Cone
‘ oe aes
Vhs
:
‘
lene
re
as
i
2 + r
; ry
ae ee
le NB ea aga
. |
ca wee ja aae .
: oa :
tt es
t %
Qiae
{ Leechs =
Ww Td Migcedtemess ¢
tit 4
' at
: nk
: ii ;
ies : ra
ly hesew ;
fer + :
iy bp ts wh
uly thretete tt
‘ : i
- ‘
pe pedy
: pbia pt *
deere gE
-++— 1 +
; ees}
shy lt deae bee
inde beet
ane
: :
‘ :
me
; i
abs ‘
sees ees
een Rien
eae ee ra
+
'
‘
meee
theta pe
+
,
. +
babes
doe
ue
i
ai
rebge ‘
vie >, ra
te beeeye
staaee
eege bee
ee
ad
a4
' ty
ee
+ ite
‘
4
=i
sit
Eat
'
ee ar
EEX.
eg peirr
, if
hia ag
OND
treet
tae
a4 ¢
rept
ide
te
\
peed
ar ‘
thre
titees
oe
ee
1
Vie ees
preter te
on
epee ’
ye
t
ere °
Sybase
e
s
te
2
agree
a eee
‘
- -
~
‘
‘
é
wt
Loe
eke
10
(2) - Alloys soluble in each other when hot, but
insoluble in each other when solid.
(3) - Alloys soluble in each other when hot, and
partially soluble in each other when cold.
Let us breéfly exauine case 1. A typicel cooling curve of
such an alloy is shown in Figure 2. Plate 1. As the alloy cools
from A to B, we have a steady drop in the curve, When the point B
is reached at a temperature T, instead of getting an evolution of
heat enough to hold the temperature constant, there is a slowing up
only, the alloy graduelly becoming solid, but not entirley so till
C is reached at a temperature T'. From C to D we nave a normal
cooling curve, From this it is shown that between C and B, there
is a gradual: precipitation of crystals and not € succen one as in
the case of a pure metal. Fijure 7,Plete 1, we have the cooling .
curves of a series of alloys of the same two metals, varying from
100 4 of A to 100 % of B. At a, the curve is similar to the one in
1, a',a",a'", are similar to thht in Figure 2, and at a"" the
same as Figure 1. If we connect b,b',b",b'",b"", and b,c',c®,c"",
b"", we will get a curve that can be used for any alloy of the
metals A and B. However in this case we are attempting to pot
both time and composition along the same line. Let us eliminate
the thme factor, ploting b's and c's directly under each other.
Our curve then is show by Figure 4, Plate 2. From this curve we
can study the mechanism of solidification. Above acb everything
is liquid and below adb ewerything is solid, while within the area
acbd we have the transformation of one state to the other. If
everything above acb is liquid we have but one phase and
FP=xe35Z-12 2
V1
Which sisaply mesns that both the concentration cad the tenanerature
of the liquid nay be changed within wide limits without disturbing
the equilibrium of the systea, or which in this case means solid-
ificetion. The sane rule apvlies below adb, excent that here both
teanerature and concentration may be cnanged without causing the
alloy to melt. Witnin the area acbd, we have two phases present,
Liquid and solid and tne system ie univariant, that is only the temper-
‘ature or the concentration may be varied arbitrarily. In Figure
4, plate 2, let us consider an alloy of the composition R. As the alloy
cools to tne poin y! crystals begin to precipitatite out, the comp-
osition of which at the teuperature y", is shown by the point x’.
If the point x' does not show the composition of the crystcl, then
sone other point on the line y'x' must show it. Let us suppose
that the composition lies somewhere to the left of x', then it
lies within the area acbd and the crystal itself must be a nart
Liquid which is impossible, while if it lies to the right of x'
it must have become solid at a higher temperature than y'.
Therefore it must have a composition shom by the voint x'. As
the te:aperature slowly drops the composition of the crystals
chenge froa x', to x*", while the composition of the liquid bath
shifts from y’ to y'", and the last drop to solidify has the
composition y'",. If the alloy is zradually cooled below adb,
difusion will gradually takes place in the solid solution, 80 that
all the crystals will have the composition R', In other words
within the area ac bd, if we vary the temperature, the concentration
will vary accordingly, or if the concentration be varied the temp-
erature will change.
Case 2-
If we have a molten alloy falling under this case and
THESIS or HAP
PLATE @.
TEMPERATURE
FIGURE 4
TEMPERATURE
COMPOS/TIOLY
EO FIaViE F
ros ZB 5
FUSIGILITY CURVE: OF A MARY
i ALLOY WHaSE- COMPONENTS ‘HARE
SOLUBLE Ii) FACH OTHE L
TIME COOLING CURVE OF A PUINARY
ALLOY WHOSE
COMPONENT S| HAE
INSOLUBLE | 1 EACH OTHER WHEN
SOLID.
<4 soe R
Se LUD fFi- 24/-1=2 *
i oe
> RQ Hee 4 wae 97
Y ‘ Hp Ly
ie ~ NS FH RW oO, 3 wf
ty * iz e i te Mee 7 F22t-2 2+ 4
; Ast
* re fs ganas ne eee / A te
Ped R SOLID eat
f ft fe |
: Gane Sb bY 3
G 4.
“hk ff. FU OSLT IO, ide ceca eae WHOSE: oor NTS Ae
PARYIALY SOLUBLE V4 ERCH eee SHOW TL
PUBS,
THE RF PPLICATION OF (THE
4
frit $4
Mls 4h
. -*
"4 cs . yt
. Ly ‘ a. .7 --A SL ,
" . - 8 o . at
. . ” * aoe .
-, a a re an -
Ls : . oy . Vo. wo ‘
eek er ee me ee ee fe tee ee em me ee ae oo
. it _ ie wee = ere emwes Be = 3-w- eo - em ee Tg hte teh ere oe
7 7.
j ’ ate
‘ “~~ . ‘ »
j :
' 1 om - -
i i “- 7 = we . ‘ vo, .
. : aX or -
? { ‘ ‘, ees Sem, - °
f i ‘ ate “y oe
} { . Te. - - ! ¢ Me ~ ° 7
7 ! : moe has . _ \ ' - “os wee, et
. “ 2 . ‘, ta -
} | : hs
.
. ‘
mo ! fo |
: : ' " : .
4 } . ! ‘
. ) , ' ‘
he ' .
‘ ’ \ e '
. i .
' ’ . :
. : . i ‘ ’ ' “
so } - - ee ‘
» . ’ . - . . :
. , . ‘
: a | “wh :
> 17 !
> * '
, 1
. . |
' - . a =e 2 6 + -@! ; Rovere ee welt wae le |
, ' - ‘
- , ' a ‘ -
J , : 7 . ’ N \ .
4 1 , * $
, . . + ‘ .
. ‘ _ 7
' ‘ - . 5 .
‘ ‘ : ‘ .
. woe ee oe - ee ’ ‘
. * \ ‘ “> \ '
. . . , j ' ss
, wo. ‘ 1
‘ ! t \ wy Py
t » - : \ .
. ' I ~ | :
1 ' ‘ ; . i (
+ ‘ : I -:N : to
. a : . 1
4 Co OO . an
‘ ‘ 4 '
; { . . \ t - N
; ; } \e ' { \ 1
t | . -
: ( : ’ 1 ' ‘
~ . ‘ ' ’ .
d ”~ \ ' . \ . .
‘ a { , . ' .
. .
/ o ; ~ ‘ 4 : “
‘ . , : 1 .
. ,
- » . ” f- ‘ ‘ -
{ , . ‘ ‘ ‘
. , ~ ‘ +
oe a ; ' . ty \
oe a 1 ros
, \ iv
4 , - . - ’ f
1 ‘ - .
’ ’ ‘ ( . aa . " . =~
' - g 7 : .
‘ ¢
: ’ ro: p!
' : 7 . : . cos
' ¢ \ . “s €
- - *
| . 7 . - “4 - - e }
© . . - 7 ' : .
, . , —— oe. , “ . * ue ‘
. é e . - a :
: 4 . . we -* \
‘ - 7 “ '
' “ . ’ “ j ; .
} , “ af °
1 ,
‘ { a“ { :
. - + o - 1
o wr’ 7 e ' , ° - ‘ :
. ~ , . - * a * '
: . , : ‘| +
; . t
' , - ‘ ° 2 ,
. oe ‘ v - t .
r _~ 1 . .
' ‘
» 4 : - . .
: , . ~ of |
* —~ os new ~ aoe - = com —~ -- = . , | ~ ,
; wey “ :
rs “ ‘
* , -* ba’ ; .
a . . Xn ' r -
. . a . a 7 somos
” 7 ~ .
as a . a” a ‘
: ‘ \ a
‘ : :
“ - , f
- at - a ome ey ee * we ee ee ee
‘ ~ “
bees
4 f .
r ’ -”
=
- . ="
. , “
, , aie
ef mo
t-
. “4.
ee
eee
tee mee eee ee
13
vlot the cooling curve, it will taxe the form given in Figure 5,
Piate 2. Here we havetwo perts to the curve where solidification
takes place, namely BC and CD. Also we find thet no matter what
the percentaze of each constituent the line CD always falls at the
same temperature,
Case 3-
This is simply a combination of case 1? and 2. That
is metal A will nold some of metal B in solution and B will hold
some of A in solution. If we plot the cooling curves of a series
of alloys of both A and B, omitting the time factor, we get Figure
6, Plate 2. It will be seen that at the point b, or the eutectic
we have a percentage of each metal where soligification takes
place instantly. At this point we have two phases in the solid
state, consisting of two solid solutions, and also one liquid
pnase,
Taxing an alloy of a composition R, above a be,
we have all liquid, therefore one pnase and 2 degrees of freedom.
When the temperature L is reached, solidification begins. This
means two phases within the area a be, and the system is univar-
ient. The first crystals that form are solid solutions of A and B
having a composition 8. If we shift the temperature to S'L', the
composition of the crystals change to 8', and the comvosition of
tne liquid bath becomes L'. If we lower tne temperature to eb
the composition of the lighid becomes b, but below eb, everytning
is solid, therefore we get sudden precipitation of the liquid
whose composition is denoted by the point b. The structural
composition of the solid material would consist of crystals
of the solid solution of B in A, these being well formed, if the
cooling tooK place slowly, and an aggregate of finely divided sol-
14
utions of B in A and A in B, the coupositions of each being shown by
point G and H. However we still have just two phases, therefore
the system is univariant.
As the iron-carbon alloys belong to the class of binary
alloys, the constituents of which are partially soluble in each
other when cold, let us examine their fusibility curves, and
anoply the phase rule to then, In this case we have several allo-
tropic changes as well as changes of state. These forus can be
zrouped under nine heads as follows-
(1) - Liquid iron
(2) - Liquid Fesc
(3) = Solid solution of FezC in Gamma iron
(4) - Solid gamma iron
(5) - Solid beta iron
(6) - Solid alpha iron (ferrtie)
(7) = Solid solution of C or Fe x0 in beta iron
(8) - Solid Feo ( cementite)
(9) - Graphite
Figure 7, Plate 3, gives the cooling curves of the iron-carbon
alloys as far as ° 4%0C. Let us apply the phase rule to it.
Above the liquidus, we have only one liquid pvhese, and consequently
2 degrecs of freedom for
Fs 3-1 = 2
Within the area LSE, we have two phases, namely liquid solution
plus solid solution (austenite) and therefore only one degree of
freedon,
Within the area LAHDS, we heve but one phase, a solid
solution and here the system is bivariant. As we reacn the point
the point D, the system becomes non-variant, for nere we have
a ee ee
oo se oe
ee eee eo ee -awe we me
——->
5 ‘
we ee es .. .
“
Tm ee
.
hoe 1 > a on
wee
. nee we ee
ao.
—
ee ee wt we
awe 04 we ee mee.
See nme es
- were -
‘
ame ee ee --
: 4
~ . . . r .
-
~ oo
-
wR ‘ eet
‘ ‘ ye ae 4
ae
ee ee
wee owe
amen en em A
~ Meee
ne ~
tT
ae
a
‘ 7 ‘ wf
, : ty
. : : 2
_— . I - + :
*% ‘ +
y ceo
-, '
.
4
‘
t _
{ ty
eee ee Bet
‘ a ~~
t
ay |
. . ¢ |
arty}
\
y
‘
. 4
Mee
4
’
\
‘
‘
‘
‘
’
‘
'
j
s
'
wee eee foes ce od
i
‘
'
t é
“se
on
_oo 7
—<
er and
.
“
”
oo
a
*
oy oe.
1
‘, : ‘
, YN
a ‘
‘
a *
a
at
Q
: By
. o .
% \ . ne =
7 . ‘ .
. .
wy
me -
‘
a rea
16
austen ite ferrite, and cementite as the three phases. In the
area DSS'F we have two phases, namely austenite and cementite
and the systen is univariant. Within the region ABH we have
beta iron and austen ite and the system is univariant, while
in the region BODH, there are also two vhases, namely aus-
tentte and alvha iron, the system still remaining univariant
while finaly below CDF, there are several possible cases.
Case 1 - The cementite formed during the solidification
and cooling remains unchanged and we have a univariant system
the phases being ferrite and cementite. The system is how-
ever in a metastable condition. This Le the condition of the
slowly cooled steels of commerce.
Case 2 = The cementite has been completely changed over
to ferrite and «srapnhite, giving a condition of stable equil-
ibrium,
Case 3 ~ The cementite has been partly decomposed into
graphite, thus giving three phases, and therefore a non-
variant system, this being the condition of the slowly
cooled gray cast iron of commerce,
17
CHAPTER II
We find that in dealing with metals, particularly
in the solid state, the time factor in cooling plays an important
part in our final structure. If we wish to obtain material
from which we can pick out the different components, by means
of a medium powered microscope, it is necessary that we allow
the metal to change its state slow enough, that the differ-
ent components have time to adjust their crystals. At tne
same time, we cannot allow te long a neriod on account of
tne expense involved. As tnis work deals only with carbon
steels, I will outline briefly the results produced upon
the metal by differant rutes of cooling.
Wnen a pe@éce of steel is in the region LSDHA
Figure 7, Plate 3, it is composed wholly of austeni.te and is
in the solid state. As it passes thru the unper critical
ronge, there is a change of state, from one phase to two and
at the point D, we have three phases present. @6n account of
the material being in the solid state, diffusion takes place
much more slowly, thus requiring a longer time for the reaction
to complete itself. This means then that we have a change in
cyystaline structure, If the steel is allowed to gradually
epproach AHD, and gradually pass thru the transformation points
sufficient time will elapse so that the new structure will
be fairly perfectly formed into its basic structural
constituents. On the other hend, if the time is greatly
shortened, then the crystals will not have time to reform
18
before the metal becomes so cold that movenent is impossible
or at least almost infinitely retarded, and we have a structure
from which the components cannot be easily picked. However
this may increase the value of the metal from a commercial
standpoint for certain uses. I have attempted to show this
in Figures 8,9,and 10, Plate 4,
Let us take for in:tance a steel whose composition
is R, Figure 11, Plate 4a, nemely 0.5 %C. As we cool it to
1, tue first change from the solid solution takes place,
ferrite being thrown out from the crystals of austenite, thus
making the remaining sucter.ite richer in Fe,C. If these
ferrite crystals are allowed to prow pradually, they will
get a chance to form a fairly cefinite structure, but if we
suddenly stop this growth, by cooling the vhole mass sudden-
lv
wv )
then they will be toe fine to distinguish cleerly by the
usual means. Ont the other nend if the growth proceeds at a
normal rete, more ferrite will be thrown down, and segregate
in large crystals, till we reach the line CD, when there will
be a sudcen precipitation of the ewtectoid mixture of 1 part
cementite and 7 parts of ferrite, forming a fine sggregate.
Figure 12, Blete 4b, shows this teking vlece in five steps,
first eustenite crystals, lastly ferrite and pearlite.
As the range of temperature becomes smaller, where precipi-
tation can take place, that is ss we approech the eutectoid
point from either cirection, the cooling thru this range
must be morc gracual if good results are to be obteined.
Fowever some pearlite is likely to be caught in very thin
films between the ferrite, even with the slowest cooling
possible.
°
tH t rit TT i
ry ! r +444 3 I
ti + +4 7 t pio |
+ _ + +Tt++ |
rT +14 THT shea
| ort ttt t t + +344 Y rr
eee! tot +4 +. saa
+444 Ly } ror oi tt
ri ao i +
Tt ei ty Pat i rh
| + + H +4 tit
t : i | t 1
Tors | 7 { eee
ms + + me ae
t q rit 4H ttt do
Troha it } } Laat
} oe +4 | + +t oo es 44 +34 444 1 i
ttt Coe " ny ; iG 1 Ce as Le
TTT t +4 ri sae ee j ; }
} i+ Hi tae | ra na
it +H roe Teas ro soe i ;
rot rot saeeee naa eeeee aes
ioooe ; hy Seenee' mw seReeee
t I } ; { rity wri | : poy
my ws 1304 tt I iid :
od ++ ; t + + . ‘ |
+++ +++ + + + + + toa + . te te +
Wo aoa Pry nas tT Tae ttt i : | | |
tit Coa i { t ; no SEEee ia i 1 ! TTT
sanen ni rr titi SSESS POR SESRESS COR SSS EDR EERE: os By i +4
I I t ; it
i 3 |
4 titit itt }
tity
| : — ~ - _ 3 |6(N
- awe
--y--
st tee ne eee eee
TUS WAY
i :
wad
g ANSARI :
a “a : : :
B
“tyy
st
we
Ws 22
PY > Gs
r ‘ ng a
; c NANA os : :
| : ‘ }
thir ) f
y yy
1m t 4 cae a “ ve
-
7 yt vires a a ' '
7 Tort root i w
o jesegassanas gaceee +. — tc oe
qo : IIIT ET
AALS J LOA fff COLI, YY 7
Leff pls YI Sd fl LLNS A LL »
tthe Vi iy YY, 4 fA Md dpb lie Y ~
% J LS At Z f fe, 7 if v4
1 { : 4 : { #. . ' car
ee — 3
OEE
ry Iss
tt OM ee OO
“GR GY CA ON GAT FRA
; HO LL
t '
y '
, € '
i
‘
5
; ; F 5
’ .
:
- - -_ - eee Sw Stet ome — me et te ee eee ee - - «- 4- ——< a —— =
a, ~
v3
-- > -
a7
tas
or)
rig te
273
2
ave
S
Qa eee NN
Las
“
WAVE
MS
—~=\
NOt au
ah
Vir.
‘,
\
Wy
«
4
<
Pe.
qa
Y
<
a?
+ %
1 oe
a ee a
© Ege
4
\ Nome
Gee ae
rea
5 te got
he 4
ad 2?
te Fee
CUGENE DIE TZOEN CO CHICARO EW TORE Noy 350
. a — ———————
eee
en ren
n
wae
meee
Zoe ve oa .
-—~
. - ’
a tty .
. ve 4
”
\ “oe we
. . . ra
- ve x,
oer. .
aor .
te 8 te ae.
' ™
eg WL Ga. ee
we
aetna = one noes
.
o ”.
~ .
.
. s
. oe
We eee Ne ee
tla ee ee
. ia
a)
ot
7 x
,
a .
a
wee +4
- -- + -A
Soe
- ~ ctw
"oy
- - an,
s
- 4.
~ : .
- ~ 4
7 Tae
- a |
: ne mee {
- sow
~ i?
. ot
rary
~ ~ y
toa
vee
~ A
- .
. - *
4
1
ome oe .
1
. ~ "
.
. oo - -
: . «
+ .
~ + - .
\ we . ~e ,
: ee ~ vat
ee ee eee oe
m- ows
+
»
ae eam eee ee
oe
:
‘ .
‘
ee
69 te a
. '
. foe
. “ONG
te Po ee © . oe meme ae: 2 OR. ee mas oH my
ee ee oe ~_— a Pom e-- +e a |
co "9 vos ~_ - coe oo ee ~~ wee e —- ~
r - ~ ee eee see -
‘7 = —-- . -- —- -
t : ~ > - ce = cote se .
,. - - - - - e- - . |
~— - - - eg. - L oe eee :
’ -
- - - - - -~ +e ~ ee - _ oo.
, * .
i « - ~ wee of motte a3
_— co RE
Roe - - 3 : - - - - --
- & =. wee . -<- ~
po “FU. 4
" - - - 2 - - wee : oy
oo, SBT Coe
bee -. - Mo eee j
3. - soe “5 - “+ oot - ’
ane Oo) eh ee Be etn Wolly wet A ee ee RS
ae .- : . ‘ wee -
- _ ~ . t
~ we eee - - - - ~ -- 4
- . om - we . 7
gore . “ s+ a . . 4
. , . . . ~ . .
woos . ve eee - - +e @
ee -- - mc o-e ry _— -- wee
a oe 1
“ . . ate Seo .e
Vo - 2. ~ --
- ae . . .@
ST, moa
Dee ~ ~- ou -- - os }
an . + ™ » - . re 4
' * . - ~ -
“os .
ee. le cee Be ee eee!
. . _ a” > ‘
a . vy - .
. r . . .
. ‘ a
. eo: -a’ . , ‘
'
cy |
7 7 ‘fs .
- .
- |
oom es "% é
-~ > ae , ?
-- c !
- ‘ $
: ~ i
wor ¢
TG
'
‘
Q)
t
i
i
t
4
‘
‘
‘
+ eee eet
-
. see
eo 1 Ger new - me me es - 7 ewe ee re me y
‘ °
1 i
, '
\
»
. : ’ . - se e
. . . : :
. wet - 3
- i , . t
rt : - ev '
? . ‘e wos - ’
, . ; ‘
: . - ve 5
q : we '
j
‘
3
'
5 .
‘
6 - ; Le
. ‘
2 . . ’
' ’ ¢
: ‘ - ‘ {
1 7 , tor ‘
. + . - 7. ‘
: |
: .
- . - - - - 2 Sarees pire e wot ‘
. ’ . : ‘
» +: . « . » & -
. id “.¥
. ~ a
x . - .
~ ~ . - . -t
, i ~ om tm 4
‘
. . . ‘ t >
‘ . 7
¢
mis tl s
PUM ENE UE TEOEN CO CHITACO KEM DBE ENO, IEG
oo
- '
, ~
”
, rot fogs My Qh or awe
ak - 4 w\ : AM :
‘ en
a - N ‘ Moe ot x oN
o af ae 7 ° - “~ *
ar X ’ “ ‘9 ‘A ,
. . ‘ ’ “
if “ * 4 : .
4 . ws >
e - . . ys
toe 1 : on + “4 "4 om ) 2
. jo . - a? . “7
- -—
‘- Dal
-
. eee wenm em ee ee en ee nm me es ee eee we we me ee me > ee Stee wee meee atte we fener fnew ee
Yr} . { a
: :
.
.
.
’
=.
2 ee wee eee
a
wwe
. os
. }
\ ’
: ; “
. 4
- 4 . “
‘ ‘ ‘ m sa
1 . ’
\ . 2
, v $ .
\ ;
% ’ .
1 ; ‘
* ‘
5 ‘, Yee mee nm - - ea ee Ae rm oe ee - ee ee aw ene sk s+ oe ~ sot .
.
‘ . 7. ‘
‘ 7
.
‘ \
‘
.
.
‘ ‘
\
2 ‘. .
.
. r
‘ , '
* : o- sd ~~ = ~~ - ~~.
‘ .
‘ 4K
‘ wr
"
4
%
iL
t \ .
' ‘ .
‘ ‘ -
‘ ;
y
, .
. ”
\
i
t 4 -°
, :
a? i .
: \ i
a ' 4
‘ .
q ‘ .
‘ ’
‘
. "
, : .
-. ’ ‘
. ,
: : ’
‘
‘ : ,
y
‘
‘ .
. ’
’
4 ‘
‘ .
. ‘
' A
.
= '
. e
. -: - .- ~~ nie oe eee wa ay ~ wee ew ue at tet mere wt ne na we we ee aot oe bee ee - wee tw lt ee oa «se
.
22
If a definite relationship can be established between
the pvercenteges of these constituents and the percent of Carbon
present, we will have a simple and accurate method of determin-
ing the vercent of carbon, Commerciel carbon steels are never
pure alloys of iron end carbon, and the emount of the imputities
affects both the critical points and the position of the
eutectoid point. However the amount of impurties is limited
by the specifications under which the steel is purchased, and
for carbon steele of fair grade, the amount allowed will not
introduce enough error to cause any trouble.
In the case of hypo-eutectoid steels tne relation-—-
Bhip between the structural and chemical composition as re=-
gards carbon, may be found as follows-
we know that the whole mass is composed of ferrite
and pearlite, therefore-
re P = Foo
Where F = the percent of ferrite and P = the percent of
pearlite. Also all the carbon is contained in the pearl-
ite, therefore&
E/129xP = C
Where E = the percent of carbon in the pearlite and © = the
percent of carbon in the steel. Also as the carbon is
present in the shane of FesC, by molecular weights Pex
contains 6.67 % C. We do not however know the exact posit-
ion of the eutectic voint, but simply that it lies very close
to 0.85 % C, the exact ~osition devending upon the chemical
composition and the neat treatment the peéce has received.
Assuming that it is 0.85 % C, we have
6.67/100 x “CO = 0.85
2)
2 C = 100/6 67x 0.85 = 12.74
or the ratio of cementite to ferrite in the pearlite is about
1 to 7, which places the eutectoid point at 0.834 %C. From
this we can easily calculate the structural composition of
a hypo-eutectoid steel. The total cexentite will be = to 15C
and there is eignt times es much pearlite as cezentite, we will
heve the total veerlite = total cementite time & or P will
equal 120 C, Figure 13, Plate 5, gives the results of plotting
the percent of pearlite for any percent of carbon. In the case
of hyper-eutectoid steels we have only free ceinentite and
pearlite and here
P + @m = 100
also B/10c P4+6,67/100 Cm = C
as all the ferrite is contained in the pearlite, and therefore
tne total ferrite = F = 7/& P,
or “p= 8/7 PF, but FP» 105 - Cm
therefore P = 3/7(100 - total Cm)
but Cm = 15 C
substituting Pe &/7(100-15C)
or P= 114-176,
These values are also to be found in Figure 153. Plate 5, as well
as the corres»nonding values for cementite, total and free.
From Figure 13, we can therefore determine the vercentage of
the structural constituents. For instance, a steel containing
1.2% C, contains $2 % P and 13 % Cm, while a steel containing
0.5% C, has 42% F ane 60 $P.
On the other hand we can reverse the use of the
diagram. That is if we know the nercentaze of the structural
Cconetituents, we can determine the vercent of Carbon. For
Bye NE OPP UB Le ca nee Ome NU 35S
sates
ae
wt ee ee
— = ene Ow
wre em ee
“eo wom
“ene
news
ee
a ed
—
eek
2)
example, a steel containinz 30 % of p and 7) st of !F, contains
0.25 0. In order to make the reading of the diazran more
Bimple Figure 14, Plate 6, has been drawn, Here the pearlite
1s siven in one continous line, as are aiso the other
constituents, in order tnet the reading may be ore easily
done.
Thus we neve a clearly established relationshinv
existing between the structural components and the percent
of carbon present. The only renaining difficulty is to
me.sure the vercentages of the structural constituents
Simdly and accurabety. I? tnis can be done, we will have
a .etnod of deter:aining the carbon content, which will
be a great aid in checking the chemical commDosition as
deter:ined by the combustion method, if indeed it does not
tell us far more tnan the latter.
.
i i. ke
‘oe ww
i en
a a
~~
a
~-
- oe’ :
--_ .
7 ot x
aoe oO ~ .
-
—
ie x,
on“ N
_eeen
~~
a
_-_—-- q
an *
aoa
-~ar* \
we OT ‘
wee
aoe -
o_o s
oa
-
ant? .
*
y
‘
. X
‘
“
>
x
'
«
x
‘ 4K
"3 ' . Ks ; oo: * .
. . . : ,t ‘ .
so ‘ .. ’ No vo a .
1 ‘ se ‘ . us \ - oo * . ?
eo . 8 - ' - - . - oa
-y an , \ , - : . rt ‘
: . 1 > . . ‘ s ‘
> 2 \ a ass > “ . 1. - ‘ ‘ .
\ a.
- . . ‘ . oe . x
: t ’ \ s ws . 4 5 . )
3 * ‘ . :
. \ . sf .e -do. » s :
. °
soa r
.,
‘
.
.
‘
1
y
‘
x
:
.
%
XN
\
*y
“
‘
ne
- .
cone
. . oan 1+
. wa « ” ‘ . Me ’
nn) ‘ * - . e ee . ‘ ’ ve
. . tae
“.. «- - - - mere . . - - ‘
‘. - 7 oF . aa : .
awe ee - : we . wee vie _
eo, nee : Yo..
“ ce ,
- . Lo. eon . ' 5 -
x . ms ay ‘
- . ow . ‘
. se
~ - . ° Tee ee eo
. oo vom
sae . . Nee
~ . a -.
. « , - “—-
rT woe
- ten, .
oa
. - om 8 F . Se ee oe Fo eee 2 ow Com Mm Let ee! O62 aN wees em rem om 6. ww. 0 ehh ben ee a a Pm me am rome Leet ates a Oe we re a ~ e
ae - . oom o-
. ‘ , \ . - 4 .
' . . : ‘ ? . ‘ - .
. ‘ \ : \ : . : c . ‘ .
. . ( x + . ,? ae .
1 . - v4 4 { a
. ve ‘ ‘ NO:
* ‘ ‘
te oe * a ' ‘
of
CoAT LK iil
In order tnat a clear uncerstanding aay be had of
tne methods used in the following work, a snort cescription
is ziven below of tne annarstus used, None was used
excent sicn as wignt be found in any fairly well ejuipped
lasrctory. The saaple to be tested was polisned in the usual
manner, using “SF elundum on broad cloth for the final volish-
ing. The soeci.sens were etcned with a 5 % solution of HNO
?
—
in St 0H Two tyves of microscones were used. The first
Will bear cescrivtion, as it was reconstructed frome a reculer
instrument. In fact this one was used for all the work,
the re.cular :netallursical microscone not beinzs delivered till
late in tne spring teri, Tnis latter machine was a Bausch
and Lomb reszular OCM with a .sechanical stage and all the necessary
attacinaents, This was sounted upon a canera made by the saue
firm, liznt being furnisned by a new tyne arc, witn an aspheric
%
coudenser. This outfit is shown in figure 1°,, Plate 27.
A picture of the nomemade apparatus is given in Figure 16, Page
27. The microscove was fitted with a Leitz dark field illu:
inator, the substazge removed and in its place was substituted
an electro naynet, caneble of :novement in two directions by
means of slow :iotion screws, Lignt was furnisned by a Leitz
Lillivut are lann, fitted with a bulls eye condenser.
The lamp and microscove were siounted uvon a board and the
imaze thrown ezeinst a clear glass, fitted into a sxetching
board, suonorted by «a nickeled rod, the board being cxoavle
* THE NEW micnoscope did. net come till
May (4, 97.
FIGURE 16 | BOTH OVTFITS MOUNTED
ON THE SAME BASE.
METALLURGICAL LAB. MAC
FIGURE 17 SAME AS ABOVE, SHOWING
DETAILS.
fo
\O
of movement on the rod, and also horizontally around it.
The tracing paper was slioped under suitable clins on
tne board and the image traced off, This formed a con-
venient and easy meynod of sketching tne structure of the
pe@ce. It was found that a magnificetion of between 100
and 200 diameters, was great enoush, if the sample had
been given proper treatment, so that the various consti-
tuents could be easily distinguished, and the outlines of
tie crystals traced with a sharp pencil.
There are several factors, which will, if not
taken care of, produce results far from right.
Case l- Improver heat treatuent.
(A) Too rapid cooling.
(B) The temperature not being carried above the
unver critical voint.
Case 2 - Improver etching and polishing of the surface.
(A) Over etching
(B) Spoiled etching solution
(C) Improper drying of the specimen
(D) Poor polishing
Case 1 — (A)
If we cool too rapidly, we will get a structure i
which shows that the ferrite, if it is a low carbon steel,
and the cementite, if it is a high carbon steel, are tnrown
out suddenly, witn the result that the crystals become
jagged and sharp and poorly formed. Tnis is shown in Figure
18,Paze 30 and in Figure 19,Page 30. These pictures show
3s medium and a low carbon steel. Bothwere cooled, by simply
30
removing the pieces from the furnace and allowing them to cool
in air. The temperature of the furnace being About 1700
degrees F.
Ficure 18
Magnification of each about x 120
The effect of under heating is shown in Figure 20, this being
the same sample as Figure 18, but the temperature at which
the sample was treated was about 1000 degrees F. It will be
noted that the structure of the ptece is the same as that of
the specimen when regeived, this being shown in Figure 21.
Figure 29 Figure 21
vo
Heated to 1000 deg. F As received
Magnification x 120
51
Case 2 - (A)(3)(C) If we allow the specimen to remain
in the etching batn for a long period of time, as the action
of this bath is one ofcerros’ on, the mere outline of the
crystals will not snow, but the different crystals tnexselves
will be eaten away, causing blacking of the surface in snots
end areas will appear nearlitic, which are really ferrite.
The actual error introduced will be shown later on. If the
etching solution has been used a nuuaber of times, it will
contain iron salts in sufficient quanity to produce
deposits uvon the surface, which will give a very different
structure tnan tne real one. Thnis will also introduce error,
Then if tue soecimen is not dried rapidly, films of oxide
Will denosit , wnich will spoil the surface for carbon
deter:sinations.
(D) If just the carbon content is desired, then the
surface need not be entirely free from scratches, all tho
very few should be present. If we desire to keep a ver&
menant record of the sauple in tne for:n of a photograph
, then the surface must be free from scratcnes.
In some of tn&B work we wished to know tne actual
size of tne crystals, and also the distance covered by tne
drawing. The method used, while not the most accurate, gave
good results and wes very simple. The draw slide frou a
slide rule was inserted under the microscope, instead of the
snecimen, and tne instrument focused on the log seale.
As tne divisions cane to a sharp point, even at 200 diametes
the distance could be accuratly :.easured between points,
and this distance measured a--inst the actuel aceale. Tne
nuuber of times »reater,the cistance between points ther
D2
one division of the log scale would give tne mugnification.
If this was found to be 160, then by a simple calculation
one inch of screen represents 1/162 of an inch on the
sxecimen or 8.0062". One square incn of the screen then
ne@ans 0.000050 sq.in. of svecizien surface. The cross
section paner usec contains 20 lines ver inch, therefore
490 areas ver sq.in. If a crystal covers 50 areas on
the section peper, the actual crystal covers 0,7 00356
times 0.125 = 0.0000043 sq.in of surface.
Or!
\N
ChActin IV
xeeping in mind the errors which can ecsily
occur, I wished to formulate a i:ethod if possible, by which
the carbon content of a given sample might be determined
by an exemination of its structure thru the microscope. This
1etnod must be simple enough to be used by the students, and
at tne seme time the results obtained must check combus-
tions run on tne s-me ssmole.
First six ty»ical sammles were selected. These
were of Carnenter steel of very good cuality, and gave a
range of carbon from low to 1.2 %, The samples were
nunbered as follows- A,D.E,F,G,H. The above samples were
analysed by cirect combustion. The absorbing medium being
Bakers Soda Li:ne for Compustions and consisting of
Nace = 37 %
Ca(OH), 2 45 5:
90 2 15 %
Filings were taken frou several bars at various places a.id
well mixed, s0 that the results obtained would be general
as far e353 nossib.s.
The results obtained are show in Table 1, Page 34,
™:4a@ pnelysis falls within the limits usually required, Then by
tie methods vreviously described the carbon percentage
was ceter-zined.
Accordingly a .1l@cze was cut et random fron onB
of the bars, marked sample A, treated and tested.
XN
Table 1
No. of % C ~ ¢C % C
Sample Trial 1 Trial 2 Average
A 0.16 0.18 I.17
D 0.28 0.31 0.30
E 0.64 O 67 0.65
F 0.78 0.80 0.75
G 0.85 0.90 0 69
H 1.19 1.21 1.20
The average percent of carbon was taken as the
standard value for checking the microscopicel analysis,
oD,
Three ps@ces were cut from A anc marked A=1,A-2,A-3. A-1
and A-2, were heated to 1700 degrees F. A-2 wes co-t*0 in
the air A-f, slowly cooled in the furnace tiru a period
of 24 hours. During the first 6 hours the temperature dropved
from 1700 to i200, during the remainder of the time to 75
degrees, The sample wes then cut in Llo,:. so that the «nterjor
might be workec with, where no cecarcurizing effects had
taken place. After cutting A~-1, the surface wes filed and
enouyh filings obteined to run a combustion. This giving
the chemical composition of the steel as close to the stru-
cture examined as possible. The snecimens were polished,
etched with 5 % HNO, and examined. The structure of A-2
was sketched upon cross section paper, within a two inch
circle. Three different voints upon the surfece were taken
, average structure being selected. Plate 7, Page 36, gives
the original drewings under figures A-1-1,A-1-2,A-1-3. In table
II will be found the results of the three dettérminations to-
gether with the results obtained by the combustion method.
Figure 22 Page 37, gives a photogrenh of the specimen.
After polishing and etching, Sample A-2, was
exaniinec and e photogrenh is giveh under Figure 25, Page 3/.
It will be seen that the crystals are distorted, and overé
lap eachother in such a manner that it would be impossible
to sketch them accuratly.
The method of determining any given area of
crystels wes as follows. The number of squeres within the
area were counted, and as each square was@ passed, a slight
mark was placed upon it with tne »-ensil. The counts were
mede in series of 100 at a ti.ce.
SUG ENE DIEZTZGEN CO_Ch
,
1 -
tok, -
ie 4
ee Y
: . -;
. * a .
= se .
° 4
Sy :
toe “a
”
- , .
~ -!@& as a - .
: .
‘ ’ .
mo » ! y .
oe ta .
. -- . a -
-
, . ,
. «
’
- .
3
he ees
aa . ? , r
+ ‘ .
. :
. . ty,
:
‘
-
a lgli a +3 4
f
.
~ :
coe me tee. ade
TAA ST.
i
. ' ews ~
oe : .
om .
- ~ -
. { .
.
ee
‘
. ,
. , re -
. a
. . ‘
\ :
' . ~ . 8
t =
po . \
. = . .
.
. ‘
1 r * ‘
~.
‘ d
» ‘ ot - me
. . ; 7
' «-
, ne! .
: -
. - ae
. s , ‘ Nar
. te
‘ aa
’ »
a y
’ ‘
wee fon +. eer
: oc
4-
-———
at
Figure 22 Fighre 25
Sample A-2
Heated to 1700 F
Cooled in air.
Sample A-1
Heated to 1700 F
Cooled in the
furnace
Figure 23 Figure 24
Sample A~3 Semple C-1
Heated to 1530 PF Heated to 1700 F
Cooled in the Cooled in air,
furnece
38
TABLE II
Semple Heated Cooled Aree of Area of %P %&%0C “ ¢
No. to in Pearlite Circle Combus,
TI-T2-Av
A-1-1 1700 Furnace 221 1259 18 0.15
A-1-2 1700 nom 198 1250 16 0,18 .16=.18=.17
A-1-3 1790 "oN 204 1250 16 0,13
From the above table it can be seen thet the arees
hed a maxium variation of 23 areas out of 221 or about 10 %,
But for 0.1 % GC it requires about 150 areas, and in the 150
there would be a maxiun variation of 15, or this would be
equivalent to 0,01 “ C. Our chemical anelysis ceme only
within 0.02 %, giving an average of 0.17 %, while the average
by the physicel method was 0. 143 “, or 0.027 % below that
obtained by thie cherictl 1) ess, This result checked very
well with the seneral analysis made, and given in table 1,
The next ps@:se A-5, was treated as follows,
"xf
eated to 1900 F and cooled ta the furnace, This treataent
took the neece only a little way above the critical point.
The staple wag polished and etched in the srsze man:er as A-1,
end the results are xivena in Plate 7, Pave 36. The aress
ares A-§-1, A-3-2,
‘a
o
of the differeit points are given in Fi
om
-
Cc
ry
_
i
NX
oe
- . ’ y, 7 s ve - ”
and A-5-%., Fisure 24%, Pase 37, gives & photogrs,
2?
From the pnotosranhs of the different saanles
taken fro bar A, it can be seen thet slow furnace coolinzg
is necesurry, and that for low carmon stecl3s the meximum
teanerabure snould be above tne upner critical point. That
tenderatures between 1500 and 1700 F, will give good results
iP the cooling be slow enough.
Next sa anples D,E,G,H were given the sane trealinent
ag oar A, and the resulta are shown in Table 3. The sys tea
of awabering the different viéces is the same as given for
sumple A. Theat is the letter stands for the bar number, the
first fizure for the pe@cre taxen from the bar, and the second
number for the ares of the a@ce,
It will be noted on examining this date, that in
b
neorly every case the % pf carbon by this method fell a
Little below that obtained by the combustion method. Also
that this di*rerence was very neerly constant except around
tne eutectic point. At this noint, and for a little way on
each side of it, it is impossible, even witn the slowest
treatment, whnicn could be economically given, to distinzuisna
eitner free ferrite or free ceaentite. I believe tne reason
tnat the nicroscovial anslysis ran a little below the cheraical
was due to the fact, that in the mechaniam of solidification
some of the nenrlite is caught between the ferrite grains
and thus lies in such fine films that it can not be measured
at the neznifisation used, In tne case of high car0a
steels, the ceuentite areas were courted, and the sane held
true here. If we sive each sample the sane heat treatuent,
for like carbon content, we will get about tne same error.
TABLE
Samvle Heated Cooled Area of
No. to in Pearlite
A-1-1 1790 Furnace 221
Aet-2 1790 "oR 198
Ae1-3 9 1750 #" ° 204
Averase
A-3-1 15090 Furnace 222
Am3-2 = 159 nn 233
A-3-5 = =1500 oon 218
Avera,e
D-1-1 1700 Furnece 452
Det-2 1720 "4 415
De1-3 1700 on 458
PVeCre e
EFe1-1 1750 Furnace 975
Eet-2 1720 mR G78
Ee1-3—= 1720 "oon G68
AVELrEA OC
Fe1-1 1720 Furnace 1141
Fel-2 17C0 rn 1144
F-1-% 1770 m4 1151
Avernve
Ill
Area of
S%ircle
12590
1290
1259
ret
‘7 P
18
16
16
18
19
16
36
2?
78
738
TT
7
3 ©
9.15
0.30
cf
(2 &
Combus,
mo
+
7
,
17
17
SN
7)
oVé,
1? yee
és ~
> en. -
VEYe xe
<3
Sample A,-
Sample D,-
Sample E,-=
Sample F,-
Sample G,-
Sanple A,-
1790 Furnace
1700 "oA
7oo 202€«©ZL Cin
1720 Furnece
1700 4
1720 "oR
The original tracings
Plate
Plate
Plate
Plate
Plate
‘Plate
Table 3, con't,
all
all
ell
1174
1184
1172
nay be found on the following Plates.
7, Page
1) Page
11,Page
11,Page
41
36
42
4
43
4d
Ad
199
100
100
od
G9
94
IAN Derr EN Ce
cHicaae
mew -ORe
NG
390
pwacens
sn
or
Ao kan obrol
“
GC eo .
. fa
e
v a
Ay
a
AY ra
Ny 3
aN
: Ss
AA
me RENN an. ol
a
Yaa an
Dy
ante
wet pre ee eee an ce
1 -
oe ~
ay: # :
Ra? NE
Ns
S x2, PA 2 f i a
7 eT “7
‘ Se :
r ore a as het ee
; sea ee ee
ae, Dae: cords a <9 a tape ee 2 E
: : ;
‘ $ t 2 a 7 z ‘ 2
1
t .
2 ’
. j :
; '
i
4 a Dens 25 é a, Ses tS “ ia tay Cae) Hg ae Selo
DAY,
VOSA AD VAL
Pk SN
FLQENE
©
Ta SECO
cn cacs
iw rere
ee.
sou
3!
“
Aan
o
1
:
:
a eT ie
no
+.
a
ag
4
in
.
if
ee eee
|
SAA TSS
etmenh seem
:
’
1
oi Sy taro:
o hate
| ee ae ee
ee
AAT OR
BWA AQ WAR
Ne
ee os eo aes
1
:
4
1
'
i
:
TAMARA AD
+
'
i
'
:
vn
ASARAWSA HG BARI
awe
S
.
'
ARRAN WL Woe
& x
*
nay
Sass eune
hey
hy) Rey
a
»
ee ee ee eee a - --
4
ney
Hts
An Tee KR WS
a YAMAZ
SN)
&
ASAE
ay
Vriwne
vee
=i
ee
esd
ee
.
ee
w
G
oe
nee tCLORN CT OH CASO VE ORE ONG
‘
. bets .
j ‘ ‘
; ; ;
} | | | |
' : ‘ os ;
i ' .
, ‘ ' :
—_— - ar ee =? Mas = >
hy ss ee
x SO ihe
ad eee 3
; ss “i a .
a oe Shae am
eid Bae Gas ie
ahr ae Cu we ANE 4
b: wee by
+t eas
tid AER Ce
: ,* = oS
u¢ Je z
, ens ss a *
: oa wes !
‘ '
Sag ‘as wie e -
ae. gash Dhl a ¢ &
‘ $i. oie : L Fi
z - ee co
» - *
: 5 - at 5. uae
ee on :
Fs io
wv?
A
WG LVDS
1 . ‘
.<
- oth ow pan gi
“ Pa ft
oo
a “
3 ee
z : !
4 a |
es a” - Pd
' ¢ gh
z é ‘
4 . <
te ov See oh 2
” Cad
‘
: 2?
is
s
e a r
© 2
e :
1 «
- ‘A
z
cha ays ax = od
' ' ; f
ve oi , a a. ; "ay 8 a :
. |
$e Sng - me wee 8 ee te et wee eos een - -——- we eee ee i ee ee ee ee ee ee ae mee ee Si -
45
The results given in Teble 3, I have plotted on
Plate 12,Page 46. Not using the aversge value for the
difference, but the maximum and minimun, for each sample.
and tnus obtaining a curve which broadens out at tne pnoints
where error occurs. The true pearlite curve is also given,
as taken from Plates 4 and 5. It will be noticed in table
3, that sample H was heated to 1-CO F instead of 179) F.
This was so it could be taken above tne Acm line in order
to set complete precivitation of the cementite.
On Pages 47 and 48, may be found vhotogranvhs
of tne various samples, showing the incresse in the amount
of vearlite present uo to the eutectoid point and the
cementite surrounding tne grains in the high carbon steel.
In orcer to use the curve given in Plate 12, vroc
proceed as follows-
Determine the area of pearlite at tnree different
points on the surface of the metal. Two values for the
percent of carbon will be found for each percent of stru&
ctural constituents. This will give sx values in all.
Take the average of tne six. If the percent of vearlite
runs over 595, count the other constituent and subtract
from 1250, or whatever tne total area of the drawing
ha»pens to be.
Now the reverse of what has been given was tried.
The cardon content of several senvles was determined in
this manner, and the result checked by the combustion
netnod. The results are viven in table 4. Page 49.
Each of tne saanles will be taken un in detail.
eta
> em ie ct te
tu
GENE DIETZOEN CO. CHICAGOWEW YORK NO. 350
_
Oe od
- ~ at ~ - . ~
“Na . wy “
toe - ‘ Me ‘ ‘ i
t ‘ ‘ ° ‘ . - .
at . an , . 1 - f,
‘ ‘ . . . . ~~
‘ ‘ x ~ . 4 . . !
> N
~ - - . . ,
a . a > * , bo, é
. . ‘ . °
2 ON ‘ tie Teel
~ ~ ‘
eoNty ved
TA Ue
™N *,
. a. ‘
. wet i,
- SNA
“NA .
a ees .
a 4
Ma Os
fos “hs te me
San. be
‘ } aN 5
~
AN
“ENON
’ ‘
$l 4
. Ct
en ne
1 SNR
ae
NA
VN.
."
as .
oe
ws
wo,
Mia
"
nN _
‘ 8 .
‘\ ‘ +
\ ‘
+ 7
a .
°
Moy
o
®.
g. .
Um
a
.
.. .
® . ‘
. oA .
aw a
wae ee ee - -- - +e wee ee fee ee ee ee mr meme em mee ee eee as eee ee wee
5 .
. LL. . cee ee ml we eee oY
. ~, " ¥ . ’
eo oe ¥ Lae xa wd+ . “ . i : a ve
Low carbon Steel, about 0.2 % X 200
Heated to 1700 F, and cooled in the furnace,
Medium Carbon steel, about 0.5 %. J 170
Heated to 1700 F, cooled in the furnace
48
Eutectoid Steel, about 0.88 % Cc, X 150
Heated to 1700 F, cooled in‘ the Furnace
High Carbon Steel, about 1.2 %, X 150
Heated to 1900 F, cooled in the furnace.
49
TASLE IV
Sample Heated Area of Area of % P % C % C bifferenee
No, to Circle P Min. MAX.Ave.
Bele? 1709 1259 37630 26 27 S27
26 eg 90
Be1-2 1722 8 1250 367 29 25.26
Be1l-3 1700 159 185 15 14 16 Aree on ede of
seugole
BaJ-1 1520 1250 22) 28. 624 .25 SAME AS pB-!
4 . 02
Ban 2 TX 125) 545 a? 023 625
Ba5-p VWJOQ a PD 14 13. 16 frea on ecie of
stiaole
Cx1-1 #1729 1250 692 55 47 .48 .48 EFS
8 12
C-1-2 1700 1259 541 Ady 238 1349
C-1-3 17200 1250 552 Ad 37 .29
Sele1 61790 1250 168 13 14-415
S-1-2 1790 1250 1620 1513-15 1H
S-1-5 1750 1250 166 13 13-415
ATe4-1 1720 1250 4o7 34 = 28 1350
AT#1-2 1790 129) 335 51 27 .28 .27) Bdeow of cnse nercaened
smiaole Bun. osced to
AT=1=% 1700 1250 495 32 226° 29 be 1.2 50
50
Case (1) Samole B
The sanmnle wes heated to 1700 F and cooled in the
furnace for 24 hours. The s»ecimen was then sawed in two and
tne tnterior examined, The structure is shown in Plate 6, Page 5]
Figures B-1-1 end B-1!-2, were portions of the surface, taken
on the enterior of the surface, and , and these results
cneck the combustion determination very closely. In Figure
B-1-4, we have a detersination which does not check, This
drawing was made on the very edge of the sample and shows
tnat the prolonged neating an an oxidizing atmosphere, caused
a decarburization of the exterior of the sample.
Case (2) Samnle B
This was off the same bar as the above sample, but
as will be seen on tiie data sheet, the sample was only heated
to 1DOOF instead of 1700 F. In this case the crystaline
structure is not as well defined, the crystals not naving
had as much time to grow, and are tnerefore somewhat
sualler. In the case of B-$-$,and B-3-2, the results check
pretty well. Good enough for commercial analysis at least,
Figure B-3-3, again shows the effect of prolonged heating
on the carbon content.
Case (3) Samnle C
This sample was taken from the forze shop stock
from some :aterial which had been giving trouble. It was su
sunnosed to be about 0.6 % C. Some of the bars would harden
allright, others would not. Some of the bars seemed to harden
better in the center than on the ends. The sanple tested
was taken from the end of one of the bars. Figure C-1-2
and C-1-%, show the true content of the bar, C-1-1
aan
i
; } TT +
wile - J - ‘ ‘ ‘ s ' AUGENE DIETZOEN co. cHICAcO-nEWVORK NO, 350
a of” it i ella aed i el eee LP, cain ah mats Be
-
» .
.
2 3
. .
a
? . ?
7 2 . :
- Ps — Be asi
‘ : iS . eee Ee, : «
$ < an TAL rod e
3 av Pag or Ss +% ”
Y eo e
a f : =
ae * . “
3 2 ri
2 2 :
: 4
a 2 f :
a i 5g
. 7 “ - Y .
“ « Ww
- 2 at
3 a Ra 1
4
¥ - FA ;
a? Z r*
, ,
~
2 *
we “
>is
.
:
ce
:
ws
-
c 53 a
“
o
“a a
&
~ ‘
~ J .
a a
% e% a
e et r
‘ ;
1 Z i A
« e % <
, ; Bs Z 5
< ‘ >
be 4
io 36 4
1 zs z S
‘ e :
i t
~ : ¢
Pe %
i 3
Sng
. :
a4
i
pe
gives a much nigner percentage tnan the other two. This
Bar:ple was taxen to snow tie effect of overetchifg, and
shows that e mistake in this direction can be made very
easily. The ot:er crawinjgs check the carbon content as
determined by the combustion method, C-1-1 and C-1-2~2
ere found on Plate 9, Page 4i¢,
Case (4) Sanple 8
This sam.le came from the Gier Pressed Steel Co.
After prover treetment, the drawinvs of the semnole were made
and are given on Plate 13, Page $5, Here care was used in
all three ereas exrmined to get characteristic portions of
the surface, The results obteined in Fi;ures S-1-1, S-1-2,
S-1-%, check very closely with the carbon content found by
the combustion method. In this particular test, The actual
time consu..ed in the overetion. after the sannle was taken
from the furnace was 18 minutes. While the time teken for
the combustion was 20 minutes. This shows that the time
consumed is just about the seme for either method.
Case (5) Semple AT
This saunle came from the Feo Notor Co. It was
teken on the ecge of a paece of cese hardened material. The
case was sunnosecd to be about 1.2 % C, but it would not
harden when quenched. On examination, as will be seen from
Figures AT-1-1, AT-1-2, AT -1-3, Plate 13, Page 53, the carbon
content of the case did not exceed 0.27 %. It wes found that
the trouble was cue to faulty pyrometers, and that the temper-
ature was not high enough. The semple suggested to the author
that here a case mignt vresent it self, where the method of
determining the D of C, might be of value. It is impossible
NeGescace Fa ORF NG ble
a
.
“/
«
«
‘
¢
'
woe
>
rc
!
~
. . «
‘ ,
oy
~S
-
.
“1
. 4
fo of 4,
7. tr
"¢
.
vt
‘
‘
.
t
“+ = 44
cor
. ec
- a ye
te le
toe
. " ao
ow e
‘
vif
\.
Ne er?
: ‘
ON 4
aan
.
.
‘
- ---
- 7 lan
.
oc
e
@
fw
a
'
\
’
as
’
@ ae
Sm
’
'
neon
y
-
~ 8
o™
™
ofan
i+y%
”
s
ot.
Ne.
‘ os
crt
“ny ~*~
f
a
54
scanple of low carbon steel showing
the effect of different rates of cooling on
the structure of the specimen,
Saaple as received X 170
=’
1
7
X 150
Quenched in water X 150
22
to determine the carbon content of the case by the combus-
tion method. for if it is only 1/16 of and inch deep, we
could not be sure that our filings: did not have some of the
softer material. Yet if this method worked, the Carbon
content could be determined at any depth in the case with
eeseand accuracy.
Consequently 4 specimens of case hardened material
were obtained, these being hereafter designated as samples
AT,HT,HS,HR. Sample AT has already been described, Sample
HR, was foundto be cased to 0.52 %C. The other two were
found to run to eutectoid composition. On Page 56 and 57
are shown photographs of these different svecimens.
Plate 14, Page 58, gives a drawing of a section
of sample HT. This drawing is a good deal larger than
the image thrown upon the screen, and was obtained by care=
fully moving the paper and the paéce in unison. Altho
this requires very careful work, and takes a little
experience, the results justy the effort. Here the specimen
was magnified 160 ) and onthe drawing the actual depth
of case has been scaled off. It will be seen that the
eutectoid composition extends in 0.0103 ". The chart at the
bottom of the plate gives the carbon content for each
half inch of depth on the drawing or for eash 0.0031"
of actual metal.
Plate 16, Page 59, shows the case of sample *$
HS-1-1, Great care was used in meking the drewing, and the
case had been so put in, that we might heve a gradual
56
Sauple AT X 50 Sample AT X 100
Slowly cooled from 1700 F Outside of case
Sample HS X 50 Sample HT X 50
Slowly cooled from 1700 F Slowly cooled from
1700 F.
2URESE FT TIOEN CG CUCARS FEM TORE NO 350
»
es
a. 3
ses ~~
eek AN
Bu inyt ee Ae
Ca x
yo
en- eee
-_—_< wee ee ef ee et ey ts eee
2a et pt - nee pa ye se we
’ ’
1
. t
$ (
oe ee eee
HI
are ae = f
: “a
7 mo
a A
: Pe
Lh wowace tase. ee
' a,
5, Seer
a
“a
esis SE bak oer Shoes
A
‘
: x
4 ela
; :
net) ee
+
i Ge
x aga
a“ an
og
Si gsdhe = Reh SoS
Lat = a9
cr. ae
Ge 5) Se
Mees ent She
' ae ny re
a ~ ae
SS =
oak. SoA
Y 48 ¥
- Zz
“a %
i a ;
} ae
' en
tt ty
bm
&
oO
27
nk
. *
wo
-
bo
we
og
wed
“4
Sy Aaa
aay
a4"
ee ee = ee
oe
ee +
ne ee ne ee ee eee
= eg
+
te.
ee
ne i
1
a4
- x “a
* A bef CE Hot
we
al
>-
>
60
increase in tne vercent of carbon from the center out.
This is also siown in sample JR, nere we again
have the effect of cecerturizing action endon firet examin
ation it avpeered as if the exterior of the case might run
as high as 1.4 %C,. But section i! anc O only had pearlitic
sreas of 529 and 443 out of GOC. This would have mede the
carbon content run up avove the limit for steel. Yerit .
looked as though a very hirh cese hac been obtained,
This tyve of exeaninstion sives one a very good
clue ab to what hes harnened to the rterte, an wet or the
meat to which the rece was carrie? with tore curburizin-:
anterial used was oro der,
.-. sak 7 4 ’
righ carbon steel,
a od
The last ssandle tried was a
1
owas supposed to de of sutectoid composition.
wile
Nowever tho totel arca of pearlite was below 199 ™, and
conpideraule cementite snowed uo, This gave anaverace nnaligis
for the eee of J6 % vearlite, or 1.2 3 0. The crawings
for this sveciuwen will »e found on Plate 17, Tase €2,
Figures (U-1-1,'-1-2- "U-1-3.
Tall
ablee 5 and 6, give the results for saaples HR
5 : : hy om
and "5, These will be found on veses ut and ©}.
PL OEN® DF ZGEN CO CCAod HEM rom NO 350
* L ~
1
1 '
’ ‘ ‘ i . ' '
, ' : \ ; ' {
t ‘i ‘ * : :
ei tae oe . 2 Sete ‘ og '
‘ ‘
3 t h ‘ ‘ | | i
‘ ' f 7
! i r : i : : 1
' ; ‘ 4 | 2
' s , ’ } '
‘ 7 7 | 1 ’ '
iene aleeieetdnadiitesieatnth eiematedntamentanallt aiianetiadiadatned neterieteaeetiaen tee ssc eee t = i _
: : 7 ' '
i
3
:
1 , ‘ > ; ;
! ; ;
! ‘ $ . e ’ ‘ . :
— ae eg Rm th er - Q- +e £ -we 1 Spee wee mee ee ' -
| : ee Se ae ae oe i |
if) 4 ; “ > ‘ ‘ i '
- eet ee ee
VR
pate s
(¢
re ae —-
ee
ee ”
uf cn
t
’ Te
nS mi f
oO —
i ahaSd se
aS
2
“OSA SSN
2
or
=
Zs
: &
e ' 1 ! Wes Tey:
2 . £ a ‘ a, © at
‘ . ‘ : ‘ x
‘ 4 ‘ i ' SP,
: ’ 3 i ' bs ae
: 1 ‘ 1 a
+ e ‘ t
Sah) sh Stim) S preyed -Setengdsases” Saiietesmiecaewseas Sisk ea Gece od de. eae 2 ie bs
} ' 1 . .
¥ 1
. : ‘ 1 ‘
‘ ' .
' 1 t ' '
i i ‘ » ’
’ ‘ . ' {
. +” . ‘ ace: ihe . > «
’ H } : e . re
. =
‘ : ' ‘ ae
+ ' ; ,
‘ i ; ! ' ¢
1 : 1 4 =
1 ‘ ‘
siete sie pte Peewee matte Ae paces t Secession, cal hhc ae as Sak eae, ade rato a oa a ae * = mbeen. 1a
POrNe
-
‘
ETZGEN CG CHICA: O-NEW YORK
NO 350
: }
' 5 \ 1 ; ‘
; \ ‘
' . : 1 ; ‘ j
¥ > . - -- es eee - 4 - ‘ . . . -
: 7 3 Mr cree > ot, : : ; ' ss |
i ' ‘ 1 ‘ ; ‘ ¢ 1
' 1 . | 1 1 . ;
i ‘ ‘ ; : : ' ! ; S a . * ; ;
| i : : , ; t : : \ :
1 i . _ ee — . 1
erage ected pean eae, sarees eased end eee ee eg ee ee eee Ce ee ap a Tee SSeareawie esa pa ih ess = (wiles re ee * i shies eS Si. Setsee Se i 4
| an: | , 7 . 7 | : |
' ! i é ; '
| | | | 3 i ; |
‘ ' £ ‘ .
\ : ' ; GQ. ise: ower al whup tt. 2 ims ius 4 ‘
: ee pe oe ies a4 = Ae A S be) * ey : i t 5
t ’ . t ‘
| 1 ; ‘ { | t m
: : . ) . ° “ 1
‘ : a fe | . ‘ | { a
:
EGGaibeewiicisetar jeaberGa simaes en Loan@nuaes antecet 3 reins Soe od) (Gee veaeeeni ee scitiaecees eer sabe eben: «al lee : wale - 1 me greeters ee =
! f ‘ ' : a nt :
' ‘ ‘ ; i> soos ‘ a {
5 a P . ’ 4 ’ . <
5 i : 7 ; 5
' % ‘ :
i ; a
Twa ae . . eS ” -. z |
; . { an ‘ 1
i ! :
, a8 |
' r
i rae ‘
- > - ee me te . fc; :
i : “7 ;
.
: a ee 3
: A
os : ; *) oo
Li 4 i 2S 4
A
i a3
: : 7 + 1
peta ade fie a eae ns Coe b&b
i i 1
' | : / ' |
: S
3 g :
: t “te at .
oy S. on if ‘ re K
4 /
H 1 : ‘
; % 4 é : : a ax |
: : : ; es ' \ be i 3 :
fk: Mramsteades” (amines PSS & Ghee’ Opn Seis We Seam ys Scud cheese wes Stem Winaesae & aa = Sees it SS “me.
: . ne ; a : 3 , |
‘ ' : i j ' 2
’ : : ; i u 35 i e i
; te ‘ "4 : : x a ‘
4) doce ae heh irae ar . noe Ps ate Oy se e onely ve i - 7 Sy p ;
: ; 3 " ral ‘ 4 '
; a ; ; c :
A ; 3 : i ;
i ' : 4 ; 4 : : : ;
‘. i ~ ' .
‘ ; ‘ S652 M ko gene oe eh. aI : - al sap Aas
¢ & ingides " mene swe ie ede 4 so see ihe cee eee oe eee wy Sele a
i ! ‘t
i ae 1 “ >
‘ ‘ a3 ; ‘ ow ; a
me sais ig soe tickih (Sa Le pag 39 .
; 1 A 4 ‘ ? os
: . ' : ! ‘ ey : ,
: ! ' :
' ‘ za ‘ mae
— ae ————— a me — ~- - ee eh a ete ee ow nee eet eS ee —-ees oo = -_ ° > = *
Section
td
TY
a’
TALL
Area or
Penarlive
64
Arca of
Sircle
“
vo P
19
14
An
Cc...
ND
XS
ee
~~
74
i . . ' ; :
‘ ae: :
' 1 ‘
' ; ‘ : > '
t . 5 : ee ‘ ee . . + a ie oe Sb ae . . a . 7 2 !
1 oe ' : ' 1 : 4 1 ae :
4 : : ' ' : a se 7 ie {
i ; ' ‘ “i : ‘ a {
: ‘ i 1 : : 2
4 54
; { ‘ i ! ; : 7 /
' ‘ 1 . ' i
. ‘ ¥ |
SS. is ~ | ; ‘ | \ ; ‘
' ‘ ‘ ' * ' :
' ‘ : ! ei
' ‘ i % | i } i r ;
5s cee dae eam ag ten esehe, “thuakitmaiica: so sieeeete lee kage ele Sr PG RUS Sty sek penW comes bnbiee Soi og ee eee oe cate Seok eet ENS Blames aon Bloe - 4° = ’ sha.) oe seri Tey achote
4
| k : a ~ i
‘ : 7 '
: x + S00 ae , ‘
‘ : F 1
'
: : ss
’ -"
: a : :
: .
\
' os,
Pe ee ee ee Ob eee rr . < :
i Be st ;
‘ fe
on oe
. of
’ o a “9 bits
-
- Pd ’
t '
, .
< : fo 2 ‘
’ -
pean >. «= A weg come wee = ,
i es
‘ «i!
+
‘ as
= cs H oF
' « a a s hg vic): Gebes i
' o ,
$ ‘
1 St coer Se geem he ‘ = a
1 i . t
: « + .
: : o ; : 4 ‘ . :
? ' . ‘ a : . ‘ ‘ i
Wis 3 8 Soe aaibug F Sixte 3 cLaae oy aor 13 x fi . a 4 :
' + . 7 ~
i . ; ‘ i 4 : rn
i ‘ we : . ‘*
e ‘ i. . om |
; ! i 4 : A ; : 4
: “A : % is + .
, : . ‘ a es ‘e A te a oh. 9 I
sat ah igh, GP ia Seas Sky fe) a beta eoSeeee> BES h“ qi Banana e = th oimbege iia, iegeres = Gece eS) od Pr = mi = i Je *
° as . on on it -..
- . ‘ ’ : o ; !
a ’ Z > ss
“1 A ! ' , ‘ : : “ :
i Eee A “ ae
‘ oe & : : ay i ‘ 4s
; aoe ' ss : a u
Se : An
* n 3 ‘ ie vs er ee is 4 4s - to: oy - .
i re ‘
ts
wae eae 7 -—— wee eee oe
4 '
-- =e a et ee wen eee eee me de ee ee
64
TAGUL Y
. , - of pt
Section Area of Area of 33 P “ C “Oo
Mecrline Jircle Min, lax, Avera oe
tJ
BA EQ) 14 19 IES a
oN -° < r
D 152 me 2% 6200 22 22
tz
)
Oo
CU
©
XS:
On
XN
©
WN
\
fo
o>)
\
G p29 650 64 oO,
Mi
Xl
©
Cy
Ne
H 412 60
JX
Nf)
J
CD
AN
I
tr
[-
.
Sh
ON
©
CD
Co
fo
TN
_y
©
~J
C3)
XA
J
Co
M 550 690 57 290 6992 9
~
)
LS)
—~d
a
" B29 690 27 730.75
©
=
WN
CON
€)
©
oJ
oo
CN
iJ
Sl
Oy
XN
\
cnx
XM
69
TABLE VI
Sample HS
Section Area of Area of % P 3 @ 20
Pearlite Section Min.-Nax, Average
A 53 890 7 080 105 Oo
B 63 80.) & 090 .110 215
C 95 800 12 .120 . 135 13
D 132 890 17 6155 165 16
E 165 aad 21 2190 .205 L229
F S44 8090 24 210 .215 Jo
G 201 800 25 »220 .230 24
H 273 G29 a4 6290 300 235
I 38 890 4a .360 379 37
K 374 &00 47 400 .410 ae
L 417 890 52 AAD L450 AES
M 430 B29 54, 460 .470 ‘I
N AST G00 65 510 .520 82
66
Samples AT HS HT
Showing cases of different carbon content.
-—- ——_r— -
Photograph showing tne effect of improper
7
polishing and etching of the specimen.
67
Conclusion-
The author hones tnat by this work he has given
a glimpse into a field of analytical work which at present
is almost untrodden. The results obtaimed compare favorably
witn those obtained by chemical means. In the case of
material such as case hardened we can get results which
it would be impossible to get by the combustion met}od. For
tne combustion would give only the average carbon content,
whicn of course would be useless,
The entire work may be suaned up in e few brief
peragraphs. We know that the amount of pearlite im a slowly
cooled steel is proportio:.al to the carbon content, and as
steel is an alloy of iron and carbon, whose cooling curves
nave been accurately deter:nined, and als» as it is an alloy,
the cooling of which can be easily controlled, it ia possible
by proper heat treatuent to cause nearly complete seperation
of the ferrite and nearlite in hypo-eutectoid and the
pearlite and cementite in hyver-eutectoid steel. This
statement does not apply to alloy steels. Also by using
prover magnification, these differnet constituents may be
fatrrly well traced off and measured. The error introduced
is low and fairly constant, so that a correction curve ay
be plotted, which will give results that check the carbon
determinations ,by the combustion method very closely.
68
In the range from 0.79 to 0.90 4 C, it will be noted that
tne constituent of which there is very little present can
not be calculated, so that within this range our error
amounts to 0.06 ‘4. However tnis gives results as close
as most snecifications read within this range. On the
wnole a mistake is less likely by this method, for here we
bring out vhysical structure, and thus the chemical
constituents, while a chemical analysis calls for the
commolete destruction of tne sa..ple, and if some constant
error should creep in, it is likely that it would not be
detected,
In order to got good results, the following
points should be observed:¢
41) - The sneciimen should be heated to at least
1700 F, and slowly cooled in the furnace.
Steels above 1.0 %, can be heated to 1990
F, witn better results.
(2) - The snecimen must be properly polished and
etched. For slowly cooled steels, a 5 %
0% in CoHsOH, fives good results,
(3) —~ The different areas traced must be repre-
sentative of the whole vitece, and not of
a varticular voint, unless we wish to examine
that particular area,
(4) A magnification of between 122 and 290 X
high enough for a properly treated sample.
69
(5) - The tracing and measuring of tne constituents
must be carefully done, and the correction
curve used to determine the percent of carbon-
Count the area of the constituent which is
less than 50 %.
All the negatives used in this work in the
making of tne photogr:ohs are on file in the author's
labermtory at M.A C.
Prefor:aed in the Metallurgical Laboratory at
Tne Michigan Agricultural College
May 1919,