Volume 15 - Casting Part 2 ppt

where ρS B and ρ* L are the respective densities of solid steel and liquid iron at the liquidus composition, CS B is the bulk carbon concentration of solid steel, and dr/dt is the rate o

Fig 7 Melting curves for 75 mm (3.0 in.) diam steel rods immersed in iron-carbon baths A, high (4.6% C)

iron-carbon bath; B, low (2.1% C) iron-carbon bath Plotted is the fractional thickness (the fraction of the rod diameter remaining undissolved) versus immersion time Source: Ref 10

The conditions existing in the bath are illustrated in Fig 8 Figure 8(a) shows the carbon concentration profiles in the liquid bath in the vicinity of a dissolving steel bar The meaning of the carbon concentration notation is indicated in Fig 8(b) As indicated in Fig 8(a), carbon diffuses across the liquid boundary layer, while equilibrium conditions are maintained at the solid/liquid interface Thus, the carbon concentrations in the solid and the liquid at the interface are the

isothermal solidus C* sS and the liquidus C* L concentrations, respectively Dissolution takes place by isothermal conversion of the steel into a liquid, that is, by a chemical melting process It is clear from Fig 8(b) that as temperature

increases, C* L decreases significantly Thus, the driving force for diffusion increases with temperature

Fig 8 Carbon concentration profiles (a) for a steel rod dissolving in an iron-carbon melt (b) Iron-carbon phase

diagram defining the carbon concentrations noted in (a) Source: Ref 10

Based on Eq 3 and Fig 4, the flux of carbon entering the steel is:

where ρS B and ρ* L are the respective densities of solid steel and liquid iron at the liquidus composition, CS B is the bulk

carbon concentration of solid steel, and dr/dt is the rate of change of the radius or thickness of the steel with time By combining Eq 6 and 7 and integrating between the limits r = r0 at t = 0 and r = 0 at t = tm, the isothermal melting time tm is given as:

In laboratory studies, Eq 8 accurately predicted steel dissolution rates in a stagnant bath (Ref 10) Research has suggested

that melting times are reduced by about a factor of two in induction-stirred melts (Ref 9) Other work has indicated that tmcan be reduced by an order of magnitude with strong convection (Ref 11)

On the basis of Eq 8 and the mass transfer correlation for rapidly dissolving rods (Ref 12), correlations were developed

relating tm to bath temperature and to carbon concentration (Ref 10) This is given in Fig 9 for a steel thickness of 2.5 mm

(0.098 in.) Because tm is directly proportional to the thickness of the steel, Fig 9 can be used to estimate tm for a wide range of steel thickness by making the appropriate thickness correction Checks of the predictions from Fig 9 in

production operations have indicated that predicted tm values are realistic (Ref 10)

Fig 9 Predicted melting times for vertical steel plates and cylinders, 2.5 mm (0.098 in.) scrap thickness and

0.1% C composition, [(%S)i = 0.03] immersed in stagnant iron-carbon baths at different bath temperatures Source: Ref 10

References cited in this section

1 R.I.L Guthrie, Addition Kinetics in Steelmaking, in Proceedings of the 35th Electric Furnace Conference,

Iron and Steel Society of AIME, 1978, p 30-41

2 F.A Mucciardi and R.I.L Guthrie, Aluminum Wire Feeding in Steelmaking, Trans ISS, Vol 3, 1983, p

53-59

3 J.W Robison, Jr., "Ladle Treatment With Steel-Clad Metallic Calcium Wire," Paper 35, presented at

Scaninject III, Part II, MEFOS, Lulea, Sweden, 1983

4 L Kalvelage, J Markert and J Pötschke, Measurement of the Dissolution of Graphite in Liquid Iron by

Following the Buoyancy, Arch Eisenhüttenwes., Vol 50, 1979, p 107-110

5 R.G Olsson, V Koump, and T.F Perzak, Rate of Dissolution of Carbon in Molten Fe-C Alloys, Trans

Met Soc AIME, Vol 236, 1966, p 426-429

6 O Angeles, G.H Geiger, and C.R Loper, Jr., Factors Influencing Carbon Pickup in Cast Iron, Trans AFS,

Vol 74, 1968, p 3-11

7 M Eisenberg, C.W Tobias, and C.R Wilke, Mass Transfer at Rotation Cylinders, Chem Eng Prog Symp

Series, Vol 51, 1955, p 1-16

8 R.G Olsson, V Koump, and T.F Perzak, Rate of Dissolution of Carbon Steel in Molten Iron-Carbon

Alloys, Trans Met Soc AIME, Vol 233, 1965, p 1654-1657

9 R.D Pehlke, P.D Goodell, and R.W Dunlap, Kinetics of Steel Dissolution in Molten Pig Iron, Trans Met

Soc AIME, Vol 233, 1965, p 1420-1427

10 R.I.L Guthrie and P Stubbs, Kinetics of Scrap Melting in Baths of Molten Pig Iron, Can Metall Q., Vol

12, 1973, p 465-473

11 K Mori and T Sakuraya, Rate of Dissolution of Solid Iron in Carbon-Saturated Liquid Iron Alloys With

Evolution of CO, J Iron Steel Inst Japan, Vol 22, 1982, p 964-990

12 P.T.L Brian and H.B Hales, Effects of Transpiration and Changing Diameter on Heat and Mass Transfer to

Spheres, AIChE J., Vol 15, 1969, p 419-425

Purification of Metals

The structure and properties of cast metals are sensitive to numerous impurities Because purification of melts generally adds considerable cost to castings, the lowest cost and surest defense against contamination is careful selection of scrap Purification is generally reserved for elements that are so pervasive that avoidance is impossible This is exemplified by sulfur and oxygen removal from cast iron and steel, respectively, and removal of alkali and alkaline earth elements from aluminum Because of their immediate importance, aspects of the physical chemistry of these processes are reviewed

Ferrous Melts

One of the most important processes involved in cast iron and steel production is desulfurization For steels, desulfurization is necessary to reduce the level of inclusions, leading to stronger and more fatigue-resistant steels For cast iron, desulfurization is practiced in the manufacture of ductile iron castings in order to develop spherical graphite morphology Ductile iron is used in applications where high fracture toughness is needed

Sulfur is removed from iron and steel when the metals are liquid Although a variety of reagents are employed to remove sulfur, namely, calcium, magnesium, sodium, and rare earths, the most important is calcium Common forms of calcium include the metal; alloys such as calcium silicon (CaSi); the oxide, calcium oxide (CaO); and the carbide, calcium carbide (CaC2) Despite the use of various forms of calcium, the governing chemical reaction in all cases appears to be the same (Ref 13, 14):

The equilibrium constant for the reaction (Eq 9) is:

9

CaS o CaO S

Requirements for Desulfurization. For reasons related to reaction kinetics and thermodynamics, the final sulfur (%S)f concentration achieved in desulfurization processes depends on three factors:

• Initial sulfur concentrations (%S)i

• Amount of desulfurizer used, usually expressed as the weight fraction of desulfurizer to metal, W

• Extraction efficiency of the desulfurizer, which is measured by the desulfurization ratio (DR), that is, the ratio of sulfur concentrations: desulfurizer to metal

The three variables can be related as follows through a mass balance on sulfur:

(% ) (% )

i f

S S

W DR

=

For liquid-desulfurizing slags that are not saturated with respect to calcium sulfide (CaS), the maximum value of DR, that

is, the equilibrium value, can be predicted from thermodynamic considerations:

14 max

15 max

and fS is the activity coefficient for 1 wt% S in the standard state

Equation 12 can be derived from Eq 10 Using known input and desired output sulfur values, Eq 11 gives the required desulfurization ratios as a function of weight fraction desulfurizer These data are plotted in Fig 10 for two cases In the electric-melting case, initial and final sulfur concentrations were assumed to be 0.03 and 0.008% S, respectively, In cupola-melting case, the equivalent concentrations were 0.10% S and 0.008% S The cupola line applies for both cupola iron and ladle-desulfurized cupola iron

Fig 10 Weight fraction desulfurizer required to achieve given desulfurization ratios Plots for electric-melted

iron and steel (curve A) assume initial sulfur concentrations [%(S)i = 0.03] different from those of melted iron [(%S)i = 0.10] (curve B)

cupola-The relationships illustrated in Fig 10 are useful in defining systems that will provide the necessary desulfurization conditions Four systems are compared in Table 3 These cover cupola- and electric-melted cast iron and electric-melted steel Also examined are two liquid slag systems: a basic cupola slag (dicalcium silicate saturated) and a CaO-saturated CaO-Al2O3 slag

Table 3 Theoretical weight ratios: desulfurizer to iron to achieve 0.008% S in various systems

Type of melt Slag composition T, K C S (×

10 4 ) (a)

fS hO (×

10 4 ) (b)

(DR) max (c)

CaO sat -Al 2 O 3 1773 59 5 5.0 417 0.027

C cast iron electric CaO sat -Al 2 O 3 1773 59 5 5.0 417 0.0065

D steel electric CaO sat -Al 2 O 3 1873 316 1 0.45 5280 0.00052

(a) CS is obtained from optical basicity correlations (Ref 15) fS is based on iron composition (Ref 16)

(b) Case A: hO is governed by Si/SiO 2 equilibrium based on respective concentration in iron and slag (Ref 17) Cases B and C: hO is governed by Si/SiO2 equilibrium with aSiO2 = 1 due to ladle exposure to air (Ref 18) Case D: hO is governed by Al/Al2O3 equilibrium with % Al = 0.05 (Ref 19), assumed no contact with air

(c) K14 and K15 data from Ref 20

Table 3 shows that:

• The best desulfurizing cupola slags have lower sulfide capacity and desulfurization ratio than slags used

in ladle desulfurization; the consequence is the need for much larger quantities of slag*

• Compared to steel, cast iron desulfurization benefits from higher fS because of the presence of relatively high concentrations of carbon and silicon in cast iron

• Ladle desulfurization systems that are exposed to air suffer higher hO and, as a result, poorer desulfurization than might otherwise be anticipated

For the cupola slag case in Table 3, the thermodynamically predicted value of (DR)max = 90 is in good agreement with measured data (Fig 11) This indicates that the cupola desulfurization process operates close to equilibrium levels Further evidence for this is given Fig 12, which plots desulfurization data for a cupola operated with varying amounts of

municipal ferrous refuse in the charge The upper portion of Fig 12 plots CS and oxygen activity The latter is expressed

as the partial pressure of oxygen The lower portion of Fig 12 is a comparison of (DR), measured and calculated The good agreement found is evidence that near-equilibrium conditions existed

Fig 11 Desulfurization ratio-basicity comparisons for cupola (closed circles) and laboratory data (open circles)

Equilibrium values are indicated by the angled line The vertical line indicates the slag basicity above which slags are saturated at 1500 °C (2730 °F), with respect to dicalcium silicate This is the point at which the observed DR should equal (DR) max Source: Ref 14

Fig 12 Slag sulfide capacities, oxygen activities, and desulfurization ratios, measured and calculated, for a

cupola operated with municipal ferrous refuse as a charge material Source: Ref 21

For the cases discussed above, the oxygen activity in cupola iron has been found to be governed by the reaction (Ref 17, 21):

Therefore, the overall cupola desulfurization reaction is:

Considerably lower sulfur could be achieved (Ref 14) if hO were governed by:

However, equilibrium for this reaction has not been observed

This discussion has concerned CaS-unsaturated slags However, a desulfurizing slag, saturated with CaS, can in many cases continue to desulfurize as long CaO is present In this case, the ultimate sulfur levels are not as low as those for

CaS-unsaturated slags Nevertheless, CaS-saturated slags can produce sulfur levels that are more than sufficient for cast iron and steel applications In fact, CaS-saturated conditions are often employed in industrial applications because much less desulfurizer is required

In addition to the CaS-saturated liquid slags, totally solid slags such as CaO and CaC2 are in the same category because CaS does not form solid solutions with these materials Also in this category are desulfurizers containing small amounts

of liquid phase These materials possess the desirable properties of liquid and solid slags That is, they combine the fast reaction rates of liquid slags with the large desulfurizing capacities (high CaO concentration) of solid slags

The final sulfur concentrations attainable under CaS-saturated conditions can be obtained from Eq 10 by setting aCaS = 1

For ladle desulfurization, the most important industrial desulfurizers are also CaO-saturated, that is, aCaO = 1 Applying the condition of double saturation to Eq 10 yields:

9

(% )

s

ho S

Fig 13 Equilibrium sulfur concentrations for CaO desulfurization calculated with Eq 19 and 20, assuming aCaO =

aCaS = aSiO2 = 1 Source: Ref 14

Fig 14 Comparisons of the sulfur concentrations in production-continuous desulfurization using CaC2 with the sulfur concentrations from Eq 19 and 20 Open circles indicate electric-melted iron Closed circles indicate cupola iron Source: Ref 14

High desulfurization rates are needed for effective desulfurization in the short times required For CaO-base desulfurizers, the presence of relatively small amounts of liquid phase (<25 vol%) significantly increases the rate of desulfurization This is illustrated in Fig 15, in which the rates of desulfurization by CaO with varying amounts of calcium fluoride (CaF2) are compared The faster rates obtained with CaF2 additions were due to the formation of a CaO-CaF2 liquid phase that provided a path for CaS diffusion from the reaction interface into the porous CaO particle This prevented the normal rapid development of an impervious CaS coating on the CaO surface (Ref 23, 24) A similar explanation, involving liquid-phase formation, was used to account for the higher CaO desulfurization rates of steel when aluminum was concurrently added (Ref 25, 26)

Fig 15 Desulfurization rates of carbon-saturated iron, containing 0.4% Si, with CaO and varying amounts of

CaF 2 at 1450 °C (2640 °F) Source: Ref 22

Another rate-controlling variable in ladle desulfurization is bath agitation For gas-stirred melts, the desulfurization rate

constant k is a function of gas flow rate Q• , that is, k ∝ Q• n

For well-dispersed solid-liquid mixtures, the rate of diffusion-controlled interphase reaction is a relatively weak function of Q• , that is, n ~0.2 to 0.4 (Ref 27, 28, 29) For a

poorly dispersed system, such as a ladle of iron with a cover slag of desulfurizer, agitation has a much greater influence

on the rate constant, with n typically in the range of 1.0 to 2.5 (Ref 27, 28, 29) This is due to the entrainment of

increasing amounts of top slag into the liquid metal with increasing Q• This effect is illustrated in Fig 16, which plots the apparent mass transfer coefficient for desulfurization versus the gas flow rate

Fig 16 Dependence of CaO desulfurization rate constant on the rate of gas flow through a reactor Source: Ref

29

Deoxidation. In the manufacture of steel, FeO-saturated conditions (~0.2% O) are approached as a result of oxygen blowing to remove carbon and silicon as the respective oxides To make a steel product, the oxygen levels are subsequently lowered to 0.005 to 0.02% O by reactive alloy additions of electropositive elements The low oxygen concentrations are necessary to maintain the number of oxide inclusions at a suitable low level, to prevent the formation

of CO blow-holes, and to ensure effective desulfurization

For single-element deoxidation, expressed by:

x y

x x y

M O

M O O

The calculated equilibrium values of oxygen solubility in liquid iron at 1600 °C (2910 °F), based on Eq 22, are given in

Fig 17 for several elements as a function of the concentration of the element, assuming aMxOy = 1 As seen among the common deoxidizers, aluminum produces the lowest oxygen Although not shown, rare-earth metals produce even lower oxygen (Ref 30) Figure 17 plots data for oxygen concentration and oxygen activity In all cases, oxygen activity decreases with increasing levels of deoxidizer, but the oxygen concentration can go through a minimum This occurs in cases where the interaction coefficient eO M is a large negative number, and as a result, fO decreases significantly even at

relatively low concentrations of M To obtain exact values for [%O] or hO, equations for K22 can be found in Ref 13 and

20

Fig 17 Deoxidation equilibria in liquid iron alloys at 1600 °C (2910 °F) Source: Ref 30

As indicated by Eq 22, the oxygen concentration can be reduced if a complex deoxidation product is formed so that

aMxOy < 1 Common deoxidation systems of this type are Si-Mn, Al-Si-Mn, or Al-CaO An advantage, in addition to lower oxygen, is that less deoxidizer is required in solution to achieve a given level of oxygen This is illustrated in Fig

18, which plots the oxygen activity of iron as a function of the concentration of CaO in the calcium aluminate inclusion Separate lines are given for aluminum concentrations ranging from 0.001 to 0.05% Al Also indicated is the oxygen activity when pure aluminum oxide (Al2O3) is the reaction product The data in Fig 18 show that an oxygen activity of 4 ppm can be produced at three aluminum levels 0.002%, 0.005%, or 0.01% depending on whether the respective deoxidation product was CaO-saturated calcium aluminate, CaAl2O4-saturated calcium aluminate, or Al2O3

Fig 18 Concentration of oxygen and aluminum in liquid steel in equilibrium with calcium aluminate at 1600 °C

(2910 °F); arrows indicate oxygen values when the reaction product is pure Al 2 O 3 Source: Ref 31

The levels of total oxygen measured in steel melts considerably exceed equilibrium levels (Ref 20) Two possible explanations are the kinetic limitations in the deoxidation reaction and the ineffective separation of deoxidation products from the melt Consideration of the first possibility suggests that the most important kinetic limitation would be oxygen and/or deoxidant solute diffusion to inclusions in the melt (Ref 20) Considering the presence of 105 to 107 inclusion particles in a cubic centimeter of melt, deoxidation to equilibrium levels of 10 to 20 ppm would require a few minutes Appreciably longer times would be needed to reduce oxygen to less than 1 ppm, as with rare-earth additions For most cases, oxide formation appears faster than oxide particle separation from the bath Thus, for applications that depend on dissolved oxygen, such as desulfurization or CO blowhole formation, equilibrium oxygen values can be assumed to exist after relatively short reaction periods For considerations of the inclusion content of steel, the kinetics of particle separation need to be considered

The limiting case of particle flotation at velocities obtained from Stokes' law calculations is shown in Fig 19 Inclusion size versus time of flotation is plotted The particles produced during deoxidation are generally small (<5 μm, or 200 μin.); accordingly, separation times are long The times indicated in Fig 19 are representative of conditions in a quiescent bath Times are considerably shorter for stirred melts because articles grow by collision and coalescence to sizes of 30 to

100 μm (Ref 20) The efficiency of coalescence of particles is material dependent The energy of adhesion of particles is determined by the wetting characteristics of the particle, measured by wetting angle For example, the force of adhesion

of Al2O3 particles with a wetting angle of 140° is twice that for SiO2 with a wetting angle of 115° On this basis, Al2O3inclusions are expected to cluster more easily than SiO2 inclusions, a phenomenon that has been observed in practice (Ref 20)

Fig 19 Calculated time of flotation of inclusions in stagnant melts as a function of inclusion size Melt depth: A,

50 mm (2 in.); B, 500 mm (20 in.); C, 2000 mm (80 in.) Source: Ref 20

Aluminum Melts

The principal alloying elements in all cast irons, namely, carbon and silicon, are always the same, but the situation for aluminum is different because there are a number of different important aluminum alloy systems, such as Al-Si, Al-Cu, Al-Mg, and Al-Zn (Ref 32) As a result, the preparation of casting alloys from aluminum scrap holds greater risk of contamination than cast iron alloy preparation In addition to the different elements present in aluminum alloy classes, minor elements are commonly present that are beneficial to some alloy systems but harmful to others These elements must also be controlled in alloy preparation

An example of minor element control is the case of silicon modifiers: sodium, strontium, antimony, and phosphorus These elements favorably alter the nucleation and growth kinetics of silicon in aluminum-silicon alloys to produce a more compact form of the precipitate Contamination of sodium- or strontium-modified alloys with alloys containing phosphorus or antimony negates the modification Therefore, control must be exercised over these elements

Another contaminant that can have positive and negative impact on aluminum castings is hydrogen Because aluminum reacts readily with atmospheric moisture, hydrogen contamination is common in aluminum alloys (Ref 32) Although the solubility of hydrogen in liquid aluminum is only of the order of 1.0 ppm, most of the hydrogen precipitates during solidification, producing 0.03 volume fraction of gas porosity This is a serious problem for castings having high strength requirements On the other hand, for castings that are not subjected to high stress, hydrogen-generated porosity is often a desirable condition because it is used to counter the large solidification shrinkage (0.065 volume fraction) associated with aluminum castings

The control of cast aluminum alloy composition is difficult Few but the largest foundries attempt to process purchased scrap; instead they rely on secondary alloy producers with special equipment to perform this function (Ref 33) Because

of the wide range of elements involved in aluminum alloy casting, the first line in composition control is the separation of scrap into alloy type and the accurate characterization of the composition of the batch by chemical analysis Next, compatible batches are combined, and necessary alloy additions are made This process avoids the difficult task of removing unwanted elements Chemical purification is commonly practiced for hydrogen and for elements that are more electropositive than aluminum, namely, the alkali and alkaline earth elements These separations are usually performed in

a single process, with small bubbles of inert or reactive gas removing the undesired elements by evaporation and/or oxidation (Ref 34)

Refining Aluminum by Evaporation Treatment. Some elements dissolved in aluminum have higher vapor

pressures than aluminum and can therefore be separated from aluminum melts by inert gas flushing or vacuum treatment For the evaporation reaction:

M K

h

and for the condition hM = 1, that is, at the hypothetical 1 wt% standard state, the vapor pressure is only a function of K

Using the thermodynamic data given in the article "Thermodynamic Properties of Aluminum-Base and Copper-Base Alloys" in this Volume, the vapor pressures of selected elements were calculated The results are plotted in Fig 20 The vapor pressure of aluminum is also shown because it represents a practical lower limit to which impurity vapor pressures can be reduced without incurring significant losses of aluminum Making this assumption with regard to limiting the impurity vapor pressure, it can be seen in Table 4 that elements with higher vapor pressures than calcium can be effectively removed

Table 4 Approximate theoretical minimum metal compositions to be realized by the vacuum treatment of aluminum

Fig 20 Calculated vapor pressures of selected elements dissolved in aluminum at the hypothetical 1 wt%

standard state Source: Ref 35

Volatile impurities approach their equilibrium vapor pressures in purge gas bubbles (Ref 35); therefore, the rate of impurity removal depends on the vapor pressure of the element and the volumetric gas flow rate (see the article "Gases in Metals" in this Volume) For elements other than hydrogen, the volumetric ratio of purge gas to impurity vapor is so large that the economics of the process are jeopardized (Ref 35) For alkali and alkaline earth metals, this situation is rectified

by adding a reactive gas to the purge gas

Refining Aluminum With Reactive Gases. Alkali and alkaline earth metals form more stable halides and oxides

than aluminum; therefore, by adding F2, Cl2, or O2 to the purge gas, these elements can be separated from aluminum Because the reaction products in this case are condensed phases, the rate of impurity removal no longer depends on but

on the gas/liquid interfacial area, where the reactions take place For the generalized reaction between a metal M and

halogen or oxygen X2 gas:

M x

M X K

26

y m a

m mx x

M

M X p

If px2 for an impurity element is less than px2 for the corresponding reaction with aluminum, then impurity removal is

theoretically possible Conversely, if p x2 for the impurity element is greater than p x2 for aluminum, the aluminum

compound will separate in preference to the impurity Figure 21 and 22 plot p x2 versus hM, respectively, for reactions with chlorine and fluorine The oxide data were not included, because oxygen is much less effective than the halides (Ref 35)

The data in Fig 21 and 22 assume aM mX2mx = 1

Fig 21 Calculated equilibria for the ternary Al-M-Cl system at 1000 K Source: Ref 35

Fig 22 Calculated equilibria for the ternary Al-M-F system at 1000 K Source: Ref 35

Table 5 compares the minimum impurity concentrations that are possible by treatment with Cl2, F2, and O2 For all the cases examined, chlorine appears to be the best reagent Chlorine is the most commonly used reagent for removing these elements from aluminum melts In practice, the thermodynamically predicted levels are not achieved (Ref 35)

Table 5 Approximate theoretical minimum contents to be realized by refining with reactive gases

to the MgCl2 produced by reactive gas separations

Fig 23 Schematic of an electrochemical magnesium separation apparatus Source: Ref 36

The open-circuit cell potential ε for this system can be estimated from the Nernst equation:

(Eq 28)

where is Faraday's constant and (aMg)cathode = 1 For producing alloys with less than 0.1% Mg, as required for a number

of important casting alloys, such as 319 and 380, Eq 28 predicts that a cell potential of 0.32 V is required (Ref 36) This voltage must not be greatly exceeded, because aluminum will be transferred from anode to cathode at a potential of 0.5 V

References cited in this section

13 R.J Fruehan, Ladle Metallurgy: Principles and Practices, Iron and Steel Society of AIME, 1985, p 7

14 S Katz and C.F Landefeld, Cupola Desulfurization, in Cupola Handbook, American Foundrymen's

Society, 1984, p 351-363

15 I.D Sommerville, The Capacities and Refining Capabilities of Metallurgical Slags, in Foundry Processes:

Their Chemistry and Physics, S Katz and C.F Landefeld, Ed., Plenum Press, 1988, p 101-133

16 G.K Sigworth and J.F Elliott, The Thermodynamics of Liquid Dilute Iron Alloys, Met Sci., Vol 8, 1974, p

298-310

17 S Katz and H.C Rezeau, The Cupola Desulfurization Process, Trans AFS, Vol 87, 1979, p 367-376

18 S Katz, D.E McInnis, D.L Brink, and G.A Wilkinson, The Determination of Aluminum in Malleable Iron

From Measured Oxygen, Trans AFS, Vol 88, 1980, p 835-844

19 R.J Fruehan, Ladle Metallurgy: Principles and Practices, Iron and Steel Society of AIME, 1985, p 8

20 E.T Turkdogan, Physical Chemistry of High Temperature Technology, Academic Press, 1980

21 S Katz and V.R Spironello, Effect of Charged Aluminum on Iron Temperature, Silicon Recovery and

Desulfurization in an Iron-Producing Cupola, Trans AFS, Vol 92, 1984, p 161-172

22 C.F Landefeld and S Katz, Kinetics of Iron Desulfurization by CaO-CaF2, in Proceedings of the Fifth

International Iron and Steel Congress, Vol 6, Iron and Steel Society of AIME, 1986, p 429-439

23 S Katz and C.F Landefeld, Plant Studies of Continuous Desulfurization with CaO-CaF2-C, Trans AFS,

Vol 93, 1985, p 215-228

24 S Katz and B.L Tiwari, A Critical Overview of Liquid Metal Processing in the Foundry, in Foundry

Processes: Their Chemistry and Physics, S Katz and C.F Landefeld, Ed., Plenum Press, 1988, p 1-52

25 J Niederinghaus and R.J Fruehan, Desulfurization Mechanisms for CaO-Al and CaO-CaS in Carbon

Saturated Iron, Metall Trans B, to be published

26 E.T Turkdogan, Physiochemical Phenomena of Mechanisms and Rates of Reaction in Melting, Refining

and Casting of Foundry Irons, in Foundry Processes: Their Chemistry and Physics, S Katz and C.F

Landefeld, Ed., Plenum Press, 1988, p 53-100

27 S Asai and I Muchi, Fluid Flow and Mass Transfer in Gas Stirred Ladles, in Foundry Processes: Their

Chemistry and Physics, S Katz and C.F Landefeld, Ed., Plenum Press, 1988, p 261-292

28 S.-H Kim and R.J Fruehan, Physical Modelling of Liquid/Liquid Mass Transfer in Gas Stirred Ladles,

Metall Trans B, Vol 18B, 1987, p 381-390

29 J Ishida, K Yamaguchi, S Sugiura, N Demukai, and A Notoh, Denki Seiko, Vol 52, 1981, p 2-8

30 E.T Turkdogan, Ladle Deoxidation, Desulfurization and Inclusions in Steel Part I: Fundamentals, Arch

Eisenhüttenwes., Vol 54, 1983, p 1-10

31 E.T Turkdogan, Slags and Fluxes for Ferrous Ladle Metallurgy, Ironmaking Steelmaking, Vol 12, 1985, p

64-78

32 Aluminum Casting Technology, American Foundrymen's Society, 1986

33 Recycled Metals in the 1980's, National Association of Recycling Industries, 1982

34 J.H.L Van Linden, R.E Miller, and R Bachowski, Chemical Impurities in Aluminum, in Foundry

Processes: Their Chemistry and Physics, S Katz and C.F Landefeld, Ed., Plenum Press, 1988, p 393-409

35 G.K Sigworth and T.A Engh, Refining of Liquid Aluminum A Review of Important Chemical Factors,

Scand J Metall., Vol 11, 1982, p 143-149

36 B.L Tiwari and R.A Sharma, Electrolytic Removal of Magnesium From Scrap Aluminum, J Met., Vol 36

(No 7), 1984, p 41-43

Note cited in this section

* The actual differences are less than those indicated Calcium sulfide has only limited solubility in CaO-Al2O3

slags As a result, a minimum W = 0.01 is needed to maintain the CaS-unsaturated condition

References

1 R.I.L Guthrie, Addition Kinetics in Steelmaking, in Proceedings of the 35th Electric Furnace Conference,

Iron and Steel Society of AIME, 1978, p 30-41

2 F.A Mucciardi and R.I.L Guthrie, Aluminum Wire Feeding in Steelmaking, Trans ISS, Vol 3, 1983, p

53-59

3 J.W Robison, Jr., "Ladle Treatment With Steel-Clad Metallic Calcium Wire," Paper 35, presented at

Scaninject III, Part II, MEFOS, Lulea, Sweden, 1983

4 L Kalvelage, J Markert and J Pötschke, Measurement of the Dissolution of Graphite in Liquid Iron by

Following the Buoyancy, Arch Eisenhüttenwes., Vol 50, 1979, p 107-110

5 R.G Olsson, V Koump, and T.F Perzak, Rate of Dissolution of Carbon in Molten Fe-C Alloys, Trans

Met Soc AIME, Vol 236, 1966, p 426-429

6 O Angeles, G.H Geiger, and C.R Loper, Jr., Factors Influencing Carbon Pickup in Cast Iron, Trans AFS,

Vol 74, 1968, p 3-11

7 M Eisenberg, C.W Tobias, and C.R Wilke, Mass Transfer at Rotation Cylinders, Chem Eng Prog

Symp Series, Vol 51, 1955, p 1-16

8 R.G Olsson, V Koump, and T.F Perzak, Rate of Dissolution of Carbon Steel in Molten Iron-Carbon

Alloys, Trans Met Soc AIME, Vol 233, 1965, p 1654-1657

9 R.D Pehlke, P.D Goodell, and R.W Dunlap, Kinetics of Steel Dissolution in Molten Pig Iron, Trans

Met Soc AIME, Vol 233, 1965, p 1420-1427

10 R.I.L Guthrie and P Stubbs, Kinetics of Scrap Melting in Baths of Molten Pig Iron, Can Metall Q., Vol

12, 1973, p 465-473

11 K Mori and T Sakuraya, Rate of Dissolution of Solid Iron in Carbon-Saturated Liquid Iron Alloys With

Evolution of CO, J Iron Steel Inst Japan, Vol 22, 1982, p 964-990

12 P.T.L Brian and H.B Hales, Effects of Transpiration and Changing Diameter on Heat and Mass Transfer

to Spheres, AIChE J., Vol 15, 1969, p 419-425

13 R.J Fruehan, Ladle Metallurgy: Principles and Practices, Iron and Steel Society of AIME, 1985, p 7

14 S Katz and C.F Landefeld, Cupola Desulfurization, in Cupola Handbook, American Foundrymen's

Society, 1984, p 351-363

15 I.D Sommerville, The Capacities and Refining Capabilities of Metallurgical Slags, in Foundry Processes:

Their Chemistry and Physics, S Katz and C.F Landefeld, Ed., Plenum Press, 1988, p 101-133

16 G.K Sigworth and J.F Elliott, The Thermodynamics of Liquid Dilute Iron Alloys, Met Sci., Vol 8, 1974,

p 298-310

17 S Katz and H.C Rezeau, The Cupola Desulfurization Process, Trans AFS, Vol 87, 1979, p 367-376

18 S Katz, D.E McInnis, D.L Brink, and G.A Wilkinson, The Determination of Aluminum in Malleable

Iron From Measured Oxygen, Trans AFS, Vol 88, 1980, p 835-844

19 R.J Fruehan, Ladle Metallurgy: Principles and Practices, Iron and Steel Society of AIME, 1985, p 8

20 E.T Turkdogan, Physical Chemistry of High Temperature Technology, Academic Press, 1980

21 S Katz and V.R Spironello, Effect of Charged Aluminum on Iron Temperature, Silicon Recovery and

Desulfurization in an Iron-Producing Cupola, Trans AFS, Vol 92, 1984, p 161-172

22 C.F Landefeld and S Katz, Kinetics of Iron Desulfurization by CaO-CaF2, in Proceedings of the Fifth

International Iron and Steel Congress, Vol 6, Iron and Steel Society of AIME, 1986, p 429-439

23 S Katz and C.F Landefeld, Plant Studies of Continuous Desulfurization with CaO-CaF2-C, Trans AFS,

Vol 93, 1985, p 215-228

24 S Katz and B.L Tiwari, A Critical Overview of Liquid Metal Processing in the Foundry, in Foundry

Processes: Their Chemistry and Physics, S Katz and C.F Landefeld, Ed., Plenum Press, 1988, p 1-52

25 J Niederinghaus and R.J Fruehan, Desulfurization Mechanisms for CaO-Al and CaO-CaS in Carbon

Saturated Iron, Metall Trans B, to be published

26 E.T Turkdogan, Physiochemical Phenomena of Mechanisms and Rates of Reaction in Melting, Refining

and Casting of Foundry Irons, in Foundry Processes: Their Chemistry and Physics, S Katz and C.F

Landefeld, Ed., Plenum Press, 1988, p 53-100

27 S Asai and I Muchi, Fluid Flow and Mass Transfer in Gas Stirred Ladles, in Foundry Processes: Their

Chemistry and Physics, S Katz and C.F Landefeld, Ed., Plenum Press, 1988, p 261-292

28 S.-H Kim and R.J Fruehan, Physical Modelling of Liquid/Liquid Mass Transfer in Gas Stirred Ladles,

Metall Trans B, Vol 18B, 1987, p 381-390

29 J Ishida, K Yamaguchi, S Sugiura, N Demukai, and A Notoh, Denki Seiko, Vol 52, 1981, p 2-8

30 E.T Turkdogan, Ladle Deoxidation, Desulfurization and Inclusions in Steel Part I: Fundamentals, Arch

Eisenhüttenwes., Vol 54, 1983, p 1-10

31 E.T Turkdogan, Slags and Fluxes for Ferrous Ladle Metallurgy, Ironmaking Steelmaking, Vol 12, 1985, p

64-78

32 Aluminum Casting Technology, American Foundrymen's Society, 1986

33 Recycled Metals in the 1980's, National Association of Recycling Industries, 1982

34 J.H.L Van Linden, R.E Miller, and R Bachowski, Chemical Impurities in Aluminum, in Foundry

Processes: Their Chemistry and Physics, S Katz and C.F Landefeld, Ed., Plenum Press, 1988, p 393-409

35 G.K Sigworth and T.A Engh, Refining of Liquid Aluminum A Review of Important Chemical Factors,

Scand J Metall., Vol 11, 1982, p 143-149

36 B.L Tiwari and R.A Sharma, Electrolytic Removal of Magnesium From Scrap Aluminum, J Met., Vol

An example of a chemical reaction evolving gas and causing porosity is the reaction of moisture (H2O) in the sand with elements in the cast iron, such as carbon, silicon, aluminum, or iron itself For example, aluminum in cast iron often causes porosity problems, and the reaction responsible for the gas evolution is:

where the underlining of aluminum (Al) indicates that it is dissolved in the metal The free energy for reaction (Eq 1) is

highly negative even for very low concentrations of aluminum, indicating that the thermodynamics for the reaction are highly favorable Similar reactions can be written for silicon, carbon, and iron For carbon, carbon monoxide is also evolved, according to the reaction:

For the casting of aluminum, the reaction is even more favorable Whether or not the reaction will actually occur depends

on many complex factors, including liquid-phase mass transfer and surface tension; therefore, its occurrence is difficult to predict However, if there is excessive moisture in the sand, the reaction is highly likely to occur

The other major cause of gas porosity is the evolution of dissolved gases during casting For example, liquid cast iron may have dissolved hydrogen and nitrogen The solubility of these gases in the solid may be less than that in the liquid, and the gases may therefore be evolved during solidification Whether or not this will occur depends on the amount of

hydrogen and nitrogen present, the alloy being cast, chemical kinetics, and the surface tension of the alloy Similarly, hydrogen can dissolve in liquid aluminum, but its solubility is much lower in solid aluminum and can cause porosity For copper, the evolution of hydrogen, water vapor, or carbon monoxide could cause gas porosity

In the article, the solubilities of the common gases present in cast iron, aluminum, and copper will be reviewed The kinetics of the relevant reactions and the reactions during solidification will be discussed Finally, possible methods of control or removal of the dissolved gases will be analyzed The discussion will primarily focus on cast iron because more

is known about the thermodynamics and kinetics of iron than other elements However, the same basic principles also apply to aluminum and copper

Cast Iron (Ref 1)

Solubility of Hydrogen and Nitrogen. The gases in cast iron that can cause porosity are hydrogen and nitrogen Castings from iron produced in a cupola are nearly saturated with nitrogen; therefore, the presence of nitrogen a is a major concern (Ref 2) Hydrogen and nitrogen dissolve in liquid iron alloys as atomic species according to:

in Fig 1 as a function of carbon content The solubility of nitrogen at 1823 K decreases from about 450 ppm for pure iron

to about 80 ppm for an alloy containing 4.5% C The solubilities of hydrogen for the same metals are 24 and 10 ppm, respectively The hydrogen solubility decreases with temperature and at eutectic temperature is 6.5 ppm; temperature has only a small effect on the solubility of nitrogen Silicon decreases the solubility of nitrogen even further For a 4C-1Si melt, the solubility of nitrogen is about 75 ppm

Fig 1 Solubility of hydrogen and nitrogen (1 atm) in iron-carbon alloys at 1823 K in the liquid and 1500 K for

austenite

When the cast iron solidifies, it forms austenite and graphite or cementite The solubility of nitrogen in austenite (Ref 3) is also shown in Fig 1 For austenite containing about 1.8% C, the solubility is about 150 ppm Therefore, when the liquid

cast iron freezes, the solubility in the solid is higher than that in the liquid; this interesting phenomenon will be discussed later in detail There is some solubility of nitrogen in cementite (Ref 4) The exact amount is not known, but is about the same as that in austenite The solubility of hydrogen in austenite is about 7 ppm; therefore, the solubility of hydrogen in the liquid and solid are about equal, and there is little segregation of hydrogen during solidification (Ref 3)

Kinetics of Gas-Liquid Reactions. The absorption of hydrogen and nitrogen is controlled by one of the following steps:

• Diffusion of the gas to the surface

• Chemical reaction on the surface

• Diffusion of the element in the liquid away from the surface

For the desorption of a gas, the same three steps are important but occur in reverse order The process need not to be controlled by only one of the above steps; it can be controlled by two or more steps in series For liquid cast iron, nitrogen

is the more important of the gases, and it will be considered in detail

Gas diffusion is usually not rate controlling for nitrogen absorption from the atmosphere because of the pressure of N2; consequently, the driving force for diffusion is high As a result, one of the other steps is slower and rate controlling However, for nitrogen or hydrogen removal by an inert gas, diffusion on N2 or H2 must be considered The flux of

nitrogen J away from the surface is given by:

p is the pressure of nitrogen in the bulk gas, and

2

S N

p is the surface pressure

of nitrogen in equilibrium with the melt where:

2

S N

where KN is the equilibrium constant for Eq 4 and fN is the activity coefficient of nitrogen with respect to 1 wt% However, in general, gas diffusion for these reactions is faster than the chemical reaction or liquid-phase diffusion and can be neglected in most cases

The rate of the chemical reaction is controlled by the dissociation of the nitrogen molecule on the surface, and the

rate is given by:

p is the equilibrium nitrogen pressure given by an expression similar to Eq 6, and (1 - θ) is the fraction of vacant sites not occupied by surface-active elements Certain elements are surface active on liquid iron in that they lower the surface tension of iron by covering most of the surface For example, oxygen and sulfur are surface active on iron, and for bulk concentrations of 0.03% O or 0.03% S, over 90%

of the surface sites will be covered by oxygen or sulfur These elements therefore retard the rate of chemical reaction At

high coverage, the surface-active element (1 - θ) is inversely proportional to the activity a of the element Therefore, the

rate is given by:

where kA is related to the adsorption coefficient of the element on the surface and the quantity kBkA/a is the overall rate constant k For dilute solutions, the activity of the surface-active element is proportional to its concentration Therefore,

the overall rate constant is inversely proportional to the weight percent of the surface-active element in iron

Liquid-Phase Mass Transfer If the rate is controlled by liquid-phase mass transfer, the flux of nitrogen atoms is

given by:

where kmN is the liquid-phase mass transfer coefficient of nitrogen and CBNand CSNare the bulk and surface concentrations

of nitrogen The integrated form of Eq 7 is:

The rate of the nitrogen reaction with iron alloys containing oxygen, sulfur, chromium, and other elements has been measured by many investigators, and there is good agreement in most cases (Ref 6, 7, 8, 9, 10, 11) However, the rate for carbon-saturated iron containing other elements was not investigated until recently (Ref 12)

In this work, an isotope exchange technique was used to measure the rate of dissociation of the nitrogen molecule (N2) In particular, the effects of carbon, sulfur, phosphorus, lead, tin, bismuth, and tellurium on the rate were investigated

Sulfur, phosphorus, lead, bismuth, and tellurium decreased the rate, while carbon and tin had no significant effect For example, the effect of sulfur is shown in Fig 2 as the rate constant versus the reciprocal of the activity of sulfur (Ref 11) The activity of sulfur is in weight percent, and the activity coefficient of sulfur in carbon-saturated iron is 6.3 The rate for carbon-saturated iron with no sulfur is about 10-5 mol/cm2 · s · atm at 1450 °C (2640 °F) For example, for 0.009% S (1/as

= 18), the rate is decreased to 8 × 10-7 mol/cm2 · s · atm Bismuth and tellurium had an even larger effect (Ref 12), as shown in Fig 3 and 4 The activities are relative to pure bismuth and tellurium, respectively, which have larger deviations from ideal behavior As little as 50 ppm Te reduced the rate by 90%, and 50 ppm Bi decreased it by 80%

Fig 2 Effect of sulfur in the rate of the nitrogen reaction on Fe-CSAT-S alloys

Fig 3 Effect of bismuth on the nitrogen reaction with Fe-CSAT-Bi alloys at 1450 °C (2640 °F)

Fig 4 Effect of tellurium on the nitrogen reaction with Fe-CSAT-Te alloys at 1450 °C (2640 °F)

On the other hand, carbon is not surface active and does not affect the rate at moderate sulfur contents (0.017% S) up to 4.5% C (Ref 12), as shown in Fig 5 Therefore, the initial rate of nitrogen formation is primarily controlled by chemical kinetics at the surface, and surface-active elements, such as sulfur and tellurium, can reduce the rate This information can

be helpful in controlling the rate of the reaction For example, if it is desired to remove nitrogen by argon gas flushing, the concentrations of these elements should be as low as possible On the other hand, it may be possible to retard nitrogen evolution during solidification by the deliberate addition of these elements

Fig 5 Effect of carbon on the nitrogen reaction on Fe-C-S alloys at constant sulfur activity

Reactions During Solidification. During the solidification of most simple iron alloys, there is enrichment of the alloying element because there is greater solubility in the liquid than in the solid The difference in the solubilities is

expressed by the partition ratio kS/L, which is determined from the slopes of the solidus to liquidus lines on the phase diagram For most simple alloys, the partition ratio is less than 1, resulting in liquid enrichment during solidification For example, for an Fe-0.02N alloy, the last portion of liquid to solidify will be enriched and will have a concentration of 0.045% N, which is the solubility of nitrogen in liquid iron at 1 atm (Ref 13) Therefore, even though the liquid alloy is far from being saturated with nitrogen, during solidification, the concentration will increase to the point where the solubility is exceeded and nitrogen gas will be evolved

Researchers have measured the partition ratio for nitrogen in high-carbon iron alloys (Ref 4) The partition ratio was found to be 1.9 and 2.2 for stable and metastable eutectic solidification, indicating that there is greater solubility in the solid than in the liquid Therefore, in this case, as the alloy solidifies, nitrogen is actually enriched in the solid, and the nitrogen content decreases in the liquid It may appear that, because there is no enrichment during solidification, the thermodynamic pressure of nitrogen cannot increase The thermodynamic pressure is defined by an expression similar to

Eq 6 Carbon greatly increases the activity coefficient of nitrogen in the liquid and decreases its solubility; therefore, the thermodynamic pressure may increase because the carbon content of the liquid is increasing during solidification

For example, for an alloy containing 3.8% C that is saturated at 1500 °C (2730 °F) with 110 ppm N just prior to eutectic solidification, the concentration in the liquid will decrease to 97.5 ppm because of enrichment in the austenite However, the solubility of nitrogen in the liquid has decreased to 90 ppm primarily because of the increase in the carbon content Consequently, the thermodynamic pressure of nitrogen exceeds 1 atm, and the evolution of nitrogen may cause a pinhole

When determining the possibility of pinhole formation, one must consider the total thermodynamic pressure of all the

gases, and if the total pressure pT exceeds 1 atm, a gas pinhole may result For example, for hydrogen and nitrogen, the total pressure is given by:

where K i is the equilibrium constant for the gas reaction and f i is the activity coefficient At the end of solidification, the concentrations will be as indicated in Eq 11 Even a small amount of hydrogen may be important For example, in the case of an Fe-3.8C alloy containing only 4 ppm H, the hydrogen content in the liquid will increase slightly, because of the enrichment, to about 4.5 ppm just prior to the eutectic reaction However, even for this small amount of hydrogen, its thermodynamic pressure is 0.4 atm Therefore, if the thermodynamic pressure of nitrogen exceeds 0.6 atm, a gas pinhole may result Calculations indicate that iron with as little as 80 ppm N and 4 ppm H may have a gas pressure exceeding 1 atm

Figure 6 shows a plot of the total pressure of hydrogen and nitrogen just prior to final solidification equal to 1 atm as a function of bulk nitrogen and hydrogen contents in an Fe-3.8C alloy If the hydrogen and nitrogen contents are such that they are below the line, the pressure will not exceed 1 atm, and a pinhole due to gas evolution will not occur; if above the line and the pressure exceeds 1 atm, a pinhole from this source is possible

Fig 6 The thermodynamic pressure of nitrogen and hydrogen at the end of solidification equal to 1 atm for an

Fe-3.8C alloy

For pinholes to develop, it is also necessary to nucleate a bubble Bubble nucleation is a complex phenomenon and is influenced by surface tension For melts with low surface tension, bubble nucleation is favored On the other hand, elements that reduce the surface tension also reduce the rate of N2 formation Therefore, it is difficult to predict the net effect on the probability of bubble formation of an alloying element that reduces the surface tension of iron

Nitrogen and Hydrogen Removal by Inert Gas Flushing. The preceding discussion indicates that nitrogen and

hydrogen evolution during solidification may be a cause of pinholes in castings Therefore, it would be desirable to remove a portion of these gases if possible One method is by argon bubbling in the desulfurization reactor or ladle The desulfurization reactor is a continuous reactor with metal flowing in and out almost continuously However, for the present calculation it will be assumed to be a batch reactor Because this is only an order of magnitude calculation, this assumption is not unreasonable

When argon is bubbled through the melt, nitrogen atoms combine on the bubble surface to form N2 The rate is controlled

by the chemical reaction on the surface and the liquid-phase mass transfer of nitrogen to the surface in series For the purpose of the present calculations, the following were assumed:

• 10 Mg (11 ton) reactor 1 m (3.3 ft) deep

• Argon flow rate of 0.005 m3/s (10 scfm) through the melt

• Initial nitrogen content of 100 ppm

• 60 mm (2.4 in.) diam bubbles

The rate constant for iron containing a relatively small amount of sulfur is about 1.5 × 10-6 mol/cm2 · s · atm, and the mass

transfer coefficient km is about 0.1 cm/s The gas velocity, retention time, and surface area were estimated as done previously (Ref 5) Equations 8 and 9 were solved simultaneously, and the results are given in Fig 7 The calculations demonstrate that the rate is truly mixed control, as indicated by the surface concentration being between zero and the bulk concentration The results indicate that it would take about 40 min to remove 20 ppm; if the argon flow were doubled, it would still take over 20 min to remove 20 ppm Although these calculations are rather crude, they indicate that it would

be difficult to remove nitrogen by argon bubbling

Fig 7 Rate of removal of nitrogen from an Fe-3.8C alloy with argon bubbling at 0.005 m3 /s (10 scfm) in a 10

Mg (11 ton) reactor

For hydrogen, the chemical reaction and mass transfer are considerably faster than for nitrogen If equilibrium between the metal and the gas bubbles leaving the system is assumed, it is possible to calculate the amount of hydrogen that can be removed by gas bubbling from thermodynamic considerations alone The equations are developed similarly to those for steel (Ref 13), taking the form:

References cited in this section

1 R.J Fruehan, in Proceedings of the Physical Chemistry of Foundry Processes Symposium (Warren, MI),

General Motors Corporation, 1986

2 S Katz and C Landefeld, General Motors Research, private communication, 1986

3 Making, Shaping and Treating of Steel, 10th ed., United States Steel Corporation, 1985

4 A Kagawa and T Okamoto, Trans Jpn Inst Met., Vol 22 (No 2), 1981, p 137

5 R.J Fruehan, B Lally, and P.C Glaws, in Proceedings of the Fifth International Iron and Steel Congress

(Washington, DC), Iron and Steel Society of AIME, 1986

6 R.D Pelke and J Elliot, Trans TMS-AIME, Vol 227, 1963, p 894

7 M Inouye and T Choh, Trans JISI, Vol 8, 1968, p 134

8 R.J Fruehan and L.J Martonik, Metall Trans B, Vol 11B, 1980, p 615

9 P.C Glaws and R.J Fruehan, Metall Trans B, Vol 16B, 1985, p 551

10 P.C Glaws and R.J Fruehan, Metall Trans B, Vol 17B, 1986, p 317

11 M Byrne and G.R Belton, Metall Trans B, Vol 14B, 1983, p 441

12 F Tsukihashi and R.J Fruehan, submitted to Trans JISI, 1987

13 R.J Fruehan, Ladle Metallurgy: Principles and Practices, Iron and Steel Society of AIME, 1985

Details of the reactions of hydrogen with aluminum alloys and the factors influencing hydrogen removal can be found in Ref 14

Hydrogen Solubility and Reactions During Solidification. Researchers have measured the solubility of hydrogen in cubic centimeters per 100 g of alloy or in weight percent in equilibrium with 1 atm of hydrogen (Ref 15, 16)

These results are given in Fig 8 Alloying elements affect the solubility of hydrogen The solubilities for selected alloys

at 750 °C (1382 °F) are given in Table 1 Generally, the common alloying elements decrease the solubility of hydrogen

Table 1 The solubility of hydrogen in aluminum and its alloys

Alloy Hydrogen solubility,

Fig 8 Solubility of hydrogen (pH2 = 1 atm) in aluminum

Because the solubility of hydrogen is significantly higher in the liquid aluminum as compared with the solid, there will be enrichment of the liquid during solidification Assuming no solid diffusion and complete liquid diffusion, at the end of solidification there will be a large increase in the hydrogen concentration Therefore, even when the hydrogen content is far below the solubility for the bulk liquid during solidification, the concentration in the enriched liquid will increase, and the solubility limit may be exceeded For example, consider pure aluminum containing 0.4 ppm H at 750 °C (1382 °F)

By the time the alloy is 90% solidified, the concentration in the remaining liquid will be about 3.6 ppm H, which will exceed the hydrogen solubility for 1 atm, resulting in the formation of hydrogen bubbles

Inert Gas Flushing. It is possible to remove hydrogen from liquid aluminum by inert gas (argon or nitrogen) flushing

As discussed previously for cast iron, it is possible to calculate the fastest possible rate, or the minimum amount of purge gas needed to remove hydrogen The general form of the equation is:

H o

STP H

kg m k

ppm s

For example, for 10 Mg (2.2 × 104 lb) of aluminum, a gas purge rate of 0.01 m3/s (0.35 ft3/s) for 100 s could reduce the hydrogen from 0.2 to 0.09 ppm As the hydrogen content decreases, it becomes increasingly difficult to remove hydrogen

(Ref 14), as indicated in Fig 9 As the hydrogen content decreases, the ratio R of the purge gas to the hydrogen gas

removed increases from about 15 at 0.4 ppm to over 500 at 0.1 ppm (1 cm3/100 g = 0.9 ppm) It should be emphasized that this calculation indicates the theoretical minimum amount of purge gas required Due to kinetic factors, the amount of purge gas can be significantly higher

Fig 9 Gas removal ratio for equilibrium in aluminum at 760 °C (1400 °F)

There are two limiting cases with respect to hydrogen removal: the thermodynamic limit discussed above and the diffusion in the liquid film boundary layer surrounding the gas bubble For the limiting case of diffusion control, the rate equations are similar to those discussed previously for nitrogen in iron, and the rate can be expressed by:

% ln

%

A mH i

K H

In general, the rate will be controlled by both diffusion and the thermodynamic limit in series In an analysis of the case of mixed control, it was found that the purging efficiency the ratio of the actual amount of gas required to the theoretical minimums given by Eq 14 depends on the bubble size and hydrogen content (Ref 17), as shown in Fig 10 Both the interfacial area and the mass transfer coefficient increase with decreasing bubble size; consequently, the rate of hydrogen removal is enhanced by having small bubbles The use of porous plugs helps improve the rate by providing small bubbles, but the bubbles generally coalesce in the bath, reducing the beneficial effect The use of an impeller with a porous plug improves the rate further by dispersing the bubbles

Fig 10 Purging gas efficiency as a function of bubble size for 0.5 cm3 /100 g and 0.1 cm 3 /100 g of hydrogen

It has also been found that the addition of chlorine (Ref 18) or freon (Ref 19) improves the rate in some cases These

halogen gases increase the mass transfer coefficient km and the rate However, if the rate is being primarily controlled by the thermodynamic limit, the addition of these gases will have little effect It is also possible that the halogen is actually taking part in a hydrogen removal reaction:

Equation 17, under some circumstances, is thermodynamically favorable This will increase the amount of hydrogen that can be present in the inert gas bubble; that is, hydrogen can be present as H2 or hydrogen chloride gas, thus increasing the rate of removal

References cited in this section

14 G.K Sigworth, Trans AFS, 1987, p 73

15 W.R Opie and W.J Grant, Trans AIME, Vol 188, 1950, p 1234

16 C.E Ramsley and H Neufeld, J Inst Met., Vol 74, 1947-1948, p 559

17 G.K Sigworth and T.A Engh, Metall Trans B, Vol 13B, 1982, p 447

18 J Botor, Metal Odlev, Vol 6, 1980 p 21

19 J Botor, Aluminum, Vol 56, 1980, p 519

Copper and Copper Alloys

Gas porosity is a problem in the casting of copper and copper alloys The dissolved gas that is most important is hydrogen However, gaseous compounds such as water vapor, carbon monixide, and sulfur dioxide can also evolve during solidification

Hydrogen Solubility and Reactions. Hydrogen dissolves in copper and copper alloys as hydrogen atoms The solubility in terms of cubic centimeters of H2 per 100 g of metal is shown in Fig 11 (1 cm3/100 g = 0.9 ppm) (Ref 20, 21, 22) Recent work (Ref 23) has confirmed the results of earlier researchers (Ref 20), and these are the current accepted values The most common alloying element for copper is tin, and as shown in Fig 12, tin decreases the solubility of hydrogen (Ref 20) For example, in a 50Sn-50Cu alloy at 1200 °C (2190 °F), the solubility of hydrogen is only about 1 ppm as compared to 6.5 ppm for pure copper On the other hand, nickel increases the solubility of hydrogen For example, for a 10Ni-90Cu alloy, the solubility is about 10 ppm

Fig 11 Solubility of hydrogen in liquid copper as a function of temperature (pH2 = 1 atm)

Fig 12 Solubility of hydrogen in copper and copper-tin alloys (pH2 = 1 atm) as a function of temperature

Hydrogen can enter the copper directly from the hydrogen in the atmosphere, but most likely it enters according to:

Equations 18 and 19 are favored by copper with low oxygen contents If there are high concentrations of alloying elements with more stable oxides than cuprous oxide, such as tin, the alloying element may react to put hydrogen into solution, for example:

Other gases that can cause porosity are sulfur dioxide and carbon monoxide, which also form during solidification according to Eq 21 and 22, respectively:

The thermodynamic pressure of carbon monoxide gas can be very high even at moderate oxygen and carbon contents because the solubility of the carbon in copper is limited and therefore its chemical activity is high

Porosity in Copper Alloys. In copper ingots, the gas porosity usually results from H2 and water evolution The evolution of as little as 1 ppm H results in a gas evolution in 44% of the volume of the metal If the sulfur content exceeds 0.05%, the evolution of sulfur dioxide may also contribute to gas porosity Only if the oxygen content exceeds 0.01% is carbon monoxide considered to be the cause of porosity (Ref 24) Because phosphorus deoxidizes copper significantly, the oxygen content of phosphorized copper is very low and water, carbon monoxide, and sulfur dioxide do not contribute

to porosity; only H2 need be of concern

Copper-zinc alloys rarely have problems associated with gas porosity, primarily because zinc dioxidizes the metal Zinc may also remove dissolved gases because of its high vapor pressure, which results in the zinc vapor flushing hydrogen out

of the melt For copper-tin alloys, hydrogen is the major concern because tin significantly increases the solubility of hydrogen

Degassing of Copper Alloys. The two primary methods of removing dissolved gases from copper alloys are oxidation-reduction and inert gas flushing Although vacuum degassing is theoretically possible, it is rarely used because

it is not cost effective

In oxidation-reduction, the first step is to remove the hydrogen by oxidation using an oxidizing slag or oxygen-rich copper Then, just prior to pouring, the melt is deoxidized by adding phosphorus or other deoxidizers Calcium boride, boron carbide, and lithium have also been used for deoxidation (Ref 25) A charcoal cover is often used for deoxidation and protection of the melt from reoxidation

For copper alloys that contain strong deoxidizers, such as phosphorus, zinc, and tin, it is not possible to remove hydrogen

by oxidation, because these elements will form stable oxides For these alloys, inert gas flushing with nitrogen is commonly used The theoretical minimum amount of inert gas is determined by using relationships similar to those derived for iron and aluminum In practice, 150 to 250 L (5.3 to 8.8 ft3) of nitrogen per 1000 kg (2200 lb) of copper alloy

is generally recommended (Ref 26)

References cited in this section

20 M.B Bever and C.F Floe, Trans AIME, Vol 156, 1944, p 149

21 A Sieverts and W Krumbharr, Ztch Phys Chem., Vol 74, 1910, p 277

22 P Roentgren and F Moeller, Metallwertschaft, Vol 13, 1934, p 81

23 E Kato, H Ueno, and T Orimo, Trans Jpn Inst Met., Vol 11, 1970, p 351

24 R.H Waddington, J Inst Met., 1948-1949, p 311

25 E.R Thews, Metall., June 1956, p 431

26 W.A Baker and F.C Child, J Inst Met., Vol 70, 1944, p 349

Overcoming Gas Porosity

Gas in cast iron, aluminum, and copper can be a major problem The porosity results from reactions with the environment

or the evolution of gases during solidification In determining if gas evolution can cause porosity, the thermodynamics for the reactions must be favorable In this article, the solubilities of the common gases and the reactions during solidification were discussed Because of enrichment during solidification, the concentration of the elements increases to the extent that the thermodynamic pressure of the dissolved gases may exceed 1 atm In addition to the thermodynamics, kinetic factors and interfacial energies determine if the gases will evolve

The most common method for removing hydrogen from aluminum and copper, which can also be used for removing hydrogen and nitrogen from cast iron, is inert gas flushing The theoretical minimum amount of flushing gas required for removing hydrogen and nitrogen can be computed In practice, the amount of inert gas required is somewhat higher because of kinetic factors The kinetics of inert gas flushing can be improved by the dispersion of small bubbles throughout the melt

References

1 R.J Fruehan, in Proceedings of the Physical Chemistry of Foundry Processes Symposium (Warren, MI),

General Motors Corporation, 1986

2 S Katz and C Landefeld, General Motors Research, private communication, 1986

3 Making, Shaping and Treating of Steel, 10th ed., United States Steel Corporation, 1985

4 A Kagawa and T Okamoto, Trans Jpn Inst Met., Vol 22 (No 2), 1981, p 137

5 R.J Fruehan, B Lally, and P.C Glaws, in Proceedings of the Fifth International Iron and Steel Congress

(Washington, DC), Iron and Steel Society of AIME, 1986

6 R.D Pelke and J Elliot, Trans TMS-AIME, Vol 227, 1963, p 894

7 M Inouye and T Choh, Trans JISI, Vol 8, 1968, p 134

8 R.J Fruehan and L.J Martonik, Metall Trans B, Vol 11B, 1980, p 615

9 P.C Glaws and R.J Fruehan, Metall Trans B, Vol 16B, 1985, p 551

10 P.C Glaws and R.J Fruehan, Metall Trans B, Vol 17B, 1986, p 317

11 M Byrne and G.R Belton, Metall Trans B, Vol 14B, 1983, p 441

12 F Tsukihashi and R.J Fruehan, submitted to Trans JISI, 1987

13 R.J Fruehan, Ladle Metallurgy: Principles and Practices, Iron and Steel Society of AIME, 1985

14 G.K Sigworth, Trans AFS, 1987, p 73

15 W.R Opie and W.J Grant, Trans AIME, Vol 188, 1950, p 1234

16 C.E Ramsley and H Neufeld, J Inst Met., Vol 74, 1947-1948, p 559

17 G.K Sigworth and T.A Engh, Metall Trans B, Vol 13B, 1982, p 447

18 J Botor, Metal Odlev, Vol 6, 1980 p 21

19 J Botor, Aluminum, Vol 56, 1980, p 519

20 M.B Bever and C.F Floe, Trans AIME, Vol 156, 1944, p 149

21 A Sieverts and W Krumbharr, Ztch Phys Chem., Vol 74, 1910, p 277

22 P Roentgren and F Moeller, Metallwertschaft, Vol 13, 1934, p 81

23 E Kato, H Ueno, and T Orimo, Trans Jpn Inst Met., Vol 11, 1970, p 351

24 R.H Waddington, J Inst Met., 1948-1949, p 311

25 E.R Thews, Metall., June 1956, p 431

26 W.A Baker and F.C Child, J Inst Met., Vol 70, 1944, p 349

For example, mechanical properties can be adversely influenced by inclusions, which act as stress raisers There is no set pattern for the effect, but some properties are more sensitive to the presence of inclusions than others Elongation or reduction in area is usually modified more significantly than ultimate tensile strength As a result, ductility specifications are common quality control indices in cast products