MAC RAE, D. R. Plasma-arc technology ror rerroillloys, Part II. INFACON 6. rmcecdi/lg.~ of II,e 6111 Imemariofll/I Ferroalloys Congress.

Cape Town. Volume I. Johannesburg. SArMM, 1992. pp. 21-35.

Plasma-arc Technology for Ferroalloys, Part II

D.R. MACRAE
Bethlehem Steel Corporation, Bethlehem, USA

This paper, the second part of a series, updates the technology with an emphasis on
the selection of plasma-smelting furnaces ror the production of ferroalloys and on
related plasma applications for melting and heating.
The configurations of transferred and non-transferred arcs are reviewed, and the
trends in the design of plasma-arc reactor systems and plasma-processing tech-
nology are outlined. Current industrial plasma installations arc listed, and plasma
reactors for the production of ferroalloys are discussed in detail. The paper ends with
a list of over one-hundred publications.

Introduction accompanying high heat fluxes while maintaining control

The article 'Application of Plasma Technology to of the chemical potential of an oxidizing, reducing, or inert
Ferroalloy Processing' I was published in 1987. It discusses gaseous atmosphere. In addition, there was the perception
the rationale and fundamentals for the produclion of fer- that the plasma state could enhance the reaction kinetics
roalloys by plasma-carbothermic smelling reduction, and owing to the existence of activated gaseous ions and radi-
describes the processing fundamentals and the process cals. Plasma-generated activated nitrogen [(N) and (NYl
developments for plasma-smelting applications, as well as species have been used to increase the dissolution of nitro-
relevant commercial developments, for ferrochromium, fer- gen in steel melts above normal equilibrium levels. This so-
romanganese, ferrosilicon, ferromolybdenum, and ferro- called 'plasma' effect has not been commercialized. Also, it
vanadium. The present paper updates the technology with was ant.icipated that. by a rapid quenching, unique reaction
an emphasis on the selection of plasma-smelting furnaces products could be produced selectively. For example, the
for the production of ferroalloys and on related plasma rapid quenching of plasma-melted zircon sands can produce
applications for melting and heating. This is therefore Part a non-equilibrium solid from the silica phase that can be
II of lhe topic. leached to produce a zirconium oxide product. But, rapid
quenching techniques in mineral processing have had very
Background limited application. In general, the application of plasma-
arc technology has been as a high-temperature, high-heal-
Over the past twenty years, plasma-arc technology has
flux source of energy. The ability to regulate the energy
emerged as an alternative to the high-temperature process-
input under controlled gas conditions favours the develop-
ing of tines and bulk malerials at power levcls from 500
ment of materials processing that eliminates or minimizes
kW to 30 MW. In contrast, the power levels of plasma
the environmental impact.
torches in powder spray-coating technology and cutting of
metals are in the range 10 to 40 kW. Commercial uses have
been established not only for the production of ferroalloys Plasma Heaters: Transferred and Non-
but also for the melting of specially steels and refractory Transferred Arcs
metals, and for temperature control of the tundish melts for Plasma technology can clearly be divided into two distinct
continuous casters. There has been a substantial investment heater configurations: heaters designed for the bulk heating
in pilot-plant development and industrial installations of of condensed phases including both melts and solids, and
plasma processes for the treatment of stainless- and carbon- heaters designed for the bulk heating of gases. Both lypes
steel plant dusts. of heaters can be used for the heating of powders through
The present era of plasma technology was initiated during the transfer of heat by melt conduction andlor gaseous con-
the 1960s as <l result of the availability of plasma heaters vection. The condensed-phase hel.lters used for the process-
that were developed to simulate re-entry conditions for the ing of melts are termed 'transferred' -arc heaters since the
testing of space-vehicle materials. These heaters operated at arc is transferred from an electrode to the workpiece such
high power levels of up to 10 to 30 MW, which provided as l.l metal melt that serves as the other electrode. The gas
the incentive for considering their usc for large-scale indus- heaters developed for the testing of re-entry materials are
trial applications. Whereas these torches were used for the termed 'non-transferred' since the arc remains within the
high-temperature heating of gases, plasma-arc healers and heater barrel solely to heat up the gas. These modes of heat-
furnaces of up to several megawatts were also being devel- ing are shown in Figure 1, and their characteristics are
oped for the argon melting of specialty metals and steels. described in Table I. The unifying characteristic is that both
The promise of plasma technology for industrial applica- heaters involve a working gas in the arc environment. For a
tions was its ability to generate high temperatures and the non-transferred arc, the working gas can be the bulk gas

PLASMA-ARC TECHNOLOGY FOR FERROALLOYS 21

 TABLE I tuyeres in shaft furnaces. The mode of arc attachment. as
1\'10DEOF ARC ATIACHMENT: TRANSFERRED AND NON-TRANSr:ERRED either transferred or non-transferred. is only the fIrst of sev-
eral design options for the plasma device or heater. In addi-
Tr;ms(cITcd Non-tmnsfcrrcd
tion to the attachment mode, design options include the
Gas now/power ratio Low High
electrode materials (e.g. copper, tungsten, graphite); water-
Gas lype Inert/reactive Inert/reactive
cooled or non-water cooled eleclrodes; a.c. or d.c. opera-
Power level 40MW 6MW tion; the electrode polarity for d.c. arcs, or for single or
CurrenL voltage limits 100 exXI A 3000A mulliphase a.c. arcs; tbe composition and feed rate of the
I {XlO V 10000 V working gas; powder feeding upstream, within, or down-
Heater efficiency 90[01(X)% 75 to 85 % stream of the fully developed arc and/or through the elec-
Reactor function Bulk hC:llinA Gas healing trode body; arc stabilization by magnetic, mechanical,
Reactor design Crucible Shaft and/or aerodynamic means. A multitude of plasma dcvices
Scale-up Graphite electrodes Multiple-heater units have been designed based on various arrays of options, and
have been patented and promoted for high-temperature pro-
Catllad. Cal~od.
cessing applications, including the production of ferro-
(a)
e (b)
o alloys. It is not the intention in this paper to critically dis-
sect all the possible devices. It is sufficient to note that the
selection of the plasma device strongly determines the re-
actor configuration and design of the process system. The
purpose of this paper is to oUlline the development of plas-
ma ferroaHoy processing and to describe the direction that
the plasma-reactor designs have taken.
A major feature that distinguishcs a plasma-arc reactor
from the conventional electric-arc furnace is the influence
of the gas phase on the reactor design 2 . This distinction is
apparent' if experiences with plasma reactors are considered
in the analysis of submerged-arc operations using hollow
electrodes through which gases and/or solids are fed into
the reaction space of the arc attachment. For the design of a
plasma-arc reactor, a convenient starting point can be a
consideration of the role of the gas phase in heat and mass
transfer and any chemical reactions involved. The factors to
be decided on are the
(I) type of gas: inert or reacti ve;
Malerial 1o (2) Ihe tlowrate: low or high;
b. hoot.d (3) the interaction with the reactants: chemical or physical;
(4) the importance as a heat-transfer medium; and
FIGURE I. Plasma heaters
(a) Transferreu mode of arc anachmcni
(5) the elTecl on arc stability.
(b) Non-transferred mode of arc attachment Inappropriate plasma reactors were designed when the
concept of 'plasma enhanced reactions' via the excited gas
that is to be heated. For a transferred arc, the relative species was in vogue and the direct contact of particulates
flowrate of the working gas injected into the heater can be 'passing through the are' was emphasized. Also, .non-trans-
much smaller. ferred gas heaters have been inappropriately used for bulk
Over the past twenty years the major trend in the design processing when the low gas f10wrate of a transferred-arc
of plasma reactors has been the increased use of trans- system would have been a distinct advantage. The trend in
Fen'ed-arc systems. Low flowrates of gas decrease the total simpLificalion is also evident in the designs of lransferred-
energy requirements. If an expensive argon atmosphere is arc reactorS in which the water-cooled tungsten cathodes
required, the low flow rates Further minimize the costs. The have been replaced by hollow (non-water-cooled) graphite
size of the downstream equipment for gas handling, cool- eleclrodes 1u • A major breakthrough in the use of transferred
ing, and clean-up is minimized, which may represent sub- heaters for the processing of solid fine paniculates was the
stantial savings in capital and operating costs. Low gas vol- development of hollow graphite electrodes by
umes may be dictated by the thermodynamic conditions MintekJASEA. This development eventually resulted in the
(gas partial pressures and temperatures) that are critical for 30 MW (40 MY A) ferroehromium installation at
processes involving metal volatilization, intermediate Middelburg Steel & Alloys (MS&A) in Krugersdorp, South
gaseous species, and condensing of products. Africa. This evolution in reactor design occurred recently in
the development of the plasma processing of electric-arc fur-
Trends in the Design of Plasma-arc Reactor nace dust, which uses hollow graphite electrodes 67 instead of
Systems water-cooled tungsten cathodes. The relative advantages of
Plasma-arc reactor systems are built around the plasma-arc graphite-electrode operation are as follows:
device, which may be referred to as a beater, torch. genera- lower consumption of gas
tor. or gun. The plasma-arc reactor, which may be a simple no source of water in the furnace reactor
meller or a more complex smelter, includes a reactor cham- no limit on current-carrying capability
ber such as a crucible or hearth furnace, or may be a space possible use of nitrogen to replace argon
such as the reaction-zone cavity formed in front of plasma capital saving on deionized watcr circuit and on instru-

22 INFACON 6
 mentation common for the development and adoption of new process-
• may operate with fonrning slags es to take upwards of ten to twenty years. This is due (Q the
no energy loss through cooling or the electrodes inherent difficulties involved in the developmcnt of high-
no skilled repair required. temperature processes, the maturity of the entrenched tech-
The stated disadvantages are that there is a consumption of nology, and the substantial capital investment that is
the graphite eleclrode, and the insulation of the eleclrode required for new facilities. In addition, plasma processing
and roof seal is critical. has had this technological 'identity crisis', which has
It is inevitable that this spread of plasma technology to involved inappropriate applications and has been viewed as
less complex high-temperature applications and the domi- 'a solution looking for a problem'. Initially, inappropriate
nance of less complex plasma devices will continue. To applicitlions grew out of overly broad and ambitious expec-
paraphrase Albert Einstein, 'Plasma reactors should be tations for a new and 'glamorous' processing tool. Two
made as simple as possible but not too simple', or, in steel- highly visible current 'problems' for plasma processing are
makers' vemacular, 'KISS' - Keep It Simple. the increasing demands for cleaner and higher-quality met-
allurgical products, and the environmental demands in the
Trends in Plasma·processing Technology treatment of hazardous materials.
The major process trend has been the increasing develop-
ment of simple heating applications. Twenty years ago the Present Industrial Plasma Installations
high expectations for plasma technology included the most The present electrical capacity for the plasma-furnace pro-
complex of high-temperature chemical-reaction systems duct ion of ferrochromium and ferromanganese is about 100
such as the smelting of chromite are. Today, the greaterMW, as shown in Table U. In addition, there is commercial
market potential for repcated installations has emphasized capacity of about 25 MW for the melting of ferroalloy
the development of simple systems for maintaining the tem- revert fines, slags, and stainless-stecl dusts. There are indi-
perature of melts or providing an incremental energy input cations of a gradual growth in plasma-smelter capacity with
at high tcmperatures, for exam pic in continuous-casting the recent upgrade in 1989 of lhe MS&A ferrochromium
tundishes and in steelmaking-refining ladles. A sign of the smelter from 14 to 30 MW, and the proposed installation of
technical maturity of plasma technology has been the a 12 MW ferrochromium smclter at MacAlloy in
repeated installation by several vendors of the sallle devel- Charleslon, S. C. The SwedeChrollle planl 01" 48 MW has
opment. such as tundish melt heating. not been operating since 1990 owing to economic condi-
The widespread use of plasma-processing technology is lions. The processing of ferroalloy and stainless fines is
just beginning. In the field of pyrollletallurgy, il is quite established technology.
TABLE II
INDUSTRIAL PLASMA INSTALLATIONS FOR THE PRODUC.TION OF METALS AND ALLOYS

Process Feed Pnxluct System Power cilpilcity Stan-up Plant location Reference
Plasma cnlcible Crore Feer ASEAlMintek 14MW t983 Middelhurg Steel 15
smelting Crushed nnes (Upgraded from & Alloys.
141030 MW) t989 Krugenodllrp. SA 18
C, Feer 12MW 1997 South Carolina Research 22
concentrates IMacAlloy plant] AulhorilY. Charleston. S.c.
Plasma blast fumace Mn ore FcMn Acrosp:uiale 3 x 1.5 MW t984 SFPO. Buulugne-sur-Mer 27
smelling 8 x 1.5 MW 1986
Crorc FeCr SKF Plasm3- 4x6MW 1988 SwedeChromc. Malmo 24.25
Cmshcd fines chrome (IWO shafls) (molhballcd Sweden
in 1990)
Iron ore Fe Acro.<ipatialc 6 x 2.0MW 1987 LorfOlllc (HFRSU), 74
Lorraine
Reduction/melting Fcrmalloy FeMn FrcitaUVocsl 3MW 1983 Samancor. 30
crushed fines Alpine (Mimck) Meycnon.S.A.
Bughouse C,. i.Mo. SKF Plasma- 3,6MW 1986 ScanDusl. Landskrona. 48
fines Fe Dust Sweden
Crushed fine~ FeCrcharg Tetronics R&D 15MW 1988 Fcrbasa, Brazil 19
prm.'css slag fcec..l
AOD/EAF dust Cr.Ni. Mo Tetronics R&D 3MW t989 British Steel. 49a
ZnO Sheffield
EAF dUSf Zn. Pb,Cd Tctronics R&D 2MW 1989 Florida Steel, 66.67
Jackson. TN
EAFdust Zn. Pb. Cd Terlol1ics R&D 3MW t989 Nucor·Yal1lato, 66.67
Cr.l.wfords\lille. AR
Stainless dust Cr. Ni Tctrnnics R&D 7MW 199t Multiser\'e. haly 49b
EAF dust Cr. Ni. ZnQ I-Iollow-grnphitc IMW 1989 MinIck. S./\. 71
c;llhode (pilot)
EAF dllst Cr. Ni Hollow-graphite I.5MW 1990 Chiba Works. Kawasaki 50
l,:.llhodc (pilot)
AutOlnobi1e- PI Tctrtlllics R&D JMW 1984 Multi-Meteo 81
emissiun catalysl.~ 75 000 tr.oz.ly Anniston. AL

PLASMA-ARC TECHNOLOGY FOR FERROALLOYS 23

 The melting and refining of alloy scrap, particularly tita- IArc Column
nium scrap and sponge, is a very promising growth area for
plasma processing. As noted in Table II, the simple plasma
beating of continuous-casting tundish melts to control the
temperature at which the steel is cast has shown significant
growth over the past five years, and promises to become
standard practice. Industrial hy-product resource, recovery,
and recycle, and the treatment of hazardous materials, have Molten Film of
spawned many plasma-based processes for which the ultra- Plasma Jet Vanadium O.r.ide t
high temperature capability, the gas environmental control, Carbon and V· C
and the ability to treat fines are unique and attractive pro- Crucible
Roof
cessing features.
in addition to these industrial installations, there are sev- \
eral plasma-based metallurgical processes now under devel-
opment including the following:
Reference
• Magnesium production
- MintekiSamancor, S.A.; 2 to 3 MW (demo) 54
- BHliton Research, 600 kVA (Cameron and
co-workers) 55,56
Relnelrillg/smelring of silicOll~l1Jetalfines
- Mintek 36
- SlNTEF 37(a), (b)
- Dow Chemical 35
- Kawasaki Steel 38 Crucible
- - Bottom
Meiring and recovelY of aLuminium metaL
FIGURE 2. Falling-film 1 MW pl11S11111 rew:.:LOr for the production of ferro-
- Hydro Quebec; PECfAlcan 63,64 vanadium 43 (a non-transferred heater operated with the arc transferred to
- Center for Materials Production, EPRI 65 the falling-film)
Recovery of zinc from residual dusts
- Plazmet (Houston, Texas) 70 of the gas phase. In comparison, for the reduction of vana-
dium oxide hy solid carhon, the function of the gas phase
Plasma Reactors for Ferroalloy Production was to stabilize the arc and transfer the energy for the
The industrial potential of plasma-arc reactor systems for reduction reaction. The carbon reductant was included in
the production of ferroalloys was established 3 with the the solids feed, and there was no stoichiometric relationship
development during the early 1970s of a I MW facility for between the gas and the vanadium oxide as there was in the
the carbothennic reduction of vanadium oxides, which pro- reduction of iron oxide4 .
duced standard-grade ferrovanadium 43 . This pilot-plant The development of the falling-film reactor for carhother-
facility operating at 0,5 MW produced ferrovanadium at a mic reduction wao;; extended to the production of ferroniobi-
projected rate of I million pounds a year. This was suffi- urn, ferrochromium, ferromolybdenum, and ferroboron, the
cient to supply 10 per cent of the US consumption, or all of processing of ilmenite ores, and the carbon reduction of
Bethlehem Steel's ferrovanadium requirements for the pro- iron oxide concentrates. In that programme, from 1967 to
duction of vanadium-alloyed steels. Bethlehem Steel had 1981, it was apparent that there were no 'plasma effects'
just developed a process for the extraction of vanadium that could be attributed to highly excited gas species, but
oxides from a South American source of iron-ore peUets. that the reactions followed conventional thermodynamic
The plasma-ferrovanadium process provided the potential analyses. The uniqueness of the reactor was as a source of
for fully integrated vanadium sourcing from the are through high temperatures and high heat fluxes. In this context, it is,
to the steel product. in fact, more suitable to refer to these types of systems as
The design of the plasma-ferrovanadium system included 'electric arc reactors' rather than 'plasma reactors'.
the 'falling film' plasma reactor, as shown in Figure 2. This The gas-stabilized non-transferred arc has inherent eco-
reactor had been developed at I MW for producing steel nomic disadvantages associated with the cost of the feed
direct in a 'one-step' process by the reduction of iron-ore gas, the increased sensible heal lost to the off-gas, and the
concentrates with mixtures of hydrogen and natural gas. increased cost of haodling and cleaning a larger volume of
The product72 was pure iron (0,006 per cent C, 0,06 per exhaust gas than just the product gas of the reaction. The
cent Si, 0,005 per cent S, 0,001 per cent P, 0,007 per cenl ferrovanadium development used argon as the arc-stabiliz-
eu). The economic incentives for this process were the low ing gas. In further developments. carbon monoxide, as the
cost of natural gas at that time and the anticipated decline in process product gas, was demonstrated to he a suitahle
the costs of electrical energy. TIle 'falling-film' plasma working gas44 . A plant installation for the recycling of car-
reactor was unique in that the basic design was a non-trans- bon monoxide product gas would require a cleaning, stor-
ferred arc heater in which the hydrogen and natural gases age, and compression loop, which increases the gas cost,
had the multiple function of stabilizing the arc as well as particularly for a smaU facility. The use of the non-trans-
heating and reducing the iron oxides. In addition, the gases ferred arc instead of a transferred arc results in higher heat-
conveyed the iron oxide into the arc environment. The ing losses to the cooling water, and therefore a less efficient
falling-film concept allowed the residence time of the iron use of the input power. For example, the efficiency of
oxide in the reactor to be independent of the residence time water-cooled torches is about 80 per cent for non-trans-

24 INFACON6
[erred-arc operations as compared with 95 per cent for
transferred-arc operations. The efficiency of transferred-arc,
graphite electrodes that are not water-cooled and have mini-
mal heat losses is nearly 100 per cent. It was also apparent
thal. although the reduction of iron oxide with hydrogen
and natural gas dictated a non-transferred design for the
heating of the reactant gases, a transferred arc to a bulk
melt could also be used for carbothennic reactions 4o .
In 1979, ASEA announced the ELRED process for the
production of iron using coal of non-metallurgical
grade 73a .b . The iron oxide fines were prereduced in a fluid-
ized bed. Final smelting reduction occurred in a d.c. furnace
with the arc transferred from a hollow graphite electrode to
the iron melt, which had a bottom anode connection. The
prereduced concentrate and char were fed through the hol-
low electrode to the arc/melt region, as shown in Figure
3 73c . The scale-up potential and relative simpljcity of the
ELRED process using a transferred-arc furnace had
eclipsed Bethlehem's programme to develop a plasma-
steelmaking process using a solid carbon reductant.
FIGURE 4. TClronics: Research & Development (TRD) sweeping-arc
Fin. chromlte feed plasma furnace at J 103 MW with gravity/roof particulate feeding for the
melting or fcrrochromium fincs l9 lllltilundish hcutill£K8.90

Feed
,
• Rotary
Cathoda Preheater
Hollow
a'.elrod.
1 rCOlumn
Flua g ....
~

Plasma
Zone
Anodlt Gas
FIGURE 3. ASEA hollow graphite -electrode reactor developed by Mintck
and MS&A up to 14 MW and 30 MW for the produclion or
= "-Inject ion

fcrrochrollliumi ~.1 K.7.k

At that time, the implications 01' the ELRED development

for the plasma production of ferroalloys were discllssed' FIGURE 5. EAFR (exlended arc-flash reactor) horizontal graphite elec-
with representatives from the National Institute for trodes with a non-transferred arc used at 100 kYA for fcrrochromium
trials 14
Metallurgy, now Minrek. By July 1979, Mintek had already
initiated programmes with MS&A for the plasma reduction tant plasma ferroalloy application and would require a fur-
of chromite ores using the transferred-arc system of nace rated at 20 10 40 MW.
Tetronics Research & Development (TRD) with its charac- The multi-megawatt scale-up of the Bethlehem plasma
teristic precessing cathode assembly, as shown in reactor, which was projected to be a lengthy and costly
Figure 4'3 Mintek also installed a 100 kYA non-transferred development, was not undertaken. A d.c. furnace, the 'plas-
a.c. furnace based on the extended arc-flash reactor l4 , rna reactor', had been operated by ELRED at Avesta in
shown in Figure 5, which had dCll10nstraded the feasibility Sweden at the 10 MW level without solids feed, and scale-
of reducing chromite-ore fines in an open-bath process 12a ,b. up to 50 MW was planned although it was never commer-
Following extensive pilot-plant test work at Mintek support- cialized. Solids feeding was demonstrated at the metallurgi-
ed by MS&A, it was announced" in December 1982 that a cal research station, MEFOS, in Lulea, Sweden, with a 3 to
plasma smelter developed by Mintek using the ASEA 4 MW hollow electrode furnace. ASEA has continued to
ELRED hollow graphite cathode principle was to be develop and market the d.c. furnace for scrap melting. Such
installed at MS&A with a power level of 12 to 14 MW for scrap melters are now being installed up to the 80 MW
(he production of ferrochromium. The upgrade 30 in 1989 by level (UNARC furnaces supplied by MAN GHH)6
MS&A from 12 to 30 MW C1ccompLished the projection 3 Commercial installations of plasma systems for the produc-
made by Minlek in 1977 that the production of ferro- tion of ferroalloys also include the SwedeChrome plant
chromium from friable chromite ores was the most impor- based on SKF gas heaters, shown in Figure 624 ,25, and the

PLASMA-ARC TECHNOLOGY FOR FERROALLOYS 25

 Industrial production using non-transferred arc technology
beyond several megawatts requires the use of multiple
torches, as shown by the SKF SwedeChrome plant24 and
the SFPO ferromanganese blast fumace 26
Aside from the flexibility that scale-up capability pro-
vides, the two outstanding advantages of the hollow
graphite electrode system are the absence of cooling water
and the low tlowrates of arc ga.'i. These design advantages
result in lower investment costs for the auxiliary systems,
both for water supply and discharge, and for the cleaning
and handling of the exhaust gas. The absence of water-cool-
ing is a particularly important safety advantage since any
possible water-melt explosion risk is eliminated. The ener-
RGURE 6. SKF 'SwedeChrome' shaft furnace using four 6 MW non-
gy-efficiency advantage of the hollow graphite electrode
[rnnsferred gas heaters with tuyere injection of solids24. 2$ system is obvious since there are no energy losses due to
the sensible heating of water and no losses of excess operat-
ing gas. This is of particular importance with the use of a
high-grade energy source such as electricity. As noted in
Figure 8 for the SKF process. the energy utilization based
on the ferrochromium and gas products is about 67 per cent.
In comparison, for the Middelburg 30 MW operation, the
energy utilization is about 85 per cent.

SWEDECHROME

r--- ./
Ferro- Ferro-
chromium chromium
Electric
FIGURE 7. Voest-Alpine (Freital) transferred-arc furnace using W;l\er-
32 % Power 45 %
cooled tungsten-cathode arc heaters for the melting of ferromanganese
fines 30 ""- 53 %
Samancor Metalloys ferromanganese plantJO based on the
Voest-Alpine transferred-arc design, shown in Figure 7 but Gas
....-
with a single vertical cathode. 35 %
The technical feasibility of the plasma carbothennic Gas
reduction of oxides to produce ferroalloys was demonstrat- Coal & 40%
ed by all the reactors sbown in Figures 2 to 7. The choice is "'-;;ot 13 % Coke
basically between operation with a transferred or non- Water
transferred arc. and whether a water-cooled electrode and 47 %
heater design is required. The primary consideration for the V Losses Ho,
selection of a plasma (or for that maner) reactor system for Losses ~ Water 3%
20 %
pilot-plant development is 'can the reactor be scaled-up for
an industrial operation?'. For the commercial production of
ferroalloys, the most successful plasma system has been
r
12 %
-...
FIGURE 8. Energy flow diagrams: SwedeChromc and MS&A plants for
demonstrated to use the basic, unsophisticated design, as
lhe produclion or rerrochromium
shown in Figure 3, of a vertical non-water-caoled, hollow Althougb industrial-worthy water-cooled heaters, both
graphite electrode (HOE) with a d.c. arc transferred to a transferred and non-transferred, have operated reliably. car-
conductive melt. This ultimate design of a plasma ferro- bon electrodes are Jess susceptible to damage resulting from
alloy reactor as demonstrated by the MS&A 30 MW facility operating conditions and rough handling during installation,
follows the 'Keep It Simple' principle. both of which can cause water leaks, electrical faults, and
Scale-up to the multi-megawatt range is the unique fea- short electrode lifetimes. This is particularly apparent in
ture of the hollow carbon d.c. electrode, particularly for tbe larger-scale operations with dusty surroundings producing
production of ferrochromium. The single carbon electrode tonnage products, in comparison wil)] small power-level
can sustain d.c. currents at lea.'it up to 100000 A. In com- torches, which are often used in ultra-clean environments
parison, the water-cooled tungsten electrodes used in the typical of the production of Jow-vo1ume specialty metal~.
Voest-Alpine and TRD systems operate at 5000 A with a
reasonable lifetime of about ISO hours and with an upper
limit of about 10000 A. For these transferred-arc systems, Recent Developments in Plasma-ferroalloy
the power levels are therefore limited to about 4 to 5 MW, Production
corresponding to operating voltages of 400 V. The SKF The plasma production of ferroalloys is a well-established
non-transferred gas heaters with water-cooled copper elec- technical fact. Some commercially driven changes have
trodes for the SwedeCbrome plant are rated al 6 to 7 MW. been noted with the expansion of the MS&A ferrochromi-

26 INFACON 6
um plant and the closing of the SwedeChrome facility. The MW (40 MVA) at a cost of $10 million'8 The 40 MVA
technology is maturing. and there is a certain allitude of plasma furnace installed at Krugersdorp replaced the 16
confidence as expressed by the plans to install a 12 MW MY A d.c. pla'ima-arc furnace. The prevailing market
ferrochromium plant in South Carolina22 • There have been demand for ultra-low silicon and phosphorus alloy and for
a number of speculative development projects. such as the low-silicon alloy. and the abundant availability of low-cost
Voest-AlpinelDow Corning pilot-plant efforts to produce feed tines and the increased thernHlI efficiency with scale-
silicon and ferrosilicon J3 • and the Davy McKee 5 MW ferro- up. were economic incentives for this further expansion.
manganese remclter11 . These efforts are to be encouraged The piam was recently bought by Samancor and is now
since the technology is still young and advances can still be called Palmiet Ferrochrome Division.
made, particularly in the understanding of how best to opti-
mize the reaclion systems and so improve the yield and Mannesmann Demag, Dalmacija Plant, Dugi Rat,
decrease the net consumption of energy. Yugoslavia
Chromium-ore fines of up to 28 per cent of the feed mix of
Ferrochromium coke, quartz, and are were fed in a test trial through a hol-
The capability for the processing of unagglomerated fines low electrode of a 20 MW submerged-arc a.c. ferrochromi-
(ore. fluxes. and reductants) is the primary economic urn furnace. The test results were encouraging in that the
advumage of plasma reactors for the production of ferro- chromium recovery increased. there was a high smelting
chromium. The charge can be fine chromium ore, concen- rate due to faster smelting in the are, and a means was
trates from friable chromite ore. or crushed ferrochromium established for the recycling of baghouse dust and ferro-
fines. In particular, where the chromium content is upgrad- chromium reverts 2C1 • No further trials have been reported.
ed by ore crushing and further beneficiation, the tine con- . Strictly speaking. this development in a submerged-arc
centrates that are produced are very suitable as feed for a furnace should perhaps not be included in a discussion of
plasma reactor. The fines feed rate must be matched to the plasma reactors; that is. there was no discussion of the arc
power level. High feed rates with respect to the input power characteristics or reactjon-zone kinetics that would have
will result in the accumulation of unreacled feed. Al low been influenced by the feed of solids or gases through the
feed rates, excess cnergy is available, and the temperature arc region. From this standpoint. the dcsignation of an elec-
will increase, which can result in furnace damage. undesir- tric-arc process as a plasma-arc process becomes purely
able reactions, and high requirements of unit energy. .~llbjective. It is quite reasonable, though, to regard this
work as an outgrowth of the MS&A d.e. hollow graphite
Middelburg Steel & Alloys (MS&A). Krugersdorp. electrode open-bath process for the production of ferro-
South Africa chromium.
The commercially successful MS&A 12 to 14 MW plant
South Carolina Research Authority, Charleston, S. C.
has been expanded to 30 MW. The 12 MW plant ASEA
d.c. hollow graphite electrode furnace was introduced in The hollow carbon electrode plasma furnace is now being
December 1983. and replaced an existing 9 MW sub- developed for domestic raw materials in the USA. In 1989.
merged-arc furnacc. The process was initially developed at thc Strategic Materials Office of the U.S. Defense Logistics
Mintek on a I MW scale, and required further improve- Agency funded the development 21 of a 1,5 MW plasma fur-
ments for industrial scale_ up IS.16. It is common in the devel- n.ace for the production of ferrochromium 21 . The facility is
opmcnt of plasma furnaces thai major difficulties are ex- slled at the MacAlloy plant in Charleston. S. C. The inten-
perienced in the conveying and feed-rate control of the tion of the programme is to establish a domestic fer-
fines, and less on electrical problems such as unstable or rochromium capability using US chromium-ore sources.
stray arcs. An extensive analysis of the sensitivity of the Low-grade US orcs can be upgraded by crushing and bene-
balance of feed rate to power as related to the arc-ultach- ficiation to produce a 40 per cent chromium oxide concen-
ment zone has been made by Barcza, CUff, and Jones". trate. The processing of these concentrates in conventional
Improvements on the MS&A 12 MW plant included beller submerged-arc furnaces would require costly agglomeration
control over the feed rate to closely match the power level, of the fines. The pilot-trials at 1.5 MW successfully COI11-
and monitoring of back-pressure in the electrodes and selec- pleted at the end of 1991" have provided the design basis
tion of the approprime diameter to minimi7..e electrode f?r a larger demonstration plant. A 12 MW facility to be
blockages. Changes were also made in the design .md opera- Sited at MacAlloy is being considered.
tion of the gas-cleaning system and the off-gas ducts. A recent evaluation concluded that Bird River chromite
Modifications to the slag chemistry also helped to decrease ore would be exceptionally well suiled to reduction in a
the carryover of dust in the exhaust. The chromium recovery plasma fumace 23 • It is apparent that the existence of plas-
was 90 to 95 per cent. The consumption of energy was ma-furnace technology for the processing of are fines will
sliglllly higher than for a submerged-arc furnace but, with stimulate reconsiderarion of other chromite-ore reserves for
scale-up, the consumption of energy would be similar to that the production of ferrochromium.
for the submerged arc furnace. Other developmental work
has included the production of a high-chromium (75 to 80 SwcdeChrome, Malmo, Sweden
per cent), high-carbon ferroalloy to fann a superchrornium In glaring contrast to the 1988 expansion of the MS&A
slag (Cr:Fe 10 to 20: I). This slag was produced in the plas- plant. the 48 MW SwedeChrome planl that started up in
ma furnace by selective reduction or iron out of chromitc are October 1987 was shut down in 19H1J, and reportedly is for
to produce a semi-stainless alloy? (20 per cent chromium). sale. Although the ferrochromium market may have been
At the end of 1988 after a five-year developmenHlI peri- the decisive factor, it is recognized that the process technol-
od. tile 12 MW (16 MVA) operation was upgraded to 30 ogy and site conditions are not comparable with those of

PLASMA-ARC TECHNOLOGY FOR FERROALLOYS 27

the MS&A plant. The SwedeChrome plant is based on SKF the gas phase. A recent analysis28 emphasized the need to
non-transferred plasma-gas heaters. Chromium-ore fines, recycle volatilized manganese vapours, e.g. with a molten-
along with a blend of lime, silica, and coal, are fed into four slag filter, and that existing transferred-arc reactors, how-
tuyere zones in lhe front of the plasma heaters at the bottom ever capable of reducing manganese ores. must be designed
of a coke-filled shaft. as shown in Figure 625 • The for condensation. Two suggested concepts are
SwedeChrome plant has a hybrid energy input of electrici- (I) the injection of manganese and coal fines into the tuyere
ty. coal, and coke. producing not only ferrochromium and level of a shaft furnace, which could be done in shafts of
hot water for district heating, but large quantities of fuel SFPO and SKF design equipped with non-transferred
gas. Based on the energy flow diagram (Figure 8), the heaters;
heater efficiency is about 80 per cent, which is not uncom- (2) a design of a non-transferred arc-gas heater with a
mon for a non-transferred arc design. This represents a pro- falling film of manganese-rich slag that is eventually
duction of about 5 MW of relatively low-grade hot water. reduced to produce manganese, the manganese product
The process economics requires a use for this excess ener- being centrifuged to the inner layer on the reactor wall
gy. A thermochemical model 26 indicated lhm the process and, presumably, being protected from vapour loss by
was flexible, permitting many process variations that the outer slag/coke layer.
allowed the optimal selection of feed materials.
Voest-Alpine: Two-stage Smelter/Shaft Reactor
Ferromanganese A burden-condensation furnace, shown in Figure 9, was
The high volatility of manganese (e.g., 174 mm Hg, 1762 0c) proposed by Voest-Alpine29 In experiments with an open-
is a critical factor in the design of a furnace or process for bath transferred-arc electrode, the manganese losses to the
the production of ferromanganese. In the conventional pro- gas phase were 40 to 50 per cent. In the new design, the
cesses, the ferromanganese blast furnace and the sub- charge is prereduced in the shaft and then descends into the
merged-arc furnace, the manganese vapours are captured by smelting zone. A water-cooled transferred-arc heater is
condensation on the burden charge of are and coke. equipped with oxygen ports for final smelting. The oxygen
is provided to optimize the decarburization, and to control
SFPO, Boulogne sur Mer, France the melt temperature and therefore produce a medium-car-
bon ferromanganese. Ore fines could be injected into the
The Societe du Ferromanganese de Paris-Outreau (SFPO)
smelting zone. It would appear lhat a hollow graphite-elec-
has three ferromanganese blast furnaces with a total capaci-
trode system could also be used in this reactor, and any
ty of 400 kt per year27 The highly endothermic carbon
safety problems associated with potential torch failure and
reduction of M.nO occurs at temperatures above 1250 °C in
water leak explosions could be avoided. The initial objec-
the lower part of the furnace, and results in a high con-
tive for this reactor is to achieve a power level of 2,5 MW
sumption of coke with the production of an off-gas rich in
at a production rate of 800 kg per hour. No data are yet
carbon monoxide. Only one-third of this off-gas can be uti-
available on this operation.
lized in the hot-blast stoves; the remainder is used for the
generation of electricity, which is a valuable byproduct dur-
ing the winter when the electricity demand is high. During
the rest of the year, the seasonal demand is low, and the rel- CO,Ar
ative value of electricity is also low. In 1984, SFPO
installed three 1.5 MW Aerospatiale plasma-gas beaters to
ulilize the generated electricity on site during periods of SiD· 2C=SiC' CO
low demand to further superheat the hot blast. For every
100°C of additional blast temperature, the calculated coke
savings were 52 kg per tonne of ferromanganese. The initial
trials were successful, and in 1986 five additional plasma
beaters were installed for a total of 8 of the 9 tuyeres having
a blast superheated by plasma heaters. The blast tempera-
ture has been increased from between 1150 and 1200 °C up
to 1500 °C, with coke savings of about 170 kg per tonne of
ferromanganese; the replacement rate, at 500 kWh per
tonne of ferromanganese. is 3 kWh per kilogram of coke.
Substantial improvements were required in the design of
the plasmaJhot-blast tuyeres and adjacent refractories to
accommodate the higher blast temperatures. Additional
design improvements to the plasma heaters resulted in an
increase of heater efficiency from 62 per cent in 1984 to 72
per cent by 1989. The lower coke rates, i.e. less shaft vol-
ume occupied by coke, resulted in an increase of 20 per
cent in productivity and also a smoother operation with bet-
ter furnace control.
This successful application of plasma technology to the
ferromanganese blast furnace has not been repeated for FIGURE 9. Voest-Alpine 'shart-hearth' 100 kW plasma furnace using
electric-furnace hearth reactors. It is evident that, for an water-cooled tungsten-cathode arc healers for the production of ferro-
open-bath plasma system, there will be manganese losses to silicon and silicon metn]J3

28 INFACON6
Remelting of Ferromanganese-metal Fines con reaction sequence involves the countercurrent flow of
The plasma remelting of ferromanganese flnes started in solid carbon and silicon dioxide, and gaseous CO and SiO.
June 1983 at the Samancor MetaJloys plant'O The trans- Silicon carbide, formed by the reaction of the carbon and
ferred-arc heater is hmited in voltage capabihty owing to SiO gases, reacts with silicon dioxide through an SiO inter-
the conductive manganese vapour. At the full current capa- mediate reaction to form silicon metal. SiO side reactions
bility of 10000 A for the tungsten cathode, the power input form excess silicon carbide and also dissociate to form
is only 3 to 4 MW. deposits of silicon and silicon dioxide, which tend to plug
Davy McKee has described a 5 MW plasma fumace 3l the furnace. In conventional submerged-arc furnaces, it is
based on the operation of a 350 kW pilot plant for the melt- necessary to physically break up the charge to aJlow the
ing of ferromanganese fines. The proprietary design fea- furnaces' gases to escape. As a result, it is necessary to use
tures a cylindrical sleeve that surrounds a d.c. transferred a burden of carefully sized quartz and coke to optimize the
arc I m in length. The fines are fed onto the sleeve, where permeability of the burden. A reactor system capable of
melting occurs, and the melted material then falls off onto processing fines, in this case inexpensive silicon dioxide as
the bulk melt. The 5 MW heater was commissioned for sand, is an attractive application for plasma reactors.
Plasma Arc Limited, Melbourne, Australia. No production
data for this operation have been published. Hollow Graphite Electrode Plasma Reactor
The current limitations of the Voest-Alpine and Davy The capability of producing ferrosilicon from a charge of
McKee plasma heaters (both with water-cooled tungsten silica and carbon fed through a hollow graphite electrode
electrodes), the complexity of the sleeve reactor, and the was demonstrated on a 2 MYA furnace operating continu-
potential for water leaks suggests that the basic hollow ously over a three-week period3o , A decrease in the total
graphite electrode should be used for the melting of ferro- energy required was attributed to the 100 per cent conver-
manganese fines. sion of the silica fed through the hollow electrode, which
was 20 per cent of the charge. The remainder of the charge
Ferrosilicon and Silicon was a conventional burden.
Each ferroalloy system has its characteristic complexity: the
chromite system requires high temperatures, high heat flux- Voest-AlpineIDow Corning Project
es, and a complex slag chemistry; manganese alloys have a An extensive development of a plasma-reactor design, simi-
vapour pressure highly dependent on the alloy component lar to Figure 9, by Voest-Alpine and Dow Coming J3 •34
and concentration; the ferrosilicon reaction system, demonstrated the production of silicon metal from quartz
Si-O-C-Fe, has complex low- and high-temperature re- and sand, and ferrosilicon from taconite fines, at 100 kW.
action sequences involving gaseous intermediates. The sili- The reactor operated with the production of silicon carbide

,
I' ,.i
I t
Step 3

SIO(gl +2C ($) =$iC(s, +CO (g.)

$iC;S) 20c0·C SIO,o> +CO,o' 2llllII "c

Step 2
N!'t eneorgy
~npu'tl
SiC (I;) +SiO (g)::=2SJ (I) +CO (0) 1. 7klJh/kg SI
pr-oaucPD:

2'S'm 2QOQ·C 2SIO,O) 2OllIl·C

s) ~·c

I S~ep 1
~Inpu'b
Neot: I!'nergy

7 .3klJh/kg SI
Si0 2 ell +51 m ~2SIO(g)
progluCPd

FIGURE 10. NTHlSlNTEF reaction sequence and reactor concept for the production of silicon metal-hearth heating with a water-cooled lungsten-cathode
transferred arc, and shaft heating with a graphite electrode38

PLASMA-ARC TECHNOLOGY FOR FERROALLOYS 29

by the reaction of coke with SiO gas in the shaft. The sili-
con carbide was periodically pushed down into the plasma Gas outlet
smelting zone to react with quartz to produce silicon metal / U\r.CO.SiOl
and to regenerate the SiG gas. The torch reliability and life
were found to be the prime operational limitations during
these trials, which used a Voest-Alpine water-cooled trans- Carbon ~_A"'"Stock Iin~
ferred-arc torch. The scale-up of this process would require heater
a practical design for the continuous transfer of silicon car- Heat insulatO(
bide from the shaft into the smelter zone.
Since 1988, Dow Coming has been developing a single-
electrode, 200 kVA submerged-arc furnace to operate under
totally sealed conditions using quartz with charcoal and
coal reductants35 . This operation is now being scaled up to Silica'
~~~~.B
~ Arc
6MW. injection electrode
noz.zle :!;l"''"-II--_ A /
--._~

./' !~~~±~~~}Hearth
Mintek: Melting and Refining of Silicon Fines
Silicon-metal fines have been successfully remelted in a Tapping
lOO kV A transferred-arc plasma furnace under an argon hole
atmosphere36• Remelting resulted in upgraded quality from
0,45 per cent AI and 0,17 per cent Ca to 0,19 per cent AI
and 0,0 I per cent Ca.

NTWSINTEF: Silicon-fines Melter and Silicon-

reduction Reactor
The Norwegian Plasma Technology Group of NTH/SIN-
TEF has proposed a three-step plasma reactor" (Figure 10)
to produce silicon metal from quartz and coke. The lower
part of the reactor has been operated under argon to remelt
silicon-metal fines 3? The reactor has three water-cooled, FIGURE 11. Kawasaki steel reactor for the production of silicon metal-
hearth heating with alSO kV A non-transferred grnphitc-electrode a.c. arc.
tungsten-cathode heaters with a total power capability of
and shaft heating with 180 kVA carbon resistance elements 38
200 kW. In silicon smelting, silicon carbide forms at the top
of the reactor by the exothermic reaction of SiO gas with
the carbon charge. The upper reactor is to be heated by a
graphite electrode, where the silicon carbide plus SiO gas
form molten silicon. Silica is fed to the bottom reactor to
reacl with the silicon melal to form the SiO gas. The bottom _Graphite
reactor will require about five times the energy input as the 11 -- Cot hode
upper reactor. No information on the progress of this re-
aCWr is avaiJable.

Kawasaki Sleel: High-purity Silicon by Carbolhermie

Reduction of Silica
The reactor-reaction scheme outlined by NTHJSLNTEF was
also followed in the Kawasaki Steel l50 kVA arc furnace",
shown in Figure II. Pure silica was injected with argon
through silica-injection tubes into the arc space between the
single-phase a.c. graphite electrodes, as shown in the cross-
section view of the hearth region. It was possible to produce
high-purity silicon al 2 kg per hour with a yield of 83 per
cent. Afler decarburization, the silicon was of solar grade.

Other Ferroalloy Studies: FeMo, FeNb,

Carbides
The promise of plasma technology 10 lead to new process-
ing: routes and improved reactor concepts, and the use of
alternative charge materials, continue to foster developmen-
tal efforts and feasibility studies for ferroalloys. Recent ---- Water-cooled
developments have included the use of a 30 kW falling-film AnOde
sleeve reactor to produce FeNb~5 and NbC 46 . Fly ash has
been used as a charge material to produce ferrosilicon 39 .
Feasibility studies continue to be made on the production of FIGURE 12. Siemens vertical open-bath graphile-electrode d.c.
ferromolybdenum from molybdenum disulphide 41.42. tmnsferred·arc funHlce from 18781 J

30 INFACON6
 TABLE III
INDUSTRIAL PLASMA INSTALLATIONS FOR THE MELTING OF METALS AND ALLOYS

Process Feed Product System Power capacity Start-up Plant location Reference
Meltinglrefining Alloy scrap Alloy steel Freitnl 20 MW. 30-ton melt 1973 VEB,8 May 1945, Freital, 51
10 MW, 10 ton 1982
Mannesman 3 MW, 100tOil melt 1986 FRG 52
Demag (Krupp) 30-ton ladle
-
FrehaJ/ 4x3MW,45ton 1983 Voest-Alpine, Linz. 53
Voest-Alpine (decommissioned Austria
in 1987)
Tj scrap/ Ti ingot Retech 2 x 300 kW 1988 Prall & Whitney, 58
sponge East Hartford, CT
Retech 2,IMW 1988 Wyman-Gordon, 59
Worchester, MA
Retcch 2,25 MW 1989 Teledyne-Allvae. 60
Monroe. NC
Stuinless scrap Ingot slabs ULVAC () x 0,4 MW 1986 Nippon Stainless Steel 61
(4 l)
Ti scrap Ingot PEClLeybold 3 x 400 kW 1990 TiMet, Henderson, NV 62
Scrap consolidation Ti scrap / Ti electrodes Retech 0,8MW 1984 Oregon Metallurgical 57
sponge 14/22/28" dia Corp., Albany, OR
Plasma blast Foundry-iron Fe Westinghouse 6x 1,5MW 1989 General MOlars, Central 77
cupola melting chips, borings Foundry Div., Defiance,
Ohio
Aerosputiale 4MW 1989 Peugeot, Sept-Fons 78

TABLE IV
INDUSTRIAL PLASMA INSTALLATIONS FOR THE HEATING OF METALS AND ALLOYS

Process Feed Product System Power capacity Start-up Plant localion Reference
Ladle heating! Steel melt Heated melt Hollow graphite 16.8 MVA 1988 CPl Steel Corp .. 83
refining electrodes 3 phase a.c. Pucblo, Colorado
Fcrroal1oys FeSi SINTEF IMW 1982 Halla Smeltverk. Norway 85
Alloy stcels Mannesman 6MW 1989 Krupp Sicgen Steclworks 86
Carbon steels Dcrnag (KnipP) 3 phase a.e.
Low/med. cilrbon Advent Proccss Hollow graphite 1992 Maynard Steel Casting Co., 84
steels Engineering electrode, I MW, Milwaukee, Wisconsin
5 ton (dcmo)
Tundish heating Steel melt Heated melt Tetronics R&D I MW, 14-lon melt 1987 Nippon Steel, 88
Hirohata Works
Tetronics R&D 350 kW, 5 ton 1988 Aichi Steel 89
Tetronics R&D 1,1 MW. 40 Lon 1989 NKK Keihin no. 3 90
TeLronics R&D 1,1 MW, 40 ton 1989 NKK Keihin no. 4 90
TeLronics R&D 1,1 MW,40ton 1990 NKK Keihin no. I 89
Tetronies R&D 1,1 MW,50LOn 1990 NKK Keihin no. 5 89
Tetronics R&D 1,1 MW, 17ton 1990 NKK Fukuyama 89
Tetronics R&D 0,8 MW, 6 ton 1990 Anva! Nyby Powder AB 89
Mannesmann 2 MW, 20 LOn 1988 Deltasidar (ltaly), 91
Dcmag (Krupp) Aosta Works.
Manncsman 2,4 MW. 80 ton 1989 Kobe. Knkogawa Works 92
Dcmag (Krupp)
Plasma Energy 4 MW, 27 ton 1988 Chaparral. 93
Corporation Midlothian, Texas 94
Plasma Energy 1,5MW 1990 FirSL Miss (SLoney Creek) 95
Corporation Steel, Hollsopple, PA
Plasma reforming: Iron oxide Fe:<OlFe SKP Plasma- 3 x2MW 1981 Hofors, Sweden 79
(<.Ie<.:oml11i~·
Production of pellets/ore prereduced Red
sinned in
reduction gas product 19R~)

Iron oxide Fc:<O/Fe Huls USCO 1986 Union Steel Corp., S.A. 80
pellets/ore prereduced (mothballe<.l
in IlJHH)
prodUCI

PLASMA-ARC TECHNOLOGY FOR FERROALLOYS 31

 Final Word on Plasma Ferroalloy Reactors 6. Nucor's second thin slab CSP planl (80 MW UNARC
As the result of the last twenty years of development, a DC arc furnace). (1991). Steel Tillles, vol. 219, no. 10,
strong preference has been expressed for the use of hollow pp. 560, 562.
graphite electrodes for the plasma production of ferroalloys 7. Slatter, D., Barcza, N.A., Curr, T.R., Maske, KU., and
by carbothermic reduction of a particulate feed. An uncriti- McRae, L.B. (1986). Technology for the production of
cal view might equate this reactor system to submerged-arc new grades and types of ferroalloys using thermal plas-
systems with solid carbon electrodes or even to the furnace ma. INFACON 86, Fourth International Ferroalloys
setupJl described by Siemens in 1878, shown in Figure 12. Congress, Sao Paulo, Brazil. Mimek Rev. no. 6, 1987,
The development of these plasma reactors has required a pp.47-59.
critical fe-examination of the fundamentals of the reaction 8. Sommerville, l.D., and Kemeny, F.L. (1991/92). The
kinetics, mass and heat transfer, slag chemistry and gas- enhancement of refining and alloying in the ladle by
phase behaviour of ferroalloy production 7 , The develop- plasma heating using drilled electrodes. Steel
ment of plasma-reactor systems for the production of fen'o- Technology International 199//92, Sterling Publications
alloys is not complete, but will continue with the improved International, London. (In press).
understanding of these systems 8 . Inevitably, the plasma- 9. Orrling, B. (1986). DC arc single electrode smelting fur-
reactor concepts, most notably the feeding of fines through nace. Shenyang International Symposium on Smelling
hollow electrodes9. IO , will be included in submerged-arc
Reduction.
systems with the emergence of improved ferroalloy-produc-
tion technology. 10. Barcza, N.A., Curr, T.R., Denton, G.M., and Hayman,
D.A. (1989). Mintek's role in the development of plas-
ma arc technology based on a graphite cathode.
Conclusions Workshop 011 Industrial Applications, Proc., Boulos,
For the smelting of Ferroalloys, the past twenty years of M.l. (ed.). Pugnochiuso (Italy), pp. 115-126.
plasma-technology development has resulted ill relatively
II. (Siemens, W.) Stansfield, A. (1914). The electric filr-
few industrial installations. For the complex smelting of
ferroalloys, the most appropriate system is the least com-
nace. McGraw-Hill Book Co., New York, pp. 4-5.
plex, Le. the d.c. transferred-arc hollow carbon electrode
Ferrochromium
reactor.
As shown in Tables" to IV, applications of plasma tech- 12. (a) Pickles, CA., et 01. (1979). A new route to stainless
nology other than ferroalloy smelting grew significantly steel by the reduction of chromite ore fines in an extend-
during the 1980s. This growth occurred mainly for relative- ed arc l1ash reactor. TrailS. ISIF, vol 18, pp. 369-382.
ly simple systems such as the remelting of Il)etal and the (b) Pickles, CA. (1977). A new electric furnace process
control of melt temperature (shown in Tables I" and IV). for the production of fcrrochroll1iulTI. Ph.D. Thesis,
These systems, particularly for reactive metals, use argon University of Toronto.
atmospheres and water-cooled systems to minjmize trace 13. Barcza, N.A., Curr, T.R., Winship, W.D., and Heanley,
contamination of the metal. It now appears that the complex CP. (1981). The production of ferrochromium in a
heaters being used for melting and heating applications will transferred arc plasma furnace. 39th Electric Furnace
be the next candidates for redesign. Califerellce, Proc., pp. 243-260.
In conclusion, the quest for simplicity of torch design and
14.Curr, T.R. (1984). The installation of a 100 kVA a.c.
wider applications of plasma technology continues.
plasma furnace at Mintek, Randburg. Mintek, Report
M135, pp. 12.
References and Bibliography
15. (a) First plasma smeller in SouLh Africa will produce
General ferrochrome at Middelburg. Etlgineeriflg and Mining
Journal, December 1982. p. 75.
1. Mac Rae, D.R. (1987). Application of plasma technol-
ogy to ferroalloy processing, Part 1. Plasma technology (b) Chrome and ferrochrome - firmly in the hands of a
in metallurgical processing, Feinman, 1. (ed.). pp. 149- few. Metall. Bulletill MOlltltly, no. 175, Jul. 1985. pp.
161. 15-19.
2. Mac Rae, D.R. (1989). Plasma arc process systems, 16. Barcza, N.A. (1986). The development of large-scale
reactors and applications. Plasma Chemisfly and thermal plasma systems. l.S. Afr. /IISt. Mill. Metall., vol.
Plasma Processillg, vol. 9, no. I, pp. 85S-118S. 86, no. 8, pp. 317-333.
3. Hamblyn, S.M.L. (1977). Plasma technology and its 17. Barcza, N.A., Curr, T.R., and Jones, R.T. (1990).
application to extractive metallurgy. Minerals Science Metallurgy of open-bath plasma processes. Pure Appl.
and Engineering, vol. 9, no. 3, pp. 151-176. Mintek, Chem., vol. 62, no. 9, pp. 1761-1772.
Report No. 1985, April 14, 1977. 18. Barcza, N.A. (1989). World's largest transferred plasma
4. Mac Rae, D.R. (1974). Criteria for plasma processing in arc fcrrochromium furnacc. Milltek Bulletin. no. 20,
extractive metallurgy. Symposium: Commercial p. I.
Potential for Arc and Plasma Processes, American 19. Smclting and melting of ferroalloys in a plasma furnace.
Chemical Society, National Meeting, Atlantic City. (1988). (1,6 MW Tetronics R&D Furnace; FeCr crushed
5. Bethlehem Steel/National Institute for Metallurgy fines; Ferbasa, Brazil) Steel Times, vol. 216, no. 6,
(1977, 1979). Plasma technical visits: March 10, 1977, p.314.
J.B. See; May 29, 1979, P. Jochens; Dec. 14, 1979, N. 20. Rath, G., Vlajcic, T., Staubner, H., and Kunze, J. (1990,
Barcza, M. Rennie, and B. Stewart. 1991). The hollow electrode - an application to process

32 INFACON6
 chrome ore fines. 48th Electric Furnace Conference, Remelting of silicon metal fines. SINTEF, Div. of
Proc., vol. 48, pp. 267-270. lrollmaker alld Steelmaker, Metall., Plasma Techn. Group, Report 2/88. 24 pp.
vol 18,no.3,p. 11. (b) Bakken, J.A., and Jensen, R. (1990). Plasma
21. Ellio". J.F. (1989). Plasma smelling le51 progmm is research at the Norwegian Insilutc of Technology.
underway. Americall Metal.. Market, vol. 97, no. 100, Proceedings, Conference on Plasma for Industry and
Suppl. Ferroalloys, p. 29A. Environment, Oxford, lhe British National Committee
22. Larson, H.R., and Perkins, B.R. (1991). Smelting fine for Electroheal, Paper 3.1,8 pp.
chrome ores in a plasma arc furnace. 49th Electric 38. Sakaguchi, Y., Fukai, M., Aratani, F. Ishizaki, M.,
Furnace Conference, Toronto. Kawahara, T.. and Toshiyagawa, M. (1991). Production
23. Healy, G.W. (1988). Carbon redUClion of chromites in of high purity silicon by carbothennic reduction of silica
Bird River and other ores and concentrates at 1200- using AC-arc furnace with heated shaft. Tet,'iU-to-
1700 DC. Canadian Metallurgical Quanerly, vol. 27. Hagane, vol. 77, no. 10, pp. 1656-1663.
no. 4, pp. 281-285. 39. Pickles, C.A., McLean, A., Alcock, C.B., and Nikolic,
24. Wikander, A.B., Herlitz, H.G., and Santen, S.O. (1987). R.N. (1990). Plasma recovery of melal values from fly-
The first year of opemtion at SwedeChrome. 45th ash. Canadian Metallurgical Quarterly, vol. 29. no. 3,
pp. 193-200.
Electric Furnace COllferellce, Proc., vol. 45, pp.
225-231. Ferromolybdenum
25. Wikander, A.B., and Rosendahl, P. (1989). The 40. Mac Rae, D.R., and Gold, R.G. (1979). Arc control in
PlasmaChrome process at SwedeChrome AB. INFA-
plasma arc reactors. U.S. Patelll 4. 234.334, appl. 10
CON, New Orleans. Jan. 1979, acc. 18 Nov. 1980. 6 pp.
26. Srinivasan, N.S., Santen, S., and Stafransson, L.1. 41. Johnston, c., and Pickles, C.A. (1990). Study of the
(1987). PlasmaCl1rome process for ferrochrome produc- reaction of molybdenum disulphide with lime and car-
tion - thermochemical modelling and application 10 pro- bon at plasma temperatures. TrailS. InSTn Min. Metall..
cess data. Steel Research, vol. 58, no. 4, pp. 151-156. vol. 99, pp. CIOo--CI04.
Ferromanganese
42.0uyang, T., and Cao, Y. (1986). Experimenlal study of
smelting ferromolybdenulll by three phase A.C. plasma
27. UIE International Union for Electroheat (1988). SFPO arc. International Conference 011 Plasma Science and
plasma-Jired ferromanganese bla,\·t furnace VIE arc Technology. Beijillg, pp. 279-284.
plasmll processes, pp. 95-98.
28. Truffaut. E. (1990). Rational use of thermal plasmas for Fcrrovanadiurn
melaHic oxides smelting. Conference on Plasma for 43. Mac Rae, D.R., Gold, R.G., Sandall, W.R., and
Industry and Environment, Oxford. The Brilish National Thompson, C.D. (1976). Ferrovanadium production by
Committee for Electroheat. Paper 1.2, 7 pp. plasma carbothennic reduction of vanadium oxide. 34th
29. Mueller. H.G.. Malzawrakos, P., and AubergeI', H. Electric Furnace COllferellce. Proc., pp. 96-100.
Smelting of ferromanganese for oxidic ores - a new 44. Mac Rae, D.R., Gold R.G., Sandall, W.R., and
way. Ibid., Paper 4.5, 10 pp. Thompson, C.D. (1976). Melhod of producing fer-
30. Samancor's chrome showing strength. (1983). Metal rovanadium. U.S. Patelll 4. 099, 958, appl. 9 Apr. 1976.
Bulletill, no. 6798, June 24, p. 15. acc. I I J ul. 1978, 6 pp.
31. Kershaw, P.. Moss, OJ., Gill, M.E., and Heggie, J.G.
(1989). A 5 MW plasma arc furnace for the production Fcrroniobium
of ferro-alloys. Proceedings, Third European Electric 45. Hilborn, M.M., Munz, RJ., Berk, D., and Drouet, M.G.
Steel Congress. Boumemouth. pp.340-344. (1989). Evaluation of ferroniobiulll carbide production
in a laboratory-scale Plasmacan furnace. Canadian
Ferrosilicon Metallllrgiclli Quarterly, vol. 28, no. 2, pp. 101-107.
32. Orrling, B. (1988). Correspondence, Pilot plant trials for 46. Munz. RJ., and Chin. EJ. (1991). The carbothermic
production of silicon metal and ferrosilicon in the reduction of niobium pentoxide and pyrochlore in a
(ASEA) direct currenl furnace, 32 pp. transferred arc argon plasma. Canadian Metallurgical
33. Dosaj. V.D., May, J.B., Raucholz, A.W., Muller, H.G.. Quarterly, vol. 3D, no., I, pp. 21-29.
and Koch, E. (1989). Pla,ma smelting of silica from Cr, Ni, Mo (Stainless-sleel Recycle Dusts)
sand and ferrosilicon from taconite. 47th Electric
Furnace Conference. Proceedings, vol. 47, pp. 43-48. 47. Herlitz. H.G .. Johansson, B.. Santell, S.O., Feinman, J.
(1987). EAF dust decomposition and metals recovery at
34. Dosaj, V.D., and Raucholz, A.W. (1989). Feasibility of ScanDust. 45111 Electric Furnace Conference, Proc.,
plasma smelting of silicon. Center for Materials vol. 45, pp. 81-89.
Produclion, Report no. 88-8, 134 pp.
48. Johansson, B., aod Lofgren, U. (1990). Operation of the
35. Dosaj, V.D., Haines, C.M., Brumels, M.D., and May, Scandust plant on stainless steel dust. Proceedings.
J.B. (1991). Silicon smelting in a closed furnace. 49th Conference on Plasma for Industry and Environment,
Electric Furnace Conference, Toronto. Oxford, The British National Committee for
36. McRae, L.B. (1983). The remelting of silicon fines. Electroheat. Paper 2.5, 13 pp.
Minlek, Report M135. 49. (a) Roddis, B.. and Cooke, R. (1990). Recycle of melals
37. (a) Bakken, J.A., Holt, NJ., and Jensen, R. (1988). from stainless steelmaking dusts by a plasma smelting

PLASMA-ARC TECHNOLOGY FOR FERROALLOYS 33

 technology. COllference on Recycling of Metalliferous (1989). Aluminium recovery from dross: comparison of
Materials, Binni1Jgham, Institution of Mining and plasma and oil-fired rotary furnaces. 9th lmemational
Metallurgy, pp. 221- 227. Symposium on Plasma Chemistry, IUPAC, Proceedings,
(b) Cbapman, C.D., and Cowx, P.M. (1991). Treatment vo/. 3, pp. 1697-1701.
of EAF dust by the Tetronics plasma processes. Steel 64. Plasma Energy Corporation and Alcan announce new
Times, vol 219, no. 6, pp. 301, 302, 304. dross processing technology. (1989190). Plasma
50. Kisimoto, Y., Tobo, H., Hara, Y., and Sakuraya, T. Quarterly, Report on PEC, FalllWinter, 2 pp.
(1990). Development of dust treatment system based on 65. Schmidt, R. (1992) Innovative methods for melting of
hollow cathode DC arc technology. Proceedings, aluminium. Plasma 20()()2, 13th Advanced Technology
Conference on Plasma for Industry and Environment, Symposium, Iron & Steel Soc., Charleston, South
The British National Committee for Electroheat, Paper Carolina.
2.4,8 pp.
Zinc
Alloy Scrap 66. Mazanek, M.S., Pargeter, l.K., and Cowx, P.M. (1990).
51. Muller, F. (1977). Plasma primary melting - a new pro- lMSrretronics thermal processing of EAF dust.
cess of steel manufacture. Neue Hut/e, vol. 22, no. 4, pp. Symposium on Industrial and Environmental
222-223. Applications of Plasma, Electric Power Research
52. Neuschutz, D., Rossner, H.-O., Bebber, H.J., and Institute, Palo Alto, 6 pp.
Hartwig, J. (1985). Development of three-phase AC 67. PockJington, D.N. and Cowx, P.M. (1990). The com-
plasma furnace at Krupp. lroll and Steel Ellgineer, vol. mercial application of plasma technology to steelworks
62,no.5,pp.27-33. dust treatment. Proceedings, Conference on Plasma for
Industry and Environment, Oxford, The British National
53. Knoppek, T., Lugscheider, W., and Steipe, O. (1985).
Committee for Electroheat, Paper 2.3, 8 pp.
Initial results of Voest-Alpine 45 t plasma furnace and
future aspects of plasma technology. Iroll and Steel 68. Pargeter, J.K., and Mazanek, M.S. (1991). lMS recy-
Engilleer, vol. 62, no. 5, pp. 23-26. cling services of the 1990s. 49th Electric Furnace
Conference, Toronto.
Magnesium 69. Complete EAF dust processing management, Davy
54. Schoukens, A.F.S. (1989). A plasma arc process for the McKee Hi-Plas System. Heckett Technology Services,
production of magnesium. Extraction Metallurgy '89 Inc., Brochure, 4 pp.
Symposium, London, Institution of Mining and 70. Mustoe, T. N., General Manager. Plazmet (Affiliate of
Metallurgy, pp. 209-223. U. S. Zinc). (1919). Goodyear Dr, Houston, TX 77017,
55. Cameron, A.M., Canham, D.L., and Aurich, V.G. personal communications, 5/90; 10/91.
(1990). Magnesium production by plasma-powered pro- 71. Schoukens, A.FX., Nelson, L.R., and Barcza, N.A.
cessing. JOIln/al of Metals, vol. 42, no. 4, pp. 46-48. (1991). Plasma-arc treatment of steel-plant dust and
56.Cameron, A.M. (1990). Technical characteristics of an zinc-containing slag: theoretical and practical considera-
economic plasma smelting process. Proceedings. tions. International Lead and Zinc Study Group
Conference, Rome.
Conference on Plasma for Industry and Environment,
Oxford. The British National Committee for
Iron
Electroheat, Paper 1.1, I I pp.
72. Gold, R.G., Sandall, W.R., Cheplick, P.G., and
Titanium Mac Rae, D.R. (1977). Plasma reduction of iron oxide
with hydrogen and natural gas at 100 kW and 1 MW.
57. Stocks, S. (1990). Plasma melting in titanium. Ironmaking and Steelmaking. no. I, pp. 10--15.
Symposium on Industrial and Environmental
Applications of Plasma, Electric Power Research 73. (a) ASEA. (1979). The ELRED Process. Annual Report,
Institute, Palo Alto, 7 pp. pp.7-8.
(b) Collin, P., and Stickler, H. (1979). ELRED - a new
58. Malley, D.R. (1990). Plasma hearth melting: a Pratt &
process for the cheaper production of liquid iron. AlSE
Whitney perspective. Advanced Aerospace Materials
Convention, Cleveland, 12 pp.
Conf., Long Beach, CA.
(c) Widell, B. (ASEA AB) (1977). Furnace and method
59. Reichman, S.H. (1989). Status of plasma melting of tita-
for the melt reduction of iron oxide. V.S. Patellt 4, 146,
mum at Wyman-Gordon. Vacuum Metallurgy Conference,
390, app/. 28 Mar. 1977, acc. 27 Mar. 1979.
Pittsburgh, PA.
74. Lorfonte (HFRSU) massive coal injection into hot blast
60. Eschenbach, R.C. (1990). Personal communication.
superheated by plasma torches. (1988). VIE Arc Plasma
61. Bhat, G.K. (1982). New developments in plasma arc Processes, lfIE International Union for Electroheat, pp.
melting. J. Vacuum Science & Technology, vol. 9. 91-93.
62. PEC and Leybold collaborate on plasma development. 75. New plasma melter fortifies General Motors' Central
(1989/90). Plasma QlIarterly, Report on PEC, Foundry Division. Iron & Steelmake,., vol. 16, no. 11,
FalllWinter, 2 pp. p.45.
76. Dighe, S.V. (199). Plasma fired cupola. Symposium on
Aluminium Industrial and Environmental Applications of Plasma,
63. Menuier, J., Zinunermann, H. t and DroLlet, M.G. Electric Power Research Institute, Palo Alto, 3 pp.

34 JNFACON6
77. Dighe, S.Y. (1991). Plasma cupola iron melling. 49th 87. Dembovsky, Y., and Motloch, Z. Potential of plasma
ElecLric Furnace Conference, Toronto. technology in the ladle refining of steel. Proceedings,
78. Pineau. M. (1990). Plasma on a found I)' cupola. Conference on Plasma for Industry and Environment,
Symposium on Industrial and Environmental Oxford, The British National Comminee for
Applications of Plasma, Electric Power Research Electroheat, Paper 1.5, 10 pp.
Institute. Palo Alto. 21 pp.
79 Herlitz. H.G., Johansson, B., and Santen, S.O. (1984). A
new family of processes based on plasma technology. Tundish Heating and Continuous Casting
Iroll and Steel Ellgilleer. vol. 61. no. 3, pp_ 39-44. 88. Kuwabara, T., Umezawa, K., Nube, T., and Fukuyama,
80. Union Steel Corporation of South Africa. (1984). M. (1987). Development of tundish plasma heater. 8th
Plasma production of reducing gas for direct reduction International Symposium 011 Plasma Chemistry, IUPAC
of iron ores. Meral Blllletin Mall/Illy. May, pp. 28, 31. Proceedillgs, vol. 4, pp. 2247-2252.
89. Tetronics R&D Co., Faringdon, U.K. (1990). Personal
Platinum and Gold
communication, May.
81. Saville, J. (1985). Recovery of PGMs by plasma arc
90. Mizushina, M., Kitano, Y., Suga, Y., Saiki, M., and
smelting (first commercial plant). 9th International
Suguro, Y. (1990). Development of tundish plasma
Precious Metals Coof., New York.
healing system. Proceedings. Conference on Plasma for
82. McRae. L.B., and Kleyensttiber. A.S.E., (1990). The Industry and Environment. Oxford, The British National
pyrometallurgieal rccovel)' of gold from leached calcine Committee for Eleclroheat.
residues. J. S. Afr. I""t. Min. Merall.. submitted 1990.
91. Hoster, T., Bebber, H,J., and Breitzmann, M. (1988).
Ladle Heating and Refining Isothermal C0111;IlUOllS casting using the Krupp A. C.
83. Oliver. J.F.. Stefanic, S.M., and Kemeny, F.L. (1988). plasma system for lUndis!? heating. Krupp Industries,
Plasma heating for ladle treatment furnaces. 46th [nc., Brussels, 8 pp.
£IecII'ic Arc Furnace Conference, Proc.. vol. 46, pp. 92. Saito, T., Matsuo, K., Fujimoto, H., Maeda, M., lmiya,
313-321. K., and Umahashi, K. (1991). Development of AC plas-
84. Kemeny, F.L. (1992). Plasma arc ladle heming and ma arc heating system in tundish. Tetsu-to-Hagane, vol.
refining for the foundry industry. Plasma 20002, 13th 77, no. 10, pp. 1649-1655.
Advanced Technology Symposium, Iron & Steel Soc.. 93. Cheek, D.L. (1988). Progress report on the plasma
Charleston, South Carolina. tundish heater at Chaparral Steel. Plasma Energy
85. Bakken. J.A., and Haraldsen, S. (1982). Immersed arc Corporation, Internal Memo, II pp.
heater. 40th Electric Arc Furnace Conference. Proc., 94. Walzer, G. (1989). Plasma tundish heating - improving
vol. 40, pp. 251-256. casting productivity at Chaparral Steel. Center for
86. Neuschutz, D., Schubert, K.-H., and Bebber, H,J. Metals Production, Mellon Inst., Pinsburgh, CMP
(1991). Metallurgical results from a 30 ton AC plasma Tecllllpplication, CMP-046, 2 pp.
ladle furnace. Steel Research. vol. 62, no. 9, pp. 95. Stoney Creek Steel: a steelmaking showcase. (1989).
390--394. lroll Age, vol. 5. no. II, pp. 33-34.

PLASMA-ARC TECHNOLOGY FOR FERROALLOYS 35