Argon Degassing

Practice And Results

by R. J. Chou let, R. L. W. Holmes, and

L. R. Chrzan

Previously reported results of gassing were 95%, 90%, and 70% for grades indicated increased drawabil-
tests on induction furnace and I-ton silicon, manganese, and aluminum, ity, reduced die wear and strand
electric furnace heats at Union Car- respectively. Close specifications for breaks, and improved welding arc
bide's Tonawanda Laboratories dem- reactive alloys, e.g. 0.01-0.02% zir- stability.
onstrated that argon degassing sig- conium, aluminum, and titanium, This paper will deal specifically
nificantly removed oxygen, nitrogen were easily met. Significant improve- with the results of the startup of
and hydrogen from steel.'-" These re- ment was observed in cleanliness the argon degassing unit at Wick-
sults indicated that the principal (Jernkontoret or Chevrolet ratings), wire Brothers, Inc., Cortland, N. Y.
benefits obtainable on a plant scale particularly in bearing, aircraft, and The major grade produced in their
would be improved steel cleanliness, forging quality steels. Very low sul- two ll-ton electric furnaces is a
yield, and alloy recoveries. Relative fur contents (0.005-0.010%) and sul- low carbon rimmed steel for wire
to vacuum degassing processes, ar- fide ratings were also obtained in products. Other grades include
gon degassing appeared to have these grades. The most striking im- C1015, C1020, C103S, and welding
certain advantages, including low provement in mechanical properties wire. A two-strand continuous cast-
operating costs independent of of steels argon degassed was evident ing billet machine has been installed
equipment utilization, low capital in transverse ductility and impact and is nearing completion of the
investment, low temperature losses, values. On several grades, transverse startup period. Wickwire manage-
simplicity of equipment and opera- ductility was equivalent to longitu- ment was aware of the unsolved
tion, and adaptability to various ex- dinal ductility. Other physical prop- metallurgical problems associated
isting shop layouts. erties generally were on the high with continuous casting of rimmed
Based on the laboratory results, side of the normal range. Prelim- steels and of the improvement in
field work began, and about 100 inary evaluation of welding wire quality and dissolved gas control as-
electric furnace heats ranging from
7 to 40 tons have been argon de-
gassed. These included bearing, air-
craft, forging, cold heading, welding
wire, Cr-Ni-Mo, Cr-Ni, Cr-Ni-Mo-V, 12 120
high manganese, rimmed, and alum-
inum-killed grades. Forty of these
heats were made at several plants.
At the request of these customers, 100
specific data and results are not in-
cluded in this report. However, the
general results are summarized as
follows: Depending on the grade, 8 80
product oxygen contents of 5 to 40 Z
ppm were obtained as compared
III Z
0 III
with 25 to 100 ppm in conventional 0 C!)
heats. In 20 heats made primarily for ~ 60 0
hydrogen removal, the average liq- c ~
to-
uid steel hydrogen content after ~
treatment ranged from 2.1 to 3.1 :J: Z
ppm. Nitrogen removal was not sig-
nificant, but furnace levels of 60 to
~
a.
40 :ea.
100 ppm were maintained in the a. a.
product as a result of protection
during tapping and teeming. Re- - -.......- -_ _ _.-120
coveries of alloys added after de-

R. J. CHOULET and R. L. W. HOLMES are

development engineers, Union Carbide Corp.,
Linde Div., Newark, N.J. L. R. CHRZAN,
formerly with the Linde Div., now is president
ARGON SCF/TON
of Missile Metals, Inc., Mountainside, N.J.
Fig. I-Argon requirement tor hydrogen and nitrogen removal at 291rF.

72-JOURNAL OF METALS, JANUARY 1966

sociated with degassing processes.'-l'
Hence, an ll-ton argon degassing
system was installed and opera~ed
on a pilot basis pending completIOn
of the casting machine installation.
To date, 60 heats have been argon
degassed and argon cast into ingot
molds. To avoid complicating the
casting machine break-in period,
only one argon degassed heat has
been cast through the machine, al-
though several degassed heats have
been cast through experimental tun-
dishes to investigate metal fiowrates
and argon shielding requirements.

DISCUSSION
Process
The mechanism of argon degas-
sing is quite similar to that of vac-
uum degassing. The theoretical
aspects of vacuum degassing, par-
ticularly with respect to carbon de-
oxidation, have been aptly discussed
in recent literature."·lO However,
there are several inherent differences
as discussed below.

Nucleation and Removal of Gases

The removal of dissolved gases
from steel by argon results from (1) Fig. 2-Argon degossing vessel.
the effect of argon bubbles in pro-
viding nucleation sites for reactions bles during degassing is not known; bon steel to an oxygen level of 40
which transfer the dissolved gases however, it has been calculated that ppm is 16.3 ft" per ton of steel, as-
from the steel to the gas phase, (2) the residence time of 2 cm CO bub- suming an initial oxygen content of
the improved reaction kinetics r e - bles nucleated at a bath depth of 3 254 ppm at a p ," of 1 atm. The cor-
sulting from stirring, and (3) the ft could be as long as 45 sec in an responding final carbon would then
driving force for these reactions pro- active boil.'· On the other hand, ex- be 0,064 %. The theoretical amount
vided by the low partial pressure perimental data show that the re- of argon required for carbon deoxi-
of CO, H2and N2 in the argon bub- sidence time of ga s bubbles in an dation of a 0.360 % carbon steel to
bles. The latter partial pressures inactive bath of the same depth is an oxy gen lev el of 14 ppm is 1.B ft"
amount to a maximum of B, 9, and 4 about 7 sec.'" Our experimental data per ton, assuming an initial oxygen
microns, respectively, in commercial and theoretical calculations indicate conte nt of 54 ppm at a P .o of 1 atm.
argon. that thiS latter residence time is The final carbon content would be
According to several authors, the sufficient for an argon bubble to be- 0,357 % ,
CO reaction is limited primarily by come saturated WIth CO. The rate The theoretical relationship fo r
the rate of nucleation of CO bubbles, limiting factor in carbon deoxidation the removal of hydrogen and nitro-
and this rate is appreciable only by argon degassing then, is the rate gen with argon degassing is shown
when gas filled refractory pores of argon introduction. It has been in Fig. 1. As shown, the amount of
serve as nucleation sites. H - 1 • For the shown that a residence time of 25 argon required to remove hydrogen
degassing reaction to occur, the CO sec for bubbles avera ging 2 cm in or nitrogen increases rapidly at low
pressure (P,u ) which is proportional diameter will permit a close ap- concentration. For example, to re-
to the product of the effective car- proach to hydrogen equilibrium.'" duce hydrogen from 5 ppm to 2.5
bon and oxygen contents of the bath, The nitrogen removal rate is very ppm requires about 46 ft" of argon
must be greater than the sum of the slow and surface active solutes, par- p er ton of steel ; however, to go from
barometric pressure, the ferrostatic ticularly oxygen and sulfur, mark- 3 ppm to 1.5 ppm would require
pressure, and the capillary over- edly decrease the rate by preferen- about 72 ft" of argon per ton of steel.
pressure. As the p ,." in the steel de- tially occupying nucleation sites, In low carbon steels, the significant
creases, the area of effective refrac- thus blocking the transfer of nitro- quantities of CO evolved act as an
tory with suitable reaction sites also gen atoms to the gas phase. It has inert gas with respect to hydrogen
decreases. These restrictions on car- been calculated that if the average and nitrogen and will reduce the
bon deoxidation are practically argon bubble size is about 2 mm, a amount of argon to remove any
eliminated in argon degassing by residence time of about 1 % sec given amount of these gases. How-
continuously injecting bubbles at would permit an approach to equi- ever, it should be noted that the de-
the bottom of the vessel. These bub- librium." However, no significant rivation of these equations is based
bles absorb CO, Hz and N o as they nitrogen removal was observed in on the assumption that the residence
rise through the bath. Argon injec- plant scale degassed heats. time of argon bubbles in the metal
tion causes considerable stirring of is sufficient for equilibrium condi-
the bath, thus improving the kinetics Theoretical Argon Requirements tions, and that chemical interaction
of the degassing reaction by (1) re- The method used to calculate the effects of other elements present
ducing the effect of diffusion bound- quantity of argon required to re- have not b een considered. 's
ary layers around the reaction sites, move dissolved gases by the above
(2) transporting the reaction sites mechanism has been previously de- Effect of Deoxidizing and Alloying
to the reactants, and (3) transport- scribed.",1. The equations used are Elements
ing the saturated bubbles away from given in the Appendix. The theoret- The effect of deoxidizing and al-
the newly injected reaction sites. The ical amount of argon required for loying elem ents on the activity of
actual residence time of argon bub- carbon deoxidation of a O.OBO% car- carbon, oxygen, nitrogen, and hydro-

JANUARY 1966, JOURNAL OF METALS-73

gen has been reported.,o,lo," For ex-
ample, the presence of alloying ele-
ments such as chromium and vana-
dium hinder carbon deoxidation
while the presence of nickel aids the
CO reaction. Elements such as tita-
nium and vanadium increase the
solubility of nitrogen in steel, thus
making its removal more difficult.
The formation of stable oxides and/
or nitrides inhibit the removal of
CO or N t. When the CO reaction is
the primary mechanism of deoxida-
tion, the presence of strong deox-
idizers such as zirconium, aluminum,
titanium, and silicon should be
avoided until after degassing. How-
ever, our experience indicates that
equal or better product oxygen con-
tents (5 to 15 ppm in some grades)
and cleanliness can be obtained by
adding various amounts of silicon
or aluminum before or during degas-
sing, but at the expense of lower al-
loy recovery. Stirring is known to
improve the kinetics of alloy deox-
idation, permitting equilibrium con-
ditions to be more closely ap-
proached.'" Extremely low oxide
contents are attainable if the oxides
formed can escape prior to solidifica-
tion and if reoxidation is prevented
during casting. Deoxidation results Fig , 3-Argon degassing.
obtained in vacuum degassed DR
heats, made with or without prior
deoxidation, have been reported by
Parke et aI., who describe the pro-
duction of cleaner steels through in-
clusion agglomeration and flotation."
It should be noted that the billet
oxygen contents obtained in argon
degassed low carbon steel, having
residual aluminum or silicon, ap-
proach the equilibrium values re-
ported by Hilty et al,"'
Equipment Design and Operation
Argon degassing is carried out by
tapping the steel into the degassing
vessel, injecting argon at the lower
sidewall of the vessel during (and
for a specified time after) tap, and
maintaining an argon atmosphere in
the ladle headspace to prevent the
pickup of oxygen, nitrogen, and hy- Fig. 4-Vessel teem stream protector.
drogen from air. Alloy additions are
made through the vessel cover open- space and tap stream protection is nozzle. When the argon flow is
ing and stirred with argon. The total dispensed at a rate of 5000 to 10,000 stopped, the gas between the shutoff
treatment time including tapping, scfh, is imbedded in the cover re- valve and the tuyere nozzle escapes
degassing, and alloying is about 6 fractory. The amount of argon w:ed into the metal until the gas pressure
min for ll-ton heats. The steel is for headspace protection depends on in the tuyere equals the ferrostatic
then argon cast into ingot molds. the number and sizes of openings in pressure. Within a few seconds,
The argon degassing equipment is the cover and the time required for metal freezes at the face of the noz-
specifically designed to fit Wickwire's degassing. Larger heats would re- zle.
physical plant and operations. Figs. quire significantly less argon per ton Improper argon injection tech-
2 and 3 show the degassing vessel for headspace protection. Headspace niques can result in poor argon ef-
during tap. The principal compo- protection is particularly critical ficiency, excessive splash, refractory
nents are the vessel, removable when degassing deslagged heats, ad- w ear, and temperature losses and
cover, and tuyere. The vessel is de- ding alloys which are strong oxide poor mixing. The angle, direction,
signed to allow a 2-ft freeboard or nitride formers, and in meeting and depth of injection and the argon
when full. The cover is refractory close specifications of reactive alloys. flow characteristics are critical fac-
lined and has openings for the fur- Argon is injected into the metal by tors in the formation of optimum
nace spout, stopper rod, and pre- means of a side-mounted steel tu- sized bubbles and the prevention of
heating burner. The spout opening yere. The tuyere nozzle has an aver- channeling. While the present tu-
is shaped to minimize open area, age life of four heats and is protected yere system is adequate, develop-
hence improve headspace protection by a refractory sleeve which is re- ment continues on improved injec-
and to direct argon over the tap placed every heat. Argon is injected tion equipment aimed at increasing
stream. An argon distribution pipe, into the metal at a maximum of 5000 both argon efficiency and degassing
through which the argon for head- scfh through a specially designed rates.

74-JOURNAL OF METALS, JANUARY 1966

 The argon casting procedure, pre- less steels and has been observed to standard 70 , 85, and 99% alumina
viously described in the literature, some degree in all of our argon de- brick recommended by refractory
includes mold purging with a low gassing trials.'"-"'-" The tendency to- firms) were evaluated at the vessel
velocity diffuser and argon shroud- ward increased skulling when slag line. After eight heats the wear
ing of the teeming stream.'" Figs. 4 casting at lower temperatures can of the 85, 90, and 99 % alumina
and 5 illustrate the argon casting apparently be offset by adequate brick was significantly less than the
equipment. It is handled so that ladle preheat and choice of refrac- 70 % alumina or chemically bonded
there is little or no interference with tories. 85 % alumina brick.
the teeming operation. Theoretical considerations of oxide
The total amount of argon, pre- Vessel Refractories stabilities and refractory reoxida-
sently used at Wickwire, in argon Agitation of the bath caused by tion rates suggest that lower oxygen
degassing ranges from 70 to 100 ft" argon injection results in increased contents would be obtained by using
per ton which includes 20 ft" per ton wear of conventional brick at the high alumina brick.'" While the num-
injected in the metal, 40 to 70 ft" per slag line. The severity of wear on ber of degassed heats is insufficient
ton for vessel headspace protection a given type of refractory depends for a statistical comparison, no dif-
and 10 ft" per ton for argon casting. on metal and slag composition, metal ference in product oxygen contents
temperature, slag volume, and argon attributable to different refractories,
injection rate, pressure, and time. has been observed in alloy steels.
At Wickwire, alloy (low FeO tap However, product oxygen contents of
slags) and low carbon (90 % de- low carbon heats degassed with high
slagged at tap) heats were argon de- alumina linings appeared to be 10-
gassed using conventional fireclay 20 % less than those of similar heats
and 70% alumina linings. The use made wi th fireclay linings.
of the latter resulted in increased At Wickwire, use of 85% alumina
temperature losses and skulling. brick at the vessel slag line and
However, these effects were mini- conventional fireclay brick in the re-
mized by high preheat temperature mainder of the vessel appears to be
and use of bonding mortar com- an acceptable compromise, consid-
patible with the brick. Use of 70 % ering costs, lining life, temperature
alumina brick (only one brand was losses, a nd quality requirements.
tested) at the slag line in argon de-
gassed alloy heats resulted in a lin- Practice
ing life equal to that obtained with The grades of steel which were
fireclay and conventional practice. degassed at Wickwire included
However, the life of the 70% alum- CI038, welding wire (MIL-E-23765,
ina lining in argon degassed low MIL Type 70S-3 and 70S-5) and
carbon heats was less than one-half modified CI006. Practices were var-
that obtained conventionally. It ied to obtain the minimum oxygen,
should be noted that the low carbon nitrogen, and hydrogen contents con-
Fig. 5-Argon Casting. heats to be degassed were tapped at sistent with temperature and com-
3075'F due to higher temperature position requirements for continuous
Temperature Losses losses incurred in the absence of a casting and with quality and cost
Temperature losses due to degas- slag cover and a minimum metal considerations. These practices were
sing, beyond those normally incur- temperature requirement of 2950°F evaluated on the basis of gas anal-
red during tap, necessitated an in- prior to continuous casting. Ingot yses, steel cleanliness, semi and fin-
crease in tap temperature of 35 to cast conventional low carbon heats ished product inspection, rolling and
60 ' F with normal vessel preheat of are tapped at 2955 ' F. In the latest wire drawing properties, and cus-
300 ' F. However, preheating the ves- degassing campaign on low carbon tomer evaluation. Liquid steel ox-
sel to 1300 ' F necessitated an increase heats, several brands of high alum- ygen samples were obtained by the
in tap temperature of only 5 to 40°F. ina brick (including chemically bomb sampling technique, while hy-
On the other hand, when a 70 % bonded 85% alumina brick, a high drogen samples were obtained by
alumina vessel lining was used density 90% alumina brick, and the refrigerated Taylor pin tech-
rather than a conventional fireclay
lining, an additional 30 to 40°F in-
crease in tap temperature was re- 50
quired. The losses caused by argon
degassing included (1) an estimated
l3 ' F due to an additional 2.5 min
vessel holding time, (2) the heat
absorbed by argon calculated to be z
4'F, and (3) radiation from the
well stirred metal to the vessel ""C)
cover estimated to be 15 to 40°F.'" >-
The variation in radiation losses is ><
due to the amount of slag present. o
If the slag is not removed prior to
degassing (low FeO slags) the loss ~
is about 15 If the heat is essentially
0. A.
des lagged prior to tap (high FeO A.
slags), the loss is about 400.

A novel effect of degassing is the

apparent increase in metal flowabil-
ity. As a result, argon degassed steel
has been cast at a temperature 30 to
70'F lower than corresponding con-
ventional grades. This phenomenon AIR ARGON AIR ARGON
is not fully understood; however, it
has been previously reported for C 1038 705-5
vacuum degassed rimmed and stain- Fig. 6--0xygen content of argon vs. air cast welding wire and Cl038 steel.

JANUARY 1966, JOURNAL OF METALS--75

nique."· '" Samples for oxygen anal- perature losses during degassing can ~ CONVENTIONAL
_ DEGASSED
ysis of the semi-finished product be practically negligible if the slag
were taken from the center billet is not removed and the vessel is ad- 120

of the first, middle, and last ingot equately preheated. Thus, these
100
of each heat. All gas analytical work grades can be degassed with little or
was done by the DCC Mining and no change in furnace practice. More- 8 80
Z

".....
w
Metals Div. Laboratory at Niagara over, if slag composition is con- Z
0
w
Falls. Oxygen and hydrogen anal- trolled, significant desulfurization
"
0 6 60
yses were determined by microvac-
uum fusion, while nitrogen was
can be obtained during degassing,
due to improved slag-metal mixing. .
'"
D

X
4 40
Z

......~
determined by micro-Kjeldahl meth-
Argon Requirements
......
~
ods. Steel cleanliness was evaluated 2 20

by an aluminate severity frequency The theoretical argon requirement

or a modified Jernkontoret rating for carbon deoxidation of CI038 or 0 0

(ASTM designation E45-63) in our welding wire grades is less than 2 ft" HYDROGEN NITROGEN

own laboratory. per ton but the actual amount used

was 10-35 fta per ton. The minimum Fig. 9-Hydrogen and nitrogen contents of
C1038 steel.
Slag Control practical argon requirement depends
Furnace slags containing large on additional oxygen sources in the mold size, teem stream characteris-
amounts of iron oxide must be pre- system ( slag, vessel atmosphere and tics, etc. Without inert gas protec-
vented from entering the degassing refractories, and iron oxide formed tion, the oxygen content of the de-
vessel in significant quantities be- in the vessel during preheat), the gassed steel increased toward the
cause iron oxide is partially reduced efficiency of argon injection, and equilibrium P eo value of 1 atm.
in the process. This increases ex- prior deoxidation practice. While When aluminum, silicon, etc. are
cessively the argon required for de- this minimum has not been estab- present, oxides are formed which
oxidation as well as making carbon lished, about 15 fta per ton is used can be trapped internally, or at the
control difficult. For example, the for welding wire grades and 20 ft" surface of the ingot. Nitrogen pick-
per ton for CI038 in the present up is usually not significant, but in
200 practice. The actual hydrogen re- the presence of aluminum or tita-
z... moval in the CI038 and MIL 70S-3 nium, it can be severe. Hydrogen
CI
grades was 35 to 75% of theoretical. pickup depends on water vapor in
)0- The theoretical amount of argon the teeming atmosphere.
X
o required to reduce the oxygen con- Figs. 6 and 7 show the difference
tent of low carbon steel to about 40 between air cast and argon cast in-
......
~
ppm is 16 fta per ton. Based on ad- gots of the same degassed heats. Air
ditional oxygen available as a result cast CI006 steels rose or bled in the
of inadequate slag removal and on molds, while argon cast CI006 steels
data from degassed open heats, it is were dead. Air cast aluminum-
estimated that several times the killed steels were dead in the mold;
above amount of argon would be however, aluminum recoveries in
required to cm' bon kiLl a CI006 heat the latter were much lower than for
at Wickwire . Injection of such a argon cast steels.
large amount of argon into an 11-
ton heat would result in excessive RESULTS
AIR
Fig. 7-0xygen content of argon vs. air cast
temperature loss. Since Wickwire re- C1038 Grade
quires a minimum temperature of Billet analyses of conventional and
C1006 steel containing 0.Q1 % silicon and
2950 F after degassing, a satisfactory
0

a rgon degassed CI038 heats are

0.01 % aluminum .
practice has been developed which shown in Table 1. In conventional
tap slags of 20 degassed low carbon combines carbon deoxidation with double slag h eats 0.035 % Al was
heats averaged 15.2%FeO. With an alloy deoxidation. For this practice added in the furnace at slag-off and
estimated slag weight of 200 lb per the amount of argon actually in- 0.10% was plunged in the ladle. As
ton, complete reduction of FeO jected is 20 fta per ton. a result of treating the reducing slag
would add 0.34 % 0 to the metal. Argon Casting with FeSi, the tap silicon was 0.05 % .
Therefore, the practice on modified Failure to protect degassed metal About 0.25 % Si was added in the
CI006 grades was to remove about during teeming will nullify the ef- ladle. An oxygen content of 28 ppm
90% of the furnace slag prior to tap fects of degassing. Results of argon was obtained with a residual alum-
by back tipping the furnace. Of the inum content of 0.082%. Degassed
casting tests on vacuum degassed
remaining slag, about 75% of the steel have been reported."' The heats were made in the same manner
FeO was reduced during degassing except no aluminum was added in
amount of air pickup depends on
and this supplied about 260 ppm ox- the furnace or vessel. Fig. 8 shows
metal composition, ladle nozzle and
ygen to the metal. When carbon de- the comparative severity ratings of
oxidation is the primary degassing f2Z:2I CONVENTIONAL oxide and sulfide inclusions for con-
mechanism, additional argon, treat- _ DEGASSED ventional and degassed heats. While
ment time, and about 0.02% Care oxygen and sulfur contents of de-
required to remove this oxygen. gassed heats were about the same as
The tap slags of CI038 and weld- those of conventional heats, metal-

..
ing wire grades contained from 1 ~2 lographic examination indicated the
to 3% FeO. Although the slag data " degassed heats were significantly
were not adequate to determine ac- >4Z cleaner. Fig. 9 shows the nitrogen
curately the amount of oxygen sup-
plied to the metal, metal carbon :S!
1M

:: ~
and hydrogen content of degassed
and conventional heats. The reduc-
analyses indicated that about 30% of ......> tion in product nitrogen content for
the FeO was reduced and this added ~
;;: Z the degassed steel can be attributed
about 70 to 200 ppm oxygen to the c< 0
u
to the reduced aluminum residual
metal. Since the amount of oxygen and argon protection during tapping
available in the metal and slag was and teeming.
relatively low in these grades, slag OXIDES SULFIDES
The production of degassed C1038
removal prior to degassing was not Fig. 8-0xide and sulfide inclusion ratings of steel without the use of aluminum
necessary. As described earlier, tem- C1038 steel. resulted in (1) savings in aluminum,

76---JOURNAL OF METALS, JANUARY 1966

(2) steel with an aluminum com- reduced oxygen content and alum- ~CONVENTIONAL

position suitable for continuous cast- inate inclusion ratings obtained in 60

_DEGASSED

ing, (3) decreased billet rejections, the degassed heats. Billets of con-
and (4) an end to customer com- ventional heats had extremely heavy
plaints of sporadic cutting tool scabs and seams resulting in high z
breakage. rejections, while billets of the de-
gassed heats had fewer and lighter ..:;o
><
WELDING WIRE GRADES scabs, very few seams, no rejections,
(M I L-E-23765) and exhibited excellent rolling char- ~
MIL Type 70S-3 acteristics. For the production of this
grade by continuous casting, Wick-
Billet analyses of conventional wire expects to make the required
and degassed 70S-3 heats are given aluminum addition with an alum- OXYGEN ALUMINATE
in Table II. Aluminum additions inum wire feeding device in the CONTI NT INCLUSIONS
were omitted in the production of mold. Fig. II-Oxygen content and aluminate inclu·
degassed heats. The silicon content sian ratings of Mil type 70S-5.
at tap was about 0.20%, depending Modified C1006 Grade
on the amount of SiMn added to 2. Addition of CaSi in the vessel
Steelmakers engaged in continuous
meet the manganese specification. during tap and degassing.
casting generally experience dif- 3. Addition of Si metal or SiMn in
The balance of the silicon, required ficulties in casting (1) rimmed steels
to meet the specification, was added the furnace as a partial block prior
and (2) steels containing more than
after degassing. It should be noted to degassing.
0.025% Al.'-" Whether or not rim- 4. Combination of 1, 2, or 3 with
that nitrogen contents in the de- med steel can be continuously cast
gassed heats were about 20 % less an aluminum addition in the vessel
successfully (with or without prior
than that of conventional and de- after degassing.
degassing) or indeed, if there is In each of these practices about 90 %
gassed heats. Fig. 10 shows the ox- any advantage in producing rimmed
ygen content and oxide severity rat- of the slag was removed in the fur-
steels if killed or semi-killed steels nace, the remaining slag was shaped
ings of conventional and degassed can be produced with the necessary
heats. Preliminary customer evalua- with Carbocite, fine ferrosilicon, or
properties, has not been estab- aluminum shot, the heat was tapped
tion of degassed welding wire, rela- lished."-l0 Aluminum is not satisfac-
tive to conventional product from at 0.09 % C and 3075°F, and the steel
tory for the required control of CO was deoxidized with about 20 ft' of
several suppliers, indicated lower evolution during continuous casting,
oxygen, nitrogen, and hydrogen and argon per ton.
because it results in clogging of tun- Fig. 12 illustrates the billet ox-
improved welding arc stability. dish nozzles. Silicon is used on many ygen content obtained in making
MIL Type 70S-5 grades but it adversely affects draw- four C1006 specifications. The man-
Billet analyses of conventional ability and cannot exceed 0.01 % for ganese ranged from 0.25 to 0.40%.
and degassed 70S-5 heats are given deep drawing sheet applications. The steels in Group I rose and bled
in Table III. In the conventional To permit evaluation of steels for when cast in 10 x 10 in. ingot molds
double slag practice for this grade, wire drawing applications contain- but have not been ruled out for con-
0.031 % Al was added at slagoff. The ing varying amounts of Al and Si, tinuous casting into 2 x 2 in. billets.
tap silicon depended on the amount the following practices were inves- Steels in Group II, III, and IV were
of SiMn added to meet the manga- tigated. fully killed in the ingot molds. Some
1. Carbon deoxidation by argon
nese specification. The final alum- of the steels in Group II were made
inum addition was made in the form degassing. with an 0.008% boron addition in
of 1fz in. diameter aluminum rods the vessel during tap and have a
~ CONVENTIONAL
stacked in the ingot molds prior to 40 _ DEGASSED
residual boron content of 0.002 to
teeming. Six degassed heats were 0.003%." Group IV steels were made

.. e.
made in the same manner except 100 ~ to an aluminum-killed specification
that the aluminum addition at slag- 30 by the addition of 1.2 lb of aluminum
off was replaced by the addition of >"
w
per ton in the vessel after degassing.
"~.. 20
110 ~
~
0.05% Si as SiMn. Fig. 11 shows the ~ ~ This compares with a normal alum-
60 ... ;:
inum addition of 4 to 6 lb per ton.
...
> Z
:Ii
40 ~;: The fully killed steels were eval-
Table I. Billet Analyses of Cl038
10
~Z
WO uated primarily on their rolling and
"v
20 drawing characteristics in the rod
Mn. Sl. S, AI. 0
ppm
and wire mills. Groups III and IV
% % % % 0 ......."'"'''--
OXYGEN OXIDE
L...-- 0
steels, which contained less than
CONTENT INCLUSIONS 0.01 % Si, exhibited unique rolling
Conventional and drawing properties including a
Six Heats 0.77 0.29 0.012 0.082 28 Fig. 10-0xygen content and oxide inclusion
Argon Degassed decrease in the number of strand
Two Heats 0.69 0.19 0.003 ratings of Mil type 70S-3.
0.014 29 breaks and die wear. Group II steels,
containing 0.03 to 0.06% Si and
Table II . Billet Analyses of Mil Type 705-3 0.002-0.003% B were processed with
no difficulty in the rod or wire mill.
C, ?O Mn, % Si, % S, % AI. % O. ppm N. ppm The quality and properties were
- -- -- - - -- - --. equivalent to good conventional
Conventional rimmed steel. Group II steels, which
Two Heats 0.10 1.10 0.55 0.012 0.06 32 105 contained no boron, did not draw as
Argon Degassed
Two Heats 0.10 1.08 0.54 0.011 0.002 23 85 well as good conventional rimmed
steels.
Table III. Billet Analyses of Mil Type 70S-5 SUMMARY AND CONCLUSIONS
1. The argon degassing process is
C,-;O Mn.% Si,% S. % AI. % O.ppm N.ppm readily adaptable to electric furnace
practice for the production of alloy,
Conventional medium, or high carbon steels. Sig-
Two Heats 0.15 1.13 0.53 0.014 0.65 57 107 nificant and consistent improvement
Argon Degassed
Six Heats 0.12 1.07 0.35 0.011 0.57 18 97 in cleanliness, workability, and
mechanical properties have been ob-

JANUARY 1966, JOURNAL OF METALS-77

 20 10 R. G. Ward: An Introduction to the
Physical Chemistry of Iron and Steel Mak-
ing. Edward Arnold Ltd .• 1964. London.
20 J. Chipman: Incomplete Mixing in the
Deoxidation of Steel. TRANS. MET. Soc. AIME.
1962. vol. 224. p. 1288.
21 D. 1. Brown: JOURNAL OF METALS. 1965.
vol. 17. no. 4. p. 428. Fig. 9.
2!l M. F. Hoffman. P. G. Bailey. R. L. W.
Holmes: JOURNAL OF METALS. 1961, vol. 13.
no. 5. PP. 345-349.
160 "J. G. Henzel and J. Keverian: AIME
I 0.001 % Si, 0.001 % AI
Electric Furnace Proceedings. 1961. vol. 19.
p. 434.

n 0.03 - 0.06 % Si, 0.002 % AI

,. D. Hall and T. McHugh: JISI. 1965. vol.
203. no. 7. P. 665.
25 G. F. Huff. G. R. Bailey. J. H. Richards:

i~~;':;~~7~F METALS. 1952. vol. 4. no. 11. p.

m <0.01 % Si, < 0.01 % AI 26 H. Epstein. J. Chipman. N. J. Grant:
JOURNAL OF METALS. 1957. vol. 9. no. 4. P. 597.

DZ: < 0.01 % Si, % AI '" W. Wilson: JOURNAL OF METALS. 1961. vol.
Z 120 0.035 13. no. 5. p. 350-352.
28 W. Crafts and C. G. Chadwick: US pat-
ent 2.357.876. Sept. 12. 1944.
&1.11
~ APPENDIX
>-
>< Theoretical Argon Requirements
o for Degassing
Oxygen Removal
~ 80 4449 Pm ,
~ v. = ----loglO -C.O --
~ 4C,-30, C, O.
- 644 (C , -C2) (1)
where
v. amount of argon required
in ft 3 per ton of steel
initial wt. % carbon
40 initial wt. % oxygen
final wt. % carbon
final wt. % oxygen
total pressure on the bubble
as it leaves the metal in
atms.
(wt. % C) (wt. % 0) 1
o m

m
P,. Kfef.
I ][ where K is the equilibrium
constant for C + 0 = CO,g),
Fig. 12-0xygen content of modified Cl006 steel.
f, and fo are the activity co-
tained. Nitrogen content was main- of the Mining and Metals Div. for efficients for carbon and
tained at furnace levels and, where their valuable suggestions and as- oxygen.
specifically desired, hydrogen con- sistance in conducting the test pro-
tent was reduced to 2.1 to 3.1 ppm. gram. Note: Equation 1 is vaJid at P,.
2. Additional features associated = 1 atm or less. In practice the Fc.
with argon degassing these grades REFERENCES at tap is > 1. However, agitation in-
include improved alloy recoveries, 1 E. C. Nelson: Factors Governing the Ex- itiates a boil which drives the Feo
tent of Deoxidation of hon Carbon Alloys
ability to meet narrow specification by Argon Rinsing, paper given at Annual towards 1 atm without need for ex-
ranges, low temperature losses, and Meeting AIME. Feb. 22. 1962. tra argon.
'E. C. Nelson: Nitrogen Removal from
low product sulfur contents. Iron-Carbon Alloys by Argon Rinsing. paper
Hyd'rogen Removal
given at Annual Meeting. AIME. Feb. 22.
3. Argon degassing practices have 1962.
been developed for producing low
carbon steels having a temperature
a R. Houston and F. S. Death: JOURNAL OF
METALS. 1963. vol. 15. no. 3. p. 205-209. V H = 3870 K2 (~- ~)
• B. M. Hamilton: Iron and Steel Eng .• H2 H,
and composition suitable for contin- 1964. vol. 41. no. 9. p. 85. ,(2)
5 F. W. Rys, Iron and Steel Eng .• 1965. vol.
uous casting. Ingot casting of these 42. no. 4. p. 77. where:
steels resulted in fully killed pro- 6 W. Gerling and K. Baur: Steel and Coal.
1963. vol. 187 (Dec. 23). p. 1222. VH = amount of argon (plus
duct which exhibits excellent rolling 7 E. Plockinger and B. Tarmann: JOURNAL
and wire drawing properties. OF METALS. 1963. vol. 15. no. 11. p. 835-839.
hydrogen) in ft 3 per ton of
4. Froduction of low carbon alum- 8 F. Jaicks. L. Kraay. M. Tenenbaum: steel
inum-killed steels by argon degas-
JOURNAL OF METALS. 1957. vol. 9, no. 8. pp.
1057-1072.
initial wt. % hydrogen
sing resulted in a significant im- 9 Continuous Casting of Steel. B.I.S.I.. Spe- final wt. % hydrogen
cial Report No. 89. 1965, London. equilibrium constant for
provement in aluminum recovery '0 G. McManus: The Big Stir in Vacuum
and rolling and wire drawing prop- Degassing. Iron Age. Sept. 10. 1964. vol. 194. H 2 ,g) = 2H
.1 J. N. Hornak: Iron and Steel Eng .• 1965.
erties. Billet oxygen contents were vol. 42. no. 6. p. 73.
Nitrogen Removal
less than 50 ppm. " Vacuum Degassing: Thirty-Three. 1965.
vol. 3. no. 4. p. 57.
13 Transactions Vacuum Metallurgy Con-
ference, American Vacuum Society, Boston, VN = 276.4 K' (-.!:.-~)
ACKNOWLEDGMENTS 1964. N, N,
"' A. J. Parke. R. F. Kowal. F. O. Alti- (3)
The authors are grateful to the more: JOURNAL OF METALS, 1965, vol. 17, no. where:
management of Wickwire Brothers, 8. pp. 897-901.
15 J. E. Fogarty. Trans. Vac. Met. Cant .•
Inc., for permission to publish this American Vacuum Society, Boston, 1964, p. VN = amount of argon (plus ni-
paper, and they express their special 432. trogen) in ft" per ton of
J.6 Electric Furnace Steelmaking, vol. II:
thanks to C. Wickwire and H. Fhelps AIME. chapters 16 and 20. Interscience Pub- steel
and many others at Wickwire as well lishers. 1963. New York. N, initial wt. % nitrogen
17 A. G. Szekely: Technical Memo B-292.
as M. F. Hoffman and S. Solomon of Union Carbide Linde Div., Tonawanda, Sept. No final wt. % nitrogen
the Linde Div.'s Newark Develop- 8. 1960. K the equilibrium constant for
18 P. Dewsnap and G. Hoyle: JISI. 1965.
ment Laboratory, and to personnel vol. 203. no. 10. p. 988 N.. g ) = 2N

78-JOURNAL OF METALS, JANUARY 1966