Fluxes - Is There A Panoramic View?

K.C. Mills and D. See

Department of Materials, Imperial College of Science, Technology and Medicine, London SW7 2BP, England

(Received January 16, 2001.)

ABSTRACT (ii) in Materials, "a substance used in soldering, welding and brazing to remove oxides from the Fluxes used in the following processes have been surfaces to be joined," and examined to determine what common features they (iii) in Metallurgy, "in liquid metal processing, a non- might have: soldering, brazing, welding, electroslag metallic material used to protect the metal and remelting (ESR), ingot and continuous casting, and the remove impurities." production of aluminium master alloys. A quick examination of the functions of fluxes used The fluxes have been inspected in terms of (i) the in a specific process (such as welding) will quickly functions that they have to perform in the process, (ii) indicate that properties such as 'promoting arc stability' the properties of the flux which are important to carry are not covered by the above definitions. out those functions, and (iii) typical compositions. It Consequently, in this review we have examined the was concluded that there were some common features following types of fluxes: for the fluxes; that they should (i) melt at least 100°C - soldering and brazing, below the metal, (ii) form a protective layer on the - welding, metal, (iii) absorb inclusions from the metal and, (iv) - electroslag refining, usually contain halides. On the other hand, frequently, - formation of master aluminium alloys containing key features were specific to the type of flux and its TiB , and application; for instance, powder consumption rates are 2 - mould and ingot fluxes for the casting of steel. essential to the performance of mould fluxes, as are factors affecting arc stability, to the performance of We have examined the above fluxes in terms of: welding fluxes. (i) the functions that the flux must perform in each application, (i) the physical properties which are important in carrying out these functions, and 1. INTRODUCTION (ii) typical compositions of the various fluxes used.

The objective of this paper is to examine fluxes used 2 SOLDERING AND BRAZING FLUXES in different industries with a view to identifying any 2.1. Soldering features which (i) are common to all fluxes and (ii) are individual to specific applications. In soldering and brazing the surfaces of the metals We start with dictionary definitions of a flux /1 /: being joined remain solid throughout the entire (i) in Chemistry, "a substance promoting the operation. The filler metal from the solder makes the melting of another substance to which it is joint by intimate attachment to the base metal surfaces. added," The filler metal must 'wet' the workpiece surface and Vol. 20. Nos. 3-4, 2001 Fluxes - Is There A Panoramic View?

must flow easily over the surfaces. These characteristics fluxes used must be capable of operating within will be promoted by low surface tensions and low temperature ranges of 600 - 800° C and 750 - 1150°C contact angle (solder on base metal) for the solder or respectively. The flux must be liquid and should not braze. The fluxes are added to facilitate this process and decompose at the operating temperature. must carry out the following tasks 121: Fluxes used in hard soldering contain low melting

(i) remove any oxide films (usually Cu20 or SnO) compounds such as B203 NaF, KF, KHF2 borofluorides, present on the workpiece surfaces and protect borates, etc. these surfaces from any further oxidation, (ii) any remaining left-over fiux or reaction products 2.3. Brazing must be easily displaced from the metal, and (iii) promote the heat transfer from heat source to the Brazing is carried out on a variety of metals: Al, Mg, joint. Ti steels and Cu/Al alloys using a variety of filler The efficiency of the flux is controlled by its ability metals, each of which requires an appropriate flux. A to promote melting and is associated closely with the variety of compounds are used in these fluxes (e.g. LiCl, reactivity of the flux. However, if the flux is reactive, it LiF, NaF, and KF. KHF2, ZnCl2, KBF4, B203 etc). can cause corrosion of the joint in time. This is a major Brazing is very similar to hard soldering and the issue, especially in the case of soldering electronic requirements of the flux are identical to those for hard components 111. In some processes the fluxes are soldering. removed by washing after soldering. Fluxes for soft soldering tend to be of two types: 2.4. Discussion of Joining Fluxes (i) those soluble in organic fluids, and (ii) those soluble in water. The temperatures used for soft and hard soldering In some cases activators are added to the flux to and for brazing vary but the role of the flux is identical enhance the reactivity of the fluxes. Activated fluxes in all these applications. First, and most importantly, the nearly always contain halides. flux must promote the wetting of the filler metal and to do this, it must (i) remove the oxide films on the base (i) Fluxes soluble in organic liquids metal surface to permit wetting and (ii) have a low These are mostly based on colophony (rosin) which is melting point to provide a liquid layer to protect the a natural product derived from pine sap. The principle base metal from further oxidation. The presence of component of colophony is abietic acid. Colophony halides in fluxes is interesting. The main function of the has a melting point (mp) of 172°C and is soluble in halides is the chemical attack of the oxide films on the acetone and alcohol. It tends to have low reactivity surface of the base metal. However, there is evidence to and activators containing halides are usually added. suggest that Group VII elements (I, Br, CI, F) are very (ii) Fluxes soluble in water surface-active and thus the presence of F or CI at the

These fluxes are frequently based on ZnCl2 (mp283°C) metal/solder (or braze) interface could reduce the

or NH4CI or on mixtures of the two components. surface tension of the solder (or braze) significantly and There is little fluxing action in the absence of encourage spreading of the solder. moisture.

2.2. Hard Soldering 3 WELDING FLUXES The principles of this process are identical to those in soft soldering but the temperatures involved are In welding the metal pieces being joined are melted higher. In hard soldering the filler metals tend to be (cf soldering,brazing). Fluxes are used to protect the silver or copper to which elements such as zinc, tin etc. metal surface from oxidation. Failure to protect the are added to reduce the melting point. Consequently, the metal can lead to a weld with gas pores and poor surface

156 K.C.Mills and D. See High Temperature Materials and Processes

appearance /3,4,5/. Fluxes are used in four welding rapid change in polarity in AC welding and thus fluxes processes: with significant F content can only be used in DC (i) Manual - Metal Arc /MMA/ shown in Figure la, welding (where F ions migrate to the anode). The ions (ii) Submerged -Arc Welding /SAW/ shown in produced in the arc also affect the droplet size. The Figure lb, basic fluxes tend to produce larger metal droplets from (iii) Flux-cored wire welding shown in Figure lc, and core wire whereas more acidic fluxes (which give less (iv) Electroslag - welding shown in Figure 1 d. oxygen protection) tend to produce 'spray transfer' 131. This may also be related to the surface tension of the Some of the requirements are common to all the core wire under the welding conditions. processes, but other tasks are specific to an individual Several types of fluxes are available for MMA welding process. Functions of welding fluxes include welding and the selection is dependent upon the type of /3,4,5/: weld required, for instance, practical (i.e. non- (i) to melt at a temperature below (e.g. 100°C) the horizontal) welding, is best performed with spray metal melting point of the base metal, transfer. The following fluxes are available: - (ii) to protect the weld from the atmosphere; fluxes (i) Acid (high Si02, FeO, MnO contents) - high Ο contain (a) carbonates and fluorides which and Η levels in welds leading to poor decompose to give C02 or F shielding and (b) in mechanised properties some cases contain metals (e.g. Al) which react (ii) Rutile - (Ti02 /30-50%/, CaO, Mn0,Si02) with oxygen to reduce the partial pressure of (iii) Basic - (30-40% CaF2, 25-40% CaO, 10-15% oxygen, S;02) low Ο and Η contents in weld metals and (iii) to form a slag crust which protects the weld from give large droplets, are difficult to use in AC subsequent oxidation during cooling, welding. (iv) to remove impurities from both the surface (e.g. rust) and the metal (e.g. inclusions) by absorbing them into the molten slag, (v) to encourage arc initiation and stability; sodium 3.2. Submerged Arc Welding (SAW) and potassium ions are beneficial for this task; In this process /3,4,5/ an arc is struck between a this is very important to the welding process, and consumable electrode and the workpiece which is (vi) to increase metal deposition rates of the metal covered with a layer of flux (Figure lb). A cavern is electrodes (in practice this is achieved by formed in the metal (Figure lb) and its shape is additions of ferrous alloy powders) and to maintained as the weld travels along the workpiece. The desulphurize the weld metal (when using basic fluxes can be produced in pre-melted or agglomerated fluxes). forms. The main tasks of the slag are (i) to form a liquid slag layer which both protects the surface of the weld and absorbs inclusions (ii) to provide a stable arc which 3.1. Manual Metal Arc [MMA] Welding is very important (iii) to form a solid slag which protects The core wire consists of a metallic electrode the weld on cooling and which is easy to remove and covered by a thick layer of flux (which is manufactured (iv) to promote heat transfer. Several types of SAW by an extrusion process) 111. The welding process fluxes are available each having its own merits and consists of an arc being struck between the core wire limitations. Acidic fluxes are based on eutectic MnO + and the work pieces (Figure la). The flux melts and the Si02 slags with additions of CaF2 etc.: these fluxes tend constituents decompose to provide (a) protection against to give high oxygen levels in the weld. The oxygen oxygen ingress and (b) arc stability. The electrode core content is reduced by additions of Ti02, Al203 and by wire is consumed in the welding process. The arc Si02 being replaced by A1203 and Ti02 in some + + stability is improved by Na , K and iron ions but F" formulations. They contain CaF2 usually around the ions interfere with efficient metal transfer during the 10% level but can be up to 20% in more basic powders.

157 Vol. 20, Nos. 3-4, 2001 Fluxes - Is There A Panoramic View? Consumable electrode φ

Parent Metal

Gas optional

\ Nozzle optional Water Solidified slag amlrH Molten Solidified weld metal Flux Flux cored Slag bath Molten Molten slag ^—'electrode (3/4*1 V» metal pool in depth) Molten metal Solid

Solid ingot Water Weldφ metal φ

Fig. 1: Schematic diagrams showing the following processes (a) Manual Metal Arc (MMA) welding (b) Submerged Arc Welding (SAW) (c) Fluxed Core-wire Welding (d)Electroslag Welding (ESW) and (e) Electroslag Refining (ESR)

These basic powders tend to provide some importance of this function. Two types of fluxes are desulphurization of the weld metal but care must be available: - taken in storage to avoid moisture pick-up, which can (i) Self shielding flux-cored wires which contain Al lead to hydrogen cracking in the weld. or Mg !0% or 50%Fe additions to deoxidise the

CaF2 environment of the weld along with either

CaC03 or Ti02 or both, providing further F and 3.3. Fluxed Core Wire Welding C02 shielding, and This is an automatic or semi-automatic process in (ii) Gas-shielded flux-cored wires, which are

which an arc is struck between the reeled wire and the protected by C02 or C02/Ar, gas atmospheres.

workpiece (Figure lc). The wire contains a core of flux. These fluxes are based on Si02 - Ti02, CaO or

Since the wire is on the outside shielding of the metals CaF2 - Si02, CaO-CaF2 - Si02-Ti02, all with is difficult and the composition of the flux reflects the additions of deoxidants such as Fe-Mn.

158 K.C.Mills andD. See High Temperature Materials and Processes

3.4. Electro-Slag Welding (ESW) This small droplet has a huge (surface area/mass) ratio and refining reactions occur rapidly as it falls through The electroslag welding (ESW) process (Figure Id) the molten slag and then splashes into the molten metal differs from the other three processes in that it does not pool. involve a continuous arc /3/. Once the flux is molten The principal functions of the flux are: and a flux pool has been established, the arc is (i) to supply heat to the system by acting as a extinguished and the process relies on resistive heating resistive medium for the passage of electrical due to the current flowing through the molten slag. The current, consumable electrode provides a stream of metal (ii) to remove impurities (such as S) by mass transfer droplets, which form a molten weld pool. (Figure Id) across the metal/slag interface, e.g. Equation 1 The weld is retained by using water cooled copper where the underline indicates the metal phase plates. It is used extensively for welding thick sections S + O2" (slag) = Ο + S2 (1) held in the vertical position. (iii) to remove inclusions by (a) dissolution of the The principal functions of the slag are: oxide particles in the flux and (b) flotation of the (i) it should melt at a temperature 100C below the particles, and metal, (iv) to isolate the refined product from both the (ii) it should provide resistive heating, and atmosphere and the water-cooled mould. (iii) it should protect the steel from oxidisation. The composition of ESW fluxes tends to be The most important of these tasks is, undoubtedly, determined by the electrical conductivity of the molten the removal of harmful impurities and the inclusions flux. The conductivity must be a balance between (i) a since both cause a marked decrease in mechanical high conductivity to provide a suitable current between strength of the alloy. The rates of the refining reactions electrode and weld pool and (ii) a sufficiently low are rapid because of the huge (surface area/mass) ratio conductivity (or high resistivity) to provide the of the metal drop as it descends through the molten slag necessary power. Slag-metal reactions such as the layer. The residence time for the descent increases desulphurization of the metal occur readily because of with increasing flux velocity but this must be balanced (a) the much longer contact times between molten metal against the fact that the kinetics of the metal/ and slag (b) the large surface area/volume ratios of slag reactions are often rate-controlled by the slag droplets which provide fast kinetics and (c) the high viscosity.

sulphur capacities provided by CaF2 - rich slags. The kinetics of the refining reactions are also

ESW fluxes tend to resemble SAW fluxes but have affected by the interfacial tension (y,ns) between metal

higher CaF2 contents to obtain the required electrical and slag. A low value of yms favours an increase in the conductivity values. mass transfer rate. The classic work at IRSID /9,10/ showed that rapid mass transfer between metal and slag was accompanied by a marked decrease in interfacial

4 ELECTROSLAG REFINING (ESR) FLUXES tension (with yms close to zero) which could readily lead to emulsification of metal in the slag (or slag in the The electroslag refining process is used for refining metal). This, in turn, leads to rapid refining but also to nickel-based superalloys and steels. It is very similar to entrapment of slag globules in the metal, which is very the electroslag welding (ESW) process in that it uses a undesirable. Fortunately, the molten flux pool allows consumable electrode of the required steel component time for any flux globules to float out of the metal. The cast by conventional metallurgical processes. The surface tensions (γ) of metals are reduced dramatically electrode is submerged in a molten slag held in a water- by increasing S contents. The S content of the pool will cooled copper mould (Figure le) which serves as the be lower than that of the drop because of the

other electrode. When an electrical current is passed, the desulphurization occurring in the pool (γ p00| > Ydr0p)· tip of the electrode melts and a small droplet is formed. Under these conditions the drop (plus any inclusions)

159 Vol. 20, Nos. 3-4, 2001 Fluxes - Is There A Panoramic View?

will not penetrate into the pool but will roll along the production of large diameter ingots. surface /11/, thereby facilitating the removal of The phase diagram of the CaF2 + CaO + A1203 inclusions into the molten flux. system indicates that there is an extensive miscibility The transport of inclusions from the metal to the gap and the compositions of ESR fluxes frequently liquid flux is also dependent upon yms. Kozakevitch and border on this miscibility gap, especially, when Olette 191 showed that removal was favoured if: compositional changes (e.g. Equations 4 and 5) are taken into account. The contours of the measured Ymi >Ysi + Yms, (2) physical properties show a remarkable curvature in the region of immiscibility (Figure 7a,b, c). The shape of where the subscript i refers to the inclusion. They these contours has been ascribed to (a) the problems of showed that the process was nearly always favoured. making property measurements on immiscible liquids However, for a refined metal in the metal pool the S and /7,8/ and (b) the tendency to make thermophysical Ο levels will be low and consequently the surface property measurements in the upper part of the crucible tension of the metal will be high. The interfacial in order to minimise the interference from the base of tension, Yms is given by Equation 4 where φ is an the crucible ie making measurements, preferentially, in interaction coefficient. the lighter of the two liquids. It was proposed that overall values for the two liquids could be obtained by Yms = Ym + Ys " Φ (2YmYs)° 5 (3) extrapolating values into the two phase region /7,8/. Fluoride emissions are a threat to health and safety

The ym term is typically 5 times that of ys so a high and can cause corrosion of plant equipment. In recent

γ,„ will tend to lead to a high y,m value. Chung and years the fluoride contents of ESR fluxes have been Cramb /12/ showed that small inclusion particles cannot decreasing to 50% and attempts have been made to puncture the interface directly and require a certain rest produce fluoride-free fluxes. The major problem is to period or residency before they can traverse the obtain a flux with the required electrical conductivity interface. for a successful operation. Thus, in summary, the key physical properties of ESR fluxes are the electrical conductivity, interfacial tension and the viscosity. ESR fluxes usually contain 5. FLUXES USED IN PRODUCTION OF high concentrations of CaF2, e.g. typical ESR fluxes ALUMINIUM-MASTER ALLOYS contain (10% CaF2+ 30% Al203);(70% CaF2+15%Al203

+ 15%CaO). Titanium boride, TiB2 and TiAl3 act as grain They are fundamentally unstable at high refiners in Al alloys. The master alloy is prepared by temperatures, since CaF2 reacts with the various oxides adding fluxes of K2TiF6 and KBF4 to the Al where it or moisture to form gaseous fluorides: reacts to form TiB2 and TiAl3 and a mixture of KF and

A1F3 The separation of reaction products from the metal

CaF2+Al203 = AlF3(g) + Ca0 (4) phase is a key step in any metal production process. It is easier to separate two liquid phases and so processes are

CaF2 + H20 = 2HF(g) + CaO (5) usually designed to produce liquid metals and liquid reaction products. Thus there is a tendency for the fluoride content of It was mentioned above that rapid mass transfer the ESR flux to decrease and the oxide content to could lead to low interfacial tension which, in turn, can increase with time. Great care must be taken on storing lead to emulsification and rapid kinetics. However,

ESR fluxes since both CaO and CaF2 tend to absorb emulsiflcation can also lead to flux entrapment. If pure moisture. This leads to hydrogen pick-up in the metal. K2TiF6 and KBF4 are used flux entrapment in the Although the diffusion of hydrogen is reasonably fast, aluminium occurs (Figure 2) which is very undesirable hydrogen pick-up remains a serious problem in the /13/. However, it is known that if the K2TiB6and KBF4

160 K.C.Mills andD. See High Temperature Materials and Processes

Fig. 2: Photograph showing fluoride emulsion in Al master alloy.

contain as little as 40ppm Ca then emulsification does 6.2. Mould fluxes for the continuous casting of not occur and the waste products float to the surface. steel The extremely low levels of Ca required strongly Mould fluxes play an important role in the indicate that this is probably associated with interfacial continuous casting of steel. Molten steel from the phenomena. It has been suggested that Group VII tundish is poured through a submerged entry nozzle into elements are highly surface active (although it is a water-cooled, copper mould where the steel freezes to difficult to keep fluorine in the metal). One possible form a solid, steel shell. The dummy-bar filling the explanation for the effect of Ca is that Ca in addition to bottom of the mould is withdrawn and the solidified Al significantly reduces the soluble fluorine shell is further cooled by passing it through water sprays concentration in Al and subsequently increases the until it has completely solidified, whereupon, it is slit surface tension and thus the interfacial tension yim of the into lengths(Figure 3a). metal phase (Al). The mould flux is continuously added to the top of the mould where it forms various layers including a liquid slag pool (Figure 3b). The mould is oscillated and 6 MOULD AND INGOT FLUXES molten slag is drawn into the channel between the shell and the mould during the downstroke of the mould. The 6.1. Ingot Fluxes slag drawn into the channel in the first moments of The traditional method of casting steel was in the casting, freezes against the mould wall and forms a solid form of ingots and ingot fluxes were used to prevent the slag layer (ca. 2-3 mm thick). A thin liquid slag layer steel from oxidisation and provide some lubrication. (0.1-0.3 mm thick)remains in contact with the shell and Since the flux must provide a reducing atmosphere for provides the necessary liquid lubrication of the steel. the entire casting period they tended to have high The mould flux must perform the following tasks : carbon contents and tended to be based on fly ash with (i) form a molten slag pool which protects the steel other oxides added to lower the melting point. However, interface from oxidation, ingot-casting has been largely superseded by the (ii) provide a liquid slag film which lubricates the continuous casting of steel. shell throughout the mould,

161 Vol. 20, Nos. 3-4, 200! Flaxes - Is There .4 Panoramic View?

Ladle.

• Powder Λ Srtered powder Mushy stag v ^ 1 γ Carbon ι Liquid «lag Tundish slag em 1 Liquid steel Mold I Water spray Crystalline and hendine dag Um • • ι . , She!

Fig. 3: Schematic diagrams showing (a) Continuous caster for steel and (b) the various powder and slag layers formed in the mould.

(iii) provide the right level of horizontal heat transfer of mould flux and reported as Q, in kg (flux) tonne"' but between the shell and the mould, and Wolf /14/ pointed out that when converted to Qs in kg 2 1 (iv) absorb non-metallic inclusions e.g. A1203, Ti02, (flux) (m of mould)" it provided a measure of the

into the molten slag pool. liquid lubrication. It was subsequently shown that Qs Failure to carry out any of these tasks satisfactorily was a sensitive function of the parameter, R, which will result in inferior product quality or process control represents the ratio of the (surface area/volume)of the problems. Providing the flux forms a molten slag pool mould (Figure 4). Values of R are in the hierarchy: thin and protects the entire steel surface from oxidation, it is slabs and billets > blooms > slabs. usually considered that the slag infiltration and the Several empirical rules have been reported for control of the horizontal heat flux are the most calculating powder consumption and these have been important functions of the flux. reviewed by Wolf /14/. Two simple rules reported by Wolf /14/ are derived from relations for optimum 6.2.1 Powder consumption casting conditions /15,16/ viz: Powder consumption is a measure of the slag infiltration into the mould / strand channel and liquid 0 5 Qs= 0.7 / η vc (7) lubrication supplied to the shell since the frictional forces are related (Equation 6) to the reciprocal of the = Qs 0.6 / η vc (8) liquid slag film thickness (d|). An analysis of plant data has shown /17,18/ that these relations provide a reasonable estimate of the

F = Tl(vc-vm)/d, (6) powder consumption with Equation 8 providing a slightly better fit Thus it is essential that the casting flux where vc and v,„ are the casting speed and the mould should have the correct viscosity for the mould velocity and η is the viscosity. Thus the frictional forces dimensions and casting speed. In practice, the increase with increasing viscosity and decreasing slag oscillation characteristics also affect the powder film thickness (or powder consumption). consumption, Qs, which increases with decreasing Originally it was used only as a measure of the cost frequency and increasing stroke length.

162 K.C.Mills andD. See High Temperature Materials and Processes

0.8

0.7

b 0-5 c ο α 0.4 Ε 3 <0 ο 0.3- ο α> 1 0.2- ο Q. 0.1 -

ο 10 20 30 40 50

Sufaoe Area toVdime Ratio, Rim1)

Fig. 4: Powder consumption as a function of the (Surface Area/Volume) ratio, R, for the mould.

6.2.2 Horizontal heat transfer in viscosity and which corresponds to the point where Longitudinal cracking of the surface and the sub- liquid lubrication breaks down. Figure 5 shows a plot of surface of the continuously-cast strand is a major Tbr as a function of viscosity (and casting speed) for (1) problem,especially for .medium-carbon (MC) steel MC, crack -sensitive steels (2) HC, sticker-sensitive grades (0.06- 0.18% C).This is the result of the 4% steel grades and (3) all other grades. Thus one can mismatch in the thermal shrinkage coefficients of the δ- predict the Tbr needed for casting the steel grade and and austenite phases of iron. This, in turn, produces casting speed. Thus from the viewpoint of avoiding both thermo-stresses, which are relieved by longitudinal longitudinal cracks and sticker-breakouts, the most cracking. The usual strategy to minimize these stresses important properties are the Tbr and the degree of is to ensure that the newly -formed shell is both as thin crystallization and consequently indicates that the and as uniform as possible. This is achieved by reducing properties of the solid slag are also of prime importance. the horizontal heat transfer through the creation of a thick, crystalline slag layer /17,18/.

In contrast, the shells of high-carbon (HC) steels 6.3. Slag entrapment (C>0.4%) tend to be mechanically-weak and sticker- breakouts (where molten steel pours from the mould) Slag- and gas- entrapment are also major sources of usually arise as a result of a lack of lubrication but can defects in continuously- cast steels. It has been shown be minimized by creating a thicker, stronger shell to that they are mainly caused by the casting conditions withstand the ferro-static pressure. In this case it is which promote turbulence in the liquid metal /19,20/. necessary to enhance the horizontal heat transfer The mould flux does play some part and the problem through the formation of a thin, glassy slag layer. The can be alleviated by increasing either interfacial tension thickness of the slag layer is partially determined by the (Yms) or the viscosity of the flux; in practice, the latter solidification or break temperature (Tbr) that is the approach is usually adopted but this can lead to a low temperature on cooling where there is a sharp increase powder consumption.

163 Vol. 20, Ν as. 3-4, 2001 Fluxes - Is There A Panoramic View?

BlIIMs Blooms Slabs LC Ο • * MC Ο [• χ IC c=> Ο • HC • + 1 0.5 0 t.5 2.0 2.5 3.0 3.5 4.0

Viscosity at 1.300 =C, dPas

Fig. 5: Break temperature as a function of the viscosity of the mould flux (at 1300°C) for slab-, bloom- and billet- casting.

7 DISCUSSION obtain a fluoride-free flux with the right electrical conductivity. Fluoride emissions with the associated This study has shown that there are functions and plant corrosion problems are also are a source of worry features of mould fluxes which are common to all in the continuous casting of the steel /21/. Consequently, fluxes, viz: some steelworks are now demanding fluxes with lower (i) they all have melting points significantly below fluoride contents and ever -more stringent legislation is that of the metal phase, likely to lead to even lower halide levels in the future. (ii) they all are expected to protect the surface of the The use of pre-fused mould fluxes would be one option metal from oxidation, for reducing halide emissions and would also help to (iii) they are expected to absorb inclusions from the reduce the absorption of moisture. metal, and (iv) they tend to contain halides and this may be related to the surface activity of the Group VII elements. CONCLUSIONS However, the individual features, which are specific to the type of flux or application, are equally, or even- 1) Halides are used to lower the melting point, more important, than those mentioned above. For viscosity and possibly the surface tension of fluxes instance, the powder consumption and the factors but the levels of halides are likely to be reduced in affecting horizontal heat transfer are of prime the future because of environmental and safety importance in continuous casting and arc stability in concerns. welding processes. 2) All fluxes have melting points lower than that of the There is one feature of fluxes, common to all fluxes, metal and the slag formed is expected to both which is becoming ever more important, namely, the protect the metal from oxidation and absorb environmental concerns, health hazards and associated inclusions from the metal. plant corrosion problems posed by the use of halides in 3) The individual functions and features specific to one fluxes. ESR slags contain up to 70 % of CaF2 and are type of flux tend to be very important, such as arc basically unstable at high temperatures. It is noticeable stability in welding or powder consumption in that the fluoride contents of fluxes have been gradually continuous casting. reduced to 50 %CaF2 and fluoride-free fluxes are 4) The high surface activity of the halides may be a currently being developed. The traditional problem is to reason for their widespread use in fluxes.

164 K.C.Mills andD. See High Temperature Materials and Processes

REFERENCES 11. T. Takasu and J.M. Toguri: Marangoni and Interfacial Phenomena in Materials Processing, 1. Dictionary of Science and Technology, Academic edited by E.D. Hondros, Μ McLean and KC Mills, Press, NY. IOM, London, 1998; 153. 2. R.J. Klein Wassink: Soldering in Electronics, 2nd 12. Y. Chung, Ph.D. thesis, Surface and Interfacial Edition, Electrochemical Publications, Ayr, Phenomena of Liquid Iron Alloys, Dept. of Mat. Scotland, 1989; Chapter 5. Sei. and Eng., Carnegie Mellon University, 1984. 3. L. Davis: An Introduction to Welding Fluxes for 13. M.S. Lee, B.S. Terry and P. Grieveson: Met. Mild and Low Alloy Steels, The Welding Institute, Trans. Β 24B, 947-953 (1993). Cambridge, 1981. 14. Μ. Wolf: Ρ roc. 2"d European Conf. Continuous 4. C.E. Jackson: "Fluxes and Slags in Welding", Casting, held in Düsseldorf, 1997. Welding Research Council Bulletin, 190, 1973. 15. Μ. Wolf: AIME Electric Furnace Proc. 40, 335- 5. M.L.E. Davis and N. Bailey: Metal Construction, 346 (1982). April, 1982; 202-209. 16. S. Ogibayashi et al: Nippon Steel Technical Report 6. W.E. Duckworth and G. Hoyle: Electroslag 34, 1-10(1987). Refining, Chapman and Hall, London, 1969. 17. K.C. Mills: Proc. of Alex McLean Symposium, 7. K.C. Mills and B.J. Keene: Intl. Metal Review 1, held in Toronto, July, 1998. 21-67(1981). 18. S. Sridhar, K.C. Mills, V. Ludlow and T. 8. K.C. Mills: "Schlaken fur das ESU (Slags for Mallaband; Proc. 3rd European Conf. Continuous ESR)'\ in Schlaken in der Metallurgy, Verlag Casting, held in Madrid, 1998. Stahleisen, VDEh, Dusseldorf, edited by K. Koch 19. S. Feldbauer and A.W. Cramb: Proc. 78,h and D. Janke, 1984; 297/311. Steelmaking Conf. TMS, Warrendale, PA, 1995; 9. P. Kozakevitch: Symp. Phys. Chem. of 655. Steelmaking, MIT Press, 1957; 134. 20. W.H. Emling et al: Proc. 78'h Steelmaking Conf 10. H. Gaye, L.D. Lucas, M. Olette and P.V. Ribaud: TMS, Warrendale, PA, 1995; 37. Can. Met. Quart. 23, 179-191 (1984). 21. A.I. Zaitsev et al. : Steel Research 65, 368 (1994).

165