Thermodynamic Fundamentals

Metallurgical thermodynamics belongs to the field of chemical thermodynamics, which is employed to predict whether a chemical reaction is feasible. It also allows quantitative calculation of the state of equilibrium of a system in terms of composition, pressure, and temperature, as well as determination of heat effects of reactions and processes. Laws of thermodynamics are exact. Therefore, calculations based on them are, in principle, sound and reliable. There are standard books dealing with the basics of chemical-cum-metallurgical thermodynamics.1,2 The following is a very brief review only, with special emphasis on topics of relevance to secondary steel making. All reactions and processes tend towards the thermodynamic equilibrium. If sufficient time is allowed, then attainment of equilibrium is possible. Steelmaking reactions and processes are very fast due to their high temperatures. As a result, some of these have been found to approach equilibrium closely within the short processing time. Examples of this in secondary steelmaking are provided in subsequent chapters. Therefore, a full knowledge of thermodynamics is required for the understanding, control, design, and development of metallurgical processes. A discussion of thermodynamics requires precise definitions of some terms. For example, a system is defined as any portion of the universe selected for consideration. The rest of the universe outside the system is known as surroundings. An open system exchanges both matter and energy, a closed system exchanges only energy, and an isolated system exchanges nothing with the ambient. The state of a system is defined at any instant by specifying all state variables and properties such as temperature, pressure, volume, surface tension, viscosity, etc. A complete listing of all the properties of a state is superfluous, because many of them are often mutually interdependent. Pressure (P), volume (V), and temperature (T) are the most common state variables. When a state is described by such variables, assumptions are made implicitly or explicitly. For example, if there is no mention of magnetic field intensity, then it implies that the magnetic effect is insignificant. Similarly, if surface tension forces are ignored, then there is the underlying assumption that surface energy is negligible. Again P, V and T are interrelated. For example, for an ideal gas, PV = nRT (2.1)

where n is the number of moles occupying volume V. In general, for a thermodynamic substance, if V/n = v, where v is molar volume, then v = f ( P, T ) (2.2)

where the R.H.S. of Eq. (2.2) denotes some appropriate function of P and T. Therefore, the state of a thermodynamic substance can be defined by any two of the above three variables, provided that the only work done is against pressure.

©2001 CRC Press LLC

It should be noted that, among these variables, V is a property that depends on the amount of substance under question. On the other hand, P and T are not dependent on mass. A variable such as volume, which depends on the amount of substance in the system, is known as an extensive variable. Variables such as temperature, pressure, etc., which do not depend on mass, are known as intensive variables. It goes without saying that, if an equation contains a variable denoting an extensive property, then there must also be a term denoting mass or mol as in Eq. (2.1). If the latter is missing, there is an implicit assumption that the extensive property is per mass/mol, such as v, in Eq. (2.2), which becomes an intensive property. Thermodynamic relations among intensive properties are of more general validity. A state can be characterized by state variables only when the system has come to equilibrium with respect to those variables. Then and only then can the state be correctly defined in terms of these variables. This also implies that the magnitudes of related intensive properties throughout the system are the same. Thermodynamic equilibrium necessarily requires the attainment of mechanical, thermal, and chemical equilibria. Mechanical interaction of a system with the surroundings is most commonly in the form of pressure. Therefore, in the absence of a field of force, mechanical equilibrium generally means pressure equilibrium, i.e., uniform pressure throughout the system. Similarly, thermal equilibrium implies uniformity of temperature, and chemical equilibrium, in a broad sense, means uniformity of chemical potential for all species in the system. At chemical equilibrium, there is no tendency for further reaction. It is possible that the system is at equilibrium with respect to some variables but not some others. This is known as partial equilibrium, and thermodynamics is capable of handling this as well. However, a precondition for handling any chemical equilibrium is the establishment of mechanical and thermal equilibria.



dE = δq – δw, for an infinitesimal change where dE = an infinitesimal change in the internal energy (E) of the system δq = an infinitesimal quantity of heat absorbed by the system δw = an infinitesimal quantity of work done by the system

For a finite change, ∆E = q – w (2.4)

The first law of thermodynamics is nothing but a statement of the law of conservation of energy. Careful experiments have revealed that q is not equal to w for many processes, apparently violating the law of conservation of energy. To make these findings conform to the law of conservation of energy, the concept of internal energy (E) was proposed. Internal energy is the energy stored in the system. In chemical thermodynamics, E is taken as the energy of atoms and molecules. Experiments have proved that E is a state property. Ignoring other fields of forces, and for a closed system, E = f ( P, T )
©2001 CRC Press LLC


are not thermodynamically reversible. and c are empirical constants. ∆H per mole) values for a variety of processes have been determined experimentally and are available in thermodynamic data books.” ©2001 CRC Press LLC . Using them. in a given state at a certain temperature. has a characteristic value of heat content or enthalpy (H). It may be stated in a variety of ways. The molar specific heat of a substance is the heat required to raise temperature of one mole of a substance by 1 Kelvin. processes taking place without any outside intervention.2. H = E + PV and hence is a state property. All we can measure are changes in E (∆E for a finite change.. dH = ( dE + P dV ) + V dP = δq + Vdp at constant P.11) (2. We do not have to consider the path.8) (2. free expansion of a gas. 2. it does not introduce any significant error if q is taken as ∆H. heat requirements of processes can be calculated.1 presents values of Cp and enthalpies of transformations for iron.Again.2. etc. heat flow. or. Specific heat at a constant pressure is given by ∂q ∂H C p =  ------ =  -------  ∂T  p  ∂T  p Experimental Cp values are expressed as functions of temperature as c C p = a + bT + ----2 T (2.e. Values of Cp may be found in standard thermodynamic data books.. Molar ∆H (i.. such as diffusion.3 STATEMENT OF THE SECOND LAW The second law of thermodynamics is based on universal experience.9) (2. at constant P. By definition. and therefore. Table 2.6) Therefore.7) (2. and a. we do not know or cannot measure the absolute value of E. dP = 0 dH = δq ∆H = q (2. 2. the following statement would be useful: “Spontaneous processes. dE for an infinitesimal change). For the purpose of the ensuing discussions. Most of the processes are carried out approximately at constant pressure. i. Differentiating.e. Even though the pressure fluctuates.10) where T is temperature in Kelvins. This is a great simplification. b. and process heat balances can be worked out.2 ENTHALPY AND SPECIFIC HEAT Every substance. q is related to the change of state property (H) and hence can be calculated from the initial and final states only.

2. DeBoer. state property) known as entropy.13) where A and B designate the initial and final states respectively. 2. where δqrev refers to δq along a reversible path. R.12) or.4 REVERSIBLE PROCESSES Heat and work are not properties of state.M. Niessen. That is why δq and δw rather than dq and dw have been employed in Eq.R. rev ∑ ----------T A→B δq = S B – S A = ∆S (2. (2. ©2001 CRC Press LLC .= dS T (2. These paths would be mostly irreversible. Cohesion and Structure Series.5 × 10 T –3 Enthalpy change (∆Η) (J mol–1) + 5105 + + 670 837 Feδ = 28. and S is a state function (i.K.70 + 19. The term reversible has been coined because such a process can be reversed along the same path without leaving any permanent change in the system or its surrounding. The following relationship has been thereby proposed. would be or can be treated as reversible.4 + 3. Mattens. Amsterdam (l988). and thus the magnitude of q and w would depend on the path that the process takes in going from an initial to the final state. and A. W. Boom.C. the path has to be defined. the concept is very useful and a key one in thermodynamics. however. This is where the concept of reversible processes has assumed importance.2. Some of them.49 + 24.R. in differential form.53 × 10 T –3 +13807 Feliq = 35. Hence. Obviously. δq rev ----------. However.284 +7. 2. A. Miedema.769 × 10–3 T Feβ = 37. Cohesion in Metals—Transition Metal Alloys. This is a great mathematical limitation.66 Feγ = 7.5 ENTROPY (S) A system may go from an initial to the final state by any of the innumerable paths available to it.74 × 10 T –3 Source: F. the system is displaced from equilibrium infinitesimally and then allowed to attain a new equilibrium. They are energy in transition.TABLE 2.1 Specific Heats and Enthalpies of Transformation for Iron Transformation Reaction Feα →Feβ Feβ →Fe Feγ →Feδ Feδ →Feliq Temperature (K) 1033 1187 1665 1809 Specific heat (Cp) (Jmol–1 k–1) Feα = 17. then again displaced infinitesimally and so on. It can be proved on the basis of Carnot’s Cycle that the quantity δqrev/T is dependent only on the initial and final states. In a reversible process.” A reversible process is very slow and impractical. it may be defined as “the hypothetical passage of a system through a series of equilibrium states.3). Thus.e. considerable effort has been made by thermodynamicists to examine under what conditions δw and δq can be related to state properties. North-Holland. No practical process is reversible in strict sense.

This allows evaluation of absolute values of entropy.1 Calculate (a) entropy (So) of 1 mole of liquid iron at 2000 K.5 × 10 dT + ----------T 1665 2000 + ∫ 1665 28.5 J mol K (Ans.66 670 -----------.dT + ----------.+ 7.e. but not in metastable vitreous phases. and (b) enthalpy change ( ∆H ° 2000 – ∆H ° 298 ) in heating 1 mole of iron from 298 to 2000 K.dT + ----------. absolute zero).e.+ 1187 T 2000 ∫ 1187 C p(γ ) ∆H γ → δ ------------.+ 1809 T ∫ 1809 C p(l) -----------.+ 24. which are also tabulated in the thermodynamic data books.2) Substituting the values of Cp and ∆H from Table 2.+ 1665 T ∫ 1665 C p(δ) ∆H m ------------. in perfectly crystalline solids.531 × 10 dT + -------------. Note that ( S° 298 ( α – Fe ) = 27. the entropy of a substance at zero Kelvins (i.dT T (E1.+ T 1809 –1 –1 ∫ 1809 35.284 13807 –3 --------------.dT + ----------------.1) Substituting the values of Cp and ∆H for various transformations from Table 2. 1033 S 2000 ( l ) = S 298 ( α – Fe ) + o o 1665 ∫ 298 C p(α) ∆H a → B -------------.15 + ∫ 298 17.g. i. e. ©2001 CRC Press LLC . Example 2.1.+ T 1033 1809 1187 ∫ 1033 C p(β) ∆H B → γ -------------.dT + ----------------.) (b) Enthalpy change in heating 1 mole of iron from 298 to 2000 K 1033 1187 H o 2000 ( l ) –H o 298 ( s ) = ∫ 298 1665 C p ( α ) dT + ∆H α → b + ∫ 1033 1809 C p ( β ) ( dT + ∆H β → γ ) 2000 + ∫ 1187 C p ( γ ) dT + ∆H γ → δ + ∫ 1665 C p ( δ ) dT + ∆H m + ∫ 1809 C p ( l ) dT (E1.1.769 × 10 dT + ----------T 1033 1665 1187 + ∫ 1033 1809 37.+ 19.49 5105 –3 -----------.4 –3 --------.According to the third law of thermodynamics.. 1033 o S 2000 ( l ) = 27.+ 3.+ T 1187 ∫ 1187 7..7 837 –3 -----. is zero if there is perfect order in that state.15J mol K ) –1 –1 Solution (a) Entropy of liquid iron at 2000 K.745 × 10 dT T = 105. and at complete internal equilibrium.dT + ----------------.

where G = E + PV – TS = H – TS (2.745 × 10 T ] dT –3 = 24971 + 5105 + 5800 + 670 + 16972 + 837 + 5957 + 13807 + 7489 = 821788 J mol (Ans.3) and (2.15) (2. then combining the Eqs.5 × 10 T ] dT + 837 –3 + ∫ 1665 2000 [ 28..14).49 + 24. (2. dH = T dS + V dP (2. if the only work done is against pressure.e.18) .769 × 10 T ] dT + 5105 + –3 ∫ 1033 37.3 CHEMICAL EQUILIBRIUM 2.14) 2. dH = dE + P dV + V dP Combining Eqs.6 COMBINED EXPRESSIONS OF FIRST AND SECOND LAWS For a reversible process and a closed system. internal energy E is expressed as a function of entropy S and volume V.2.1033 1187 H o 2000 ( l ) –H o 298 ( s ) = ∫ 298 1665 [ 17.17) G is known as Gibbs free energy.16) (2. Gibbs. (2.3. Differentiating Eq.17).7 + 19. therefore.14). i. dE = T dS – P dV again. (2.) –1 2.8) and (2.14).1 FREE ENERGY AND CRITERION OF EQUILIBRIUM In Eq.66 dT + 670 + ∫ 1187 1809 [ 7.531 × 10 T ] dT + 13807 –3 + ∫ 1809 [ 35. which is a state property from the definition of G. defined a new function G. both of which are independent state variables.13) we obtain Eq. Experimental control of temperature and pressure is easier. dG = dE + P dV + V dP – T dS – S dT ©2001 CRC Press LLC (2. (2.284 + 7.4 + 3. (2.

T < 0. l. and m denote the number of moles of each. (i. and for a reversible process (or at equilibrium). at constant temperature and pressure. aA + bB = lL + mM (2. The magnitude of a variable for any standard state is indicated by a superscript o. B.19) Therefore.25) .e. (dG)P. the Gibbs free energy provides us with a criterion to predict equilibrium or possibility of occurrence of a spontaneous process at constant T and P. The free energy change for reaction represented by Eq.23) may be expressed as ∆G = ( 1G L + mG M ) – ( aG A + bG B ) (2. b. The standard state could thus be a pure solid or liquid or ideal gas at 1 atmosphere of pressure. which occurs at a temperature T.2 ACTIVITY. it can be shown that dG < V dP – SdT (2.18). L. under constant temperature and pressure.T = 0.20) (2. (∆G)P. natural or irreversible) process would occur if (dG)P. (∆G)P..23) Here A.3.14) and (2. (2.T < 0. for a finite process For an irreversible (spontaneous) process. dG = V dP – S dT at equilibrium. The word isothermal implies that the initial temperature at the beginning of the reaction and the final temperature (when equilibrium is reached) are the same. The standard state is the stablest state of the pure substance at the same temperature (T) and at a pressure of 1 atmosphere. then combining Eqs. G i = free energy of species i at its standard state f a i = ----i = activity of species i at partial molar free energy G i o fi fi = the fugacity of i at the state under consideration o = the fugacity at its standard state fi ©2001 CRC Press LLC o o (2.e. a spontaneous. i. for a finite process (2.For a closed system.22) Thus. It is not necessary that the temperature remain unchanged throughout the progress of the reaction. It can be shown that G i – G i = RT ln a i where. and M are general symbols of chemical species and a. (2. if the only work done is against pressure.e.24) where G i is the partial molar free energy of the species i. 2.T = 0. i.21) (2. EQUILIBRIUM CONSTANT Consider the following isothermal reaction.

the same temperature for reactants and products. Equation (2.e. (2.27) has been derived assuming an isothermal condition.. then for the reaction of Eq.24) through (2.. Activity versus composition relations for all species involved should be known. (2.27) is the basis for prediction of the feasibility of reactions. Since changes in pressure as encountered in metallurgy do not affect thermodynamic properties significantly.23) at temperature T.e.26) where ∆G is the standard free energy change of reaction represented by Eq. Thermodynamic predictions and calculations can be made if the following conditions are satisfied: 1. The process should take place isothermally (i. (2. standard atmosphere = 760 mm Hg).T is negative. the condition that P should be constant is of no importance in situations we normally encounter. ∆G = ( lG L + mG M ) – ( aG A + bG B ) o o o o o o (2. fugacity equals partial pressure. Equation (2. By definition. Combining Eqs. Combining this with Eq. and also that thermodynamic equilibrium prevails. If it is further assumed that the reaction is isobaric. A reaction is spontaneous or feasible if ∆(G)p. ∆G = ∆G + RT ln Q o (2. then ∆(G)P. ©2001 CRC Press LLC .29) Q is called the activity quotient. The standard free energy change of reaction (∆Go) should be available. Equation (2.28). aL ⋅ aM -------------a b a A ⋅ aB 1 m ∆G = ∆G + RT ln o (2. P = constant restriction shall be omitted from here on.28) where aL ⋅ aM Q = --------------a b a A ⋅ aB 1 m (2. the initial and final temperature should be the same) and the temperature should be known. (2. Hence.For ideal gases. the initial and final pressures are the same. 2. expressed in atm (i. 3. (2.23).T is positive. ∆Go = –RT ln[Q]e = –RT ln K (2. i.30) where K is the value of the activity quotient at equilibrium.e.27) or.20).30) is used to calculate the equilibrium condition of a reaction.e. If all reactants and products are at their standard states.T = 0 from Eq.. K is known as the equilibrium constant. activity ai is 1 when species i is at its standard state. i..26). It is impossible when ∆(G)p.

data at steelmaking temperatures in standard compilations are available in the form of Eq.4 ∆G0 FOR OXIDE SYSTEMS In secondary steelmaking.M + O 2 ( g ) = ---M X O Y Y Y (2. T ∆H = ∆H ( at T 1 ) + o f o f ∫ ∆C p dT o T1 (2. we primarily encounter formation or decomposition of inorganic oxides. then the free energy change is related to temperature as ∆G f = ∆H f – T∆S f o o o o o (2. This would make it convenient to compare the data for different oxides. (2. The standard free energies of formation reactions. in the absence of any phase transformation between T and T1.3–6 Some values of ∆G f for compounds of interest in secondary steelmaking are presented in Appendix 2. A number of books carry compilations o of such data. where the formation reaction involves only one mole of oxygen. If the metal. o These are called standard free energies of formation ( ∆G f ) . free energy data shown in diagrams would be for a reaction such as Eq.1. ©2001 CRC Press LLC .31) where M denotes a metal.32) where ∆H f and ∆S f are standard heat and entropy of formation. (2. therefore.34) where ∆H f and ∆S f are standard heat and standard entropy of formation at temperature T. X and Y are general symbols for oxide stoichiometry. Traditionally.dT T o T1 T (2. This allows us to express dependence of ∆G f on temperature as: ∆G f = A + BT o (2.31).35) where A and B are constants. and o ∆C p is the difference of specific heats of products and reactants at standard states. are now known for all inorganic compounds of interest in secondary steelmaking. a brief write-up is presented on free energies of oxide systems.33) ∆S = ∆S ( at T 1 ) + o f o f o o ∆C p ∫ ---------. A more precise representation of ∆G f as a function of T is ∆G f = A + BT + CT ln T o (2. one may assume that ∆H f and ∆S f are o essentially independent of temperature. for the limited temperature range of steelmaking. o Equation (2.35).36) However. oxygen and oxide are in their standard states. The values o o o of ∆C p are generally very small and. representing formation of compounds from the elements. According to Kirchoff’s law. Therefore. Consider formation of an oxide from the elements represented by the following general reaction: 2X 2 -----.2.35) is an approximate one. respectively.

Appendix 2. then ∆Go does not provide a correct guideline. (2. if they are not pure (e.27). pure substances) in accordance with free energy criteria [Eq. These will require knowledge of activity as a function of composition. FIGURE 2.g.1 provides values of A and B for oxides as well as some other compounds of o importance in secondary steelmaking. since the free energy of formation of the former is more negative as compared to that of the latter. and hence the reaction is feasible if all reactants and products are at their respective standard states (i.A1 2 O 3 + Si 3 3 o (2. ∆G f values are per gm mol of O2.. Al2O3 is stabler than SiO2. (2. For example.1 presents a diagram for oxides.37) ∆Go [for the reaction of Eq.22)].37)] = ∆G f (2/3Al2O3) – (SiO2) is a negative quantity. However. This normalization allows us to compare stabilities of oxides directly from such figures. and we have to find out ∆G by using Eq.1 Standard free energy of formation for some oxides. Quantitatively speaking. we are interested in the following reaction: 4 2 -. (2.Al + SiO 2 = -. Figure 2. present as solution in molten iron or slag).. ©2001 CRC Press LLC .e.

40) Activities in slag systems use conventional standard states as reference. Then.2. concentration that is available for chemical reaction. pure Fe is the conventional standard state for liquid steel. ai = Xi (2. by convention.5 are ©2001 CRC Press LLC . activity of a component i in a gas mixture is equal to its concentration. Hence. by definition. For example. CaO-SiO2. since this is its standard state. MnO-SiO2. Fe-Ni. However. The composition of a solution can be altered significantly during processing only if mixing and mass transfer are rapid.5. FeO-SiO2. by definition. i. SiO2. known as the activity coefficient (γ). is 1. For example. Similarly. For example. It may get coated by another solid such as Ca2SiO4 or CaS during steel processing. pure SiO2 is the standard state for a slag containing silica. a pure element or compound constitutes its conventional standard state. during the short processing time. 2 isoactivity lines for CaO and Al2O3. activity versus composition relationships are required here for equilibrium calculations. Al2O3 has been written as AlO1. industrial slags are multicomponent systems. solid CaO remains pure and its activity. an ideal solution obeys Raoult’s law. This is because molecular mass of CaO. molten Fe-Mn. liquid steel contains variable concentrations of impurities and alloying elements. and molecules are free.39) where Wi denotes weight percent and Mi molecular mass of species i. where pi is partial pressure of i in atmosphere.2. For a binary solution (i.+ --------M Cr M Ni M Fe X Cr (2. Departures from Raoult’s law are quantified using a parameter. Hence.+ -------. The liquid field is bounded by liquidus lines. and aFe = 1 for pure iron. show negative departures. etc. Molten slag is also a solution of oxides with a variety of compositions. ai = pi. In the conventional standard state. Hence. Liquid Fe-Cu exhibits positive departure. Hence. its composition does not change.. Here. Molten Fe-Si. let liquid steel contain chromium and nickel. XCr is to be calculated from weight percent composition as follows. and AlO1.e. Also.5 ACTIVITY–COMPOSITION RELATIONSHIPS: CONCENTRATED SOLUTIONS Crudely speaking. a particle of CaO will remain CaO as long as it does not dissolve in slag. the gases behave as ideal. which states. containing two species such as Fe + Ni or CaO + SiO2).. Solid state diffusion is very slow. activity is dimensionless. In this diagram.38) where Xi is mole fraction of solute i in the solution. activity is a measure of “free” concentration in a solution. In metallurgical processing. For example. Figure 2. They either exhibit positive or negative departures from it. W Cr --------M Cr = -----------------------------------------W Cr W Ni W Fe --------. presentation of activity versus composition diagrams is more complex and different from that of a binary solution.3 shows values of activity of SiO2 in CaO-SiO2-Al2O3 ternary system at 1550°C (1823 4 These are in the form of isoactivity lines for SiO . this is illustrated in Figure 2. FeO-MnO solutions are ideal. As already stated. Numerically. there would be diagrams presenting K). However. which is defined as: γ i = a i /x i (2. Similarly. Most real solutions do not obey Raoult’s law.e.

4 ©2001 CRC Press LLC .3 Activity of SiO2 in CaO-SiO2 – Al2O3 ternary system at 1823 K. the liquid at various locations on liquidus is saturated with compounds as shown.FIGURE 2.2 Raoult’s law and real systems showing positive and negative deviations. FIGURE 2.

Solution We are to consider the following reaction: Fe(s) + CO2 (g) = (FeO) + CO(g) For the reaction of Eq. aFe may be taken as 1. ∆G = + 13.3) As discussed earlier. 60. oxidation of Fe is not possible.8 – 161. this is quite unsatisfactory.38 kJ mol–1 = –15. being equal to 56.30) at equilibrium .27). (b) At equilibrium. p CO × ( a FeO ) 0 ∆G = ∆G + R × T ln ---------------------------[ a Fe ] × p CO2 (E2. thermodynamic calculations can be performed properly if the activity vs.1). ∆G = ∆G f [ CO ( g ) ] + ∆G f [ FeO ( s ) ] – ∆G f [ CO 2 ( g ) ] o o o o (E2.45. respectively. (2. solid iron would remain essentially pure in a limited time period. (E2. but the most comprehensive is the Slag Atlas. Example 2. ∆Go at 1300°C (1573 K) = –249. Also. slags are multicomponent and not ternary. and the p CO / p CO2 ratio in gas is 20/1. since its melting point is 1368°C.1) (E2. composition relationship can be expressed by equations. since 1. Slag activity data are available from several sources. and 2.1. Predict whether it is possible to oxidize iron. and 51. This allows easier interpolations and extrapolations of laboratory experimental data in a composition regime. ∆Go = –RT ln K where p CO × ( a Feo ) K = ln ----------------------------[ a Fe ] × p CO2 ©2001 CRC Press LLC (2.3 + 395.2 Solid iron is in contact with a liquid FeO-CaO-SiO2 slag and gas containing CO and CO2 at 1300°C.close.36 kJ mol–1 Since ∆G is positive.7 = –15. So. With the help of Appendix 2. Going through the calculations. the mole fraction scale is approximately the same as the weight fraction scale. calculate equilibrium value of p CO / p CO2 ratio in gas.2) The standard state for FeO is solid pure FeO. The activity of FeO in slag is 0.7 However. Therefore.38 × 103 J mol–1 (a) From Eq.

slags are really ionic liquids.1 A NOTE ON SOLUTION MODELS FOR MOLTEN SLAGS Whitley8 made the earliest effort in this direction. He assumed the slag to consist of 2CaO · SiO2.9 utilized it for reaction of sulfur between liquid steel and slag and obtained the analytical relation for the equilibrium constant as follows: log K h. the other group of models has been termed as ionic models.44) where I has been termed the conversion factor of the activity coefficient from the hypothetical regular solution to the real solution. This is a useful equation.314 J mol–1 K–1. ideal solution) among cations and ideal mixing among anions but no interaction between cations and anions. namely. S = ∑ X′i log K h. However. i K h.S denotes the equilibrium constant for sulfur reaction between metal and slag containing several cations. ©2001 CRC Press LLC . (Note that R = 8. Flood et al. then for a real solution. the general form of the equation for the regular solution model is10 RT lnγ i = ∑ jα ijX j + ∑ j∑ k( αij + αik – α jk ) X j X k 2 (2. It allows i calculation of Kh. who assumed ideal mixing (i. this approach was later extended to the reaction of phosphorus as well.) o 2. known as interaction energies between subscripted solutes. do not exist as such.S i i (2.Using the value of ∆G .20.S in slag from knowledge of K h. For a binary solution. The last assumption is too simplistic and has not been accepted. etc. Slag modeling for thermodynamic calculations is of considerable interest to steelmaking. Some recent studies10 indicate efforts to apply the approach of Flood et al. constituted the basis for some later models. 3CaO · P2O5. For a multicomponent solution. However.41) where Kh. RT lnγ i = ∑ jα ijX j + ∑ j∑ k( αij + αik – α jk ) X j X k + I 2 (2. where some kind of ionic structure is assumed.S of various cations.42) where X2 is the mole fraction of component 2 in the binary 1–2. Hence. X′ i is an electrically equivalent fraction of i among all cations.. and α is a constant.5. Ban-ya11 has recently summarized mathematical expression of slag-metal reactions in steelmaking processes by quadratic formalism based on regular solution model. In contrast to these models. This provides another approach. The first ionic model of salt melts is that of Temkin (1945). SiO2. Analytical relations based on a regular solution model have proved to be the most popular among structure-independent predictions. the p CO ⁄ P CO2 ratio at equilibrium with Fe and the slag turns out to be 7.e. the first assumption. thermodynamic predictions are independent of structural considerations. i denotes a cation.43) where α values are constants. Of course. If the melt is not a strictly regular solution. to estimate “free CaO” in slag as an index of aCaO. and compounds like CaO. etc. ideal mixing among cations and among anions separately. with refinements.S is the equilibrium constant if i is the only cation in slag. the regular solution model predicts RT lnγ 1 = αX 2 2 (2.

αAl-Fe = –41000. X CaO = 0. αCa-Al = –154810. These have yielded some values of αij and I. Solutes in dilute binary solutions obey Henry’s law.). CaO = 50. Composition of slag in weight percent is as follows: MnO = 4. I = –32470 + 26.0384. in dilute solution of species i in liquid iron. (2.3 Calculate γi in a multicomponent solution of slag at 1873 K. J Performing calculations on the basis of Eq. X FeO = 0. Deviation from Henry’s law occurs when the solute concentration increases. 2. where concentration of solutes (carbon. oxygen etc.091.X SiO2 = 0. αMn-Si = –75310. Since we are interested in finding values directly in weight percent. αSi-Fe = –41840 For reaction MnO(s) = MnO (regular solution).Experimental data of various slag-metal and slag-gas equilibria for many slag compositions were statistically fitted with Eq. activities of dissolved elements in liquid steel are expressed with reference to Henry’s law and not Raoult’s law.2339.44). Al2O3 = 35.) are mostly below 1 wt.1 ACTIVITIES WITH ONE WEIGHT PERCENT STANDARD STATE Liquid steel comes primarily in the category of dilute solution.163 (Ans.39): Mole fractions of various species in slag are X MnO = 0. X Al2 O3 = 0. αMn-Al = –83680.6 ACTIVITY–COMPOSITION RELATIONSHIPS: DILUTE SOLUTIONS 2. (2. not mole fraction.6. Therefore. Solution From Eq (2.0284 Interaction energies between various cations are11 αMn-Ca = –92050. FeO = 3 Take MnO as species i.45) where γ i is a constant.11 Example 2.44) and using the above data.14T. αAl-Si = –127610. αCa-Fe = –31380. the composition scale is weight percent.6058. γMnO = 0. SiO2 = 8. ©2001 CRC Press LLC . which is stated as follows: ai = γ i Xi o o (2.% or so except for high alloy steels. With these modifications. αMn-Fe = 7110. αCa-Si = –133890.

then hi = fi Wi (2. For example.5 kJ mol o –1 (2.% standard state of Si dissolved in liquid iron.54) With reference to Eq. std. state for Si in liquid iron o o o (2. ∆G 49 = – 571. SiO2 dissolved in slag is indicated by (SiO2). pct.+ --------M Si M Fe X Si (2. Si ( 1 ) + 0 2 ( g ) = SiO 2 ( s ).49) with pure liquid silicon as standard state. pct. state for Si in liquid iron o m o m where G i = G i – G i is known as the partial molar free energy of mixing of solute i into a solution. However. if Henry’s law is obeyed.52) = ∆G 49 – [ G Si ] at 1 wt.25).53) (2. state Si in liquid iron = RT ln γ Si [ X Si ] at 1 wt. std. at 1 wt. (2. in Fe-Si binary.51) (2.4 kJ mol * o –1 (2. If Henry’s law is obeyed by species i. + 0 2 ( g ) = SiO 2 ( s ). (2.47) (2.48) It is to be noted that the standard free energy change for reaction is not going to be the same if the standard state is changed. from Eq.55) * Note: Si dissolved in liquid metal is denoted either as [Si] or Si. at 1600°C. pct.1. hi = 1. then hi = Wi 2. Again. ©2001 CRC Press LLC . ∆G 50 = –406. This is because. std. If Henry’s law is not obeyed by species i. pct.46) hi is activity and fi is the activity coefficient in the so-called one weight percent standard state.%. G Si = RT ln [ a Si ] at 1 wt.50) ∆G 49 and ∆G 50 are related to each other as o o ∆G 50 = ∆G 49 – [ G Si + – G Si ] at 1 wt. Again. it can be shown that fi is related to Raoultian activity coefficient γi as: γi f i = ---o γi (2. for the 1 wt. o m (2.pct. [ Si ] wt. W Si -------M Si = -----------------------W Si W Fe -------.39).

e. e i is called the interaction coefficient. which is defined as ∂ ( log f i ) j e i = -------------------∂W j Wj→O (2. In the Henry’s law region of Fe-C binary (i.57) 0.. mass% j FIGURE 2.55).02 at WSi = 1.56) where e values are constants.4 illustrates this for activity of carbon and oxygen in liquid iron at 1833 K. i.On the basis of Eq.05 Te As Al Sb Cu Ni Se S (< 4 % ) Sn (< 4 % ) W Cr LOG f N j 0 Ti Mn Mo -0. without any other added element). ©2001 CRC Press LLC . etc. j. m Appendix 2. are solutes. In the presence of a third element in liquid iron solution.10 C Si P Co 0. C.25 × 10–3. and B.2 SOLUTE–SOLUTE INTERACTIONS IN DILUTE MULTICOMPONENT SOLUTIONS It has been found that solutes in a multicomponent solution interact with one another and thus influence activities of other solutes. o o 2..e. then log f i = e i ⋅ W B + e i W C + … + e i W i + e i W j + …… B C i j j (2.4 Influence of alloying elements on the activity coefficient of nitrogen dissolved in molten iron at 1823 K.. fC = 1.. XSi = 0. i. (2.6. (2. Noting that γ Si = 1. Figure 2. It has been derived that if. describing the influence of solute j on fi..05 V 0 1 2 3 4 5 ALLOYING ELEMENT. A is solvent (Fe in case of liquid steel). in a dilute multicomponent solution. fC keeps changing systematically.50) was obtained. log fC = 0.2 presents values of G i for some solute in liquid iron.. the value of ∆G 50 in Eq.

434 × 10 ------------------j Mi Mi i Wi → O (2. and this model has been applied to ternary iron alloys containing these elements to high concentrations. sometimes.2) (E4. The figure is based on several data sources and taken from the review by Iguchi.+ 1. ∂ ( log f j ) i e j = -------------------∂W i j M –2 M i – M = e i ⋅ ------j + 0. The first approach is application of quadratic formalism. C. Equation (2.58) Appendix 2. Figure 2. N. It is. 5 percent N2.4 Liquid steel is being degassed by argon purging in a ladle at 1873 K (1600°C). = -. Example 2. On the other hand.H 2 ( g ) 2 for which 1905 log K H = ----------. nitrogen. and Yj is atom ratio of j in a ternary containing i and j. S. originally proposed by Darken and applied to several binary systems by Turkdogan and Darken. However. the reaction may be written as 1 [ H ] wt. weight percent of the added element exhibit good linear behavior over a long range. all right for dilute solutions of liquid iron. 2 percent manganese.591 T Again. The steel contains 1 percent carbon.5 shows its application to the effect of iron on the activity coefficient of nitrogen in Cr-Fe-N ternary melts. in general. Again. This is demonstrated by Figure 2. The gas bubbles coming out of the bath have 10 percent CO. It predicts linearity between log ψ with Yj.12 In such cases. p H2 K H = ---------[ hH ] ©2001 CRC Press LLC 1⁄2 (E4. etc. (2. at equilibrium. may be treated as interstitial atoms. where ψi is a modified activity coefficient of i.5 percent silicon. Assuming these to be at equilibrium with molten steel.3 presents values of interaction coefficient for some common elements dissolved in liquid iron. calculate the hydrogen. and 0. and oxygen concentrations in steel in parts per million (ppm). The second approach is application of the interstitial solution model originally proposed by Chipman.56) contains only first-order interaction coefficients.10 Ban-ya examined its use in Fe-C-j ternary melts. even here. especially the work of Ban-ya and his coworkers. who had been active in this field for about two decades.e i is known as the self interaction coefficient and has a non-zero value only if Fe-i binary deviates from Henry’s law. Iguchi12 has recently reviewed the subject.56) may be fairly all right up to reasonably high concentrations of solutes. Good linear relation up to a high concentration of iron may be noted. Eq. 5 percent H2.pct. The total gas pressure may be taken as 1 atm.1) (E4. if solute–solute interactions are not significant. Solution (a) For hydrogen.3) .13 Elements P. The following two approaches have been seriously explored. log fi vs. and the rest Ar. second-order interaction coefficients j ( r i ) are to be employed.4 for nitrogen dissolved in liquid iron.

05 atm.002 × 2 + 0.06 × 1 – 0.69 × 10–4 percent = 4.) ©2001 CRC Press LLC . p H 2 = 0. log f H = e H ⋅ W C + e H ⋅ W Mn + e H ⋅ W Si C Mn Si (E4. and O on fH as negligible.4) Assume interactions of dissolved H. log fH = 0.027 × 0. This is justified in view of their j very small concentrations.513 × 10–4 = fH · WH Again.5 Effect of iron on the activity coefficient of nitrogen in liquid chromium. and KH = 405.3. Taking values of e i from Appendix 2.FIGURE 2.58 at 1873 K So.69 ppm (Ans.13 Now. hH at equilibrium = 5.5 putting in values. WH = 4. N.

82. and O on fC as negligible. WN = 0.01. the reaction may be written as 1 [ N ] wt. P CO 0.8) Proceeding as before. So.11) (E4.%.063 T Proceeding as for hydrogen.pct.1 –4 h O = ---------------------.N 2 ( g ) 2 for which 518 log K N = -------. the reaction may be written as [C]wt. hN = fN · WN and log f N = e N ⋅ W C + e N ⋅ W Mn + e N ⋅ W Si C Mn Si (E4.+ 2. P CO K O = -------------------[ hC ] [ hO ] hC = fC · WC = fC · 1 log f C = e C ⋅ W C + e C ⋅ W Mn + e C ⋅ W Si C Mn Si (E4.31 × 10 [ K O ] [ hC ] 419 × 1. = -. N.10) (E4.82. in 1 weight percent standard state Now.) (c) For oxygen.7) (E4.% = 77 ppm (Ans.9) (E4. at equilibrium.+ 1.= 1.003 T Again.= ------------------------. fC = 1.(b) For nitrogen.5) (E4.) Putting in values. + [O]wt. assume interactions of H. and hC = 1. = CO(g) for which 1160 log K O = ----------.6) (E4.%.0077 wt. [hN] = 0.12) (As in previous cases.82 ©2001 CRC Press LLC .

61) (2. we have followed the approach in which the overall free energy change had been employed as the criterion for assessing the feasibility of a process. we say that. for equilibrium The chemical potential approach has the following advantages: 1. and electricity flows from a higher electrical potential to a lower ©2001 CRC Press LLC (2. for the transfer to be feasible thermodynamically.1 × 10–4 wt.59) where µ i (I) and µ i (II) denote the chemical potential of species i in phases I and II.1 ppm (Ans. On the basis of Eq.25). then µ i (I) = µ i (II) O O (2.421). ai(I) > ai(II). and. µ i (I) = µ i (II) (2. µ i is identical with partial molar free energy of solute i in a solution ( G i ) . µ i (I) > µ i (II) and for equilibrium. Just as heat flows from a higher heat potential (temperature) to a lower heat potential.) C 2.% = 4. Then. ai(I) = ai(II).Again.60) (2. WO = 4.13) Putting in values (taking e O = –0.64) . WO = hO/fO and. for transfer from phase (I) to (II). log f O = e O ⋅ W C + e O ⋅ W Mn + e O ⋅ W Si C Mn Si (E4. If the standard state of i is the same in both the phases. µ i (I) = µ i (I) + RT ln ai (I) O (2. (2.63) so. There is an alternative approach based on chemical potential.62) µ i (II) = µ i (II) + RT ln ai (II) O O where µ i denotes the chemical potential of i at its standard state. We can visualize a process better because of the similarity of the concept to some common physical processes. respectively.7 CHEMICAL POTENTIAL AND EQUILIBRIUM So far. Suppose that an element i is to be transferred from phase I to phase II.

then it is µ O2 in both slag and metal. ∆G 69 o (2. it is sufficient to find out the chemical potential of the species concerned only. ©2001 CRC Press LLC . (2. we are primarily concerned about the transfer of oxygen. one should consider the reaction O2(g) = 2[O]. we consider the reaction 2 [ Fe ] + O 2 ( g ) = 2 ( FeO ). then o µ O2 = RT ln p O2 (2. If slag-metal equilibrium exists. because it is not dependent on the composition of iron.67) Hence. Since aFe ≈ 1 in ironmaking and steelmaking. Eq. and µ O2 is set equal to zero.70) (2. then the environment must have a higher chemical potential than oxygen as compared to that in the impure metal.one.66) Calculation of the oxygen potential in a gas phase is relatively simple.71) gives the µ O2 in slag.= – 2RT ln [ h O ] + RT ln p O2 p O2 2 (2. 2. [ hO ] o ∆G h = – RT ln -----------.65) Since a O2 may be equated to p O2 for ideal gas. Suppose we know µ i(I).7. The chemical potential of oxygen (O2) is expressed as µ O2 = µ O2 + RT ln a O2 = µ O2 + RT ln p O2 o o (2. if impurities in a metal are to be preferentially oxidized in refining. For liquid metals. µ O2 ( metal ) = ∆G h + 2RT ln [ h O ] o (2. µ O2 ( slag ) = RT ln p O2 = ∆G 69 + 2RT ln ( a FeO ) O (2. Similarly. If another phase II is brought in contact with it. For the reduction of a metal from its oxide. 2. It is not necessary to bother about the overall reaction. For this.68) For liquid slag in ironmaking.69) 2 ( a FeO ) o ∆G 69 = – RT ln K 69 = – RT ln --------------------2 [ a Fe ] p O2 where p O2 is partial pressure of O2 in equilibrium with [Fe] and (FeO). in the same way a chemical species i is transferred spontaneously from higher µ i to a lower µ i. the reaction environment must have a lower chemical potential than oxygen.1 CHEMICAL POTENTIAL OF OXYGEN In refining processes. all we have to do is to calculate µ i(II) to find out direction of transfer of i.71) If slag-metal equilibrium does not exist.

(E5. and the chemical potential of nitrogen as per Example 2. Eq. because oxygen exists as ions in slag and as dissociated atoms in metal. since the FeO concentration is low. (2. this is unimportant. CO2 – ∆G f . Example 2. µ O2 = – 369.5 Calculate the chemical potential of oxygen in a CO/CO2 gas mixture and slag as given in Example 2.2: Consider the reaction.3) = 2[(–396.1.1). ©2001 CRC Press LLC . There µ O2 is determined more by the following reaction: 4 (FeO) + O2(g) = 2(Fe2O3) (2.4. 2CO ( g ) + O 2 ( g ) = 2CO 2 ( g ) p CO2 o ∆G = – RT ln ---------------------2 p CO × p O2 2 (E5.1) equilibrium p CO = RT ln ( p O2 ) e – 2RT ln ---------2 p CO where ( p O2 ) e is in equilibrium with CO and CO2.CO ] o o o (E5.* ∆G = 2 [ ∆G f .67 kJ mol O 2 –1 (Ans.32 kJ mol–1 Substituting in Eq. As a matter of fact..2) (E5.2). Since p CO 20 1 o --------.35 × 10–3 × 1573)] = –291. Fe3+ ions) also. µ o2 = RT ln ( p O2 ) e = ∆G + 2RT ln  -----  20 1 p CO2 At 1300°C (1573 K).72) However.Of course.4) * From data in Appendix 2.46 + 0. The above analysis does not mean that oxygen is present as O2 in the slag or metal. for the reaction of Eq. (E5. primary steelmaking slags contain Fe2O3 (i.71) is applicable. it is far from being so.. But.) (E5.08 × 10–3 × 1573) + (118.0 + 84. in secondary steelmaking.2.= ----.e. for thermodynamic calculations and concepts. Solution (a) Calculation of µ O2 in CO/CO2 gas mixture for the problem in Example 2.

Steelmakers had always paid attention to it.5) (E5.6) From Appendix 2.45 (given).05 atm in Example 2. (2.) o –1 µ O2 in slag is different from µ O2 in gas. the numerical value of basicity was taken as the CaO/SiO2 ratio. 2∆G FeO ( l ) = – 320. aFeO = 0. or excess base. (c) Calculation of µ N2 for Example 2.46 kJ/mol–1 O2 (Ans. p N2 = RT ln p N 2 = RT ln (0. from Eq. 2.45 ) o (E5. ©2001 CRC Press LLC .) Since liquid steel and nitrogen in exit gas are at equilibrium. because they are not at equilibrium. µ N2 = –46.8 SLAG BASICITY AND CAPACITIES Basicity of a slag increases with increased percentages of basic oxides in it. ∆G FeO ( 1 ) = – 238.2: Consider the reaction 2Fe(s) + O 2 (g) = 2 ( FeO ) Since aFe = 1. It is an important parameter governing refining. µ N2 in liquid steel also shall be the same.4.57 kJ mol Therefore.05) At T = 1873 K. µ O2 = 2∆G FeO ( 1 ) + 2RT ln ( 0. µ O2 (in slag) = –341.71). from Eq.1. In the early days. modified ratio.07 + 49.4: Since p N2 = 0.6). (E5.(b) Calculation of µ O2 in FeO-CaO-SiO2 slag for the problem in Example 2.45 × 1 0 T kJ mol o –3 –1 At 1573 K.65 kJ mol–1 N2 (Ans.

and other methods of estimation based on refractive index.17–19 Some metallurgical slags contain fluorides or chlorides. fluoride or chloride ions also ought to be considered. Λth (for slag/glass) = Σ X i ′ Λth.74).1 OPTICAL BASICITY A breakthrough came with the development of the concept of optical basicity (Λ) in the field of glass chemistry by Duffy and Ingram14 in 1975–76. Other. The most widely employed method of estimation is on the basis of Eqs. (2.Since a basic oxide (e. Estimated values of Λi have been questioned.= 1. where experimental data were not available.74) is not applicable for these from theoretical considerations.75). Λ is an expression for lime character.26 ) Λi (2. the following empirical correlation was proposed by Duffy et al. numerous investigators have applied it to metallurgical slags for a variety of correlations.8. This was tantamount to suggesting that each oxide is characterized by a unique value of Λi.. assignment of correct values of Λth to transition metal oxides such as FeO.35 ( α i – 0.15 From then on. Ca2+. Slags are opaque.i). Pb2+ in CaO. and Pb2+ in a sample. electronic polarizability.74) and (2. In an oxide medium. the concentration of free O2– increases with increasing basicity. This relationship has allowed estimation of Λi for a variety of oxides where experimental data are not available from the values of αi. it has not been accepted by recent investigators. (2. However. Hence. Therefore. Considerable efforts have been made to evaluate Λi for common fluorides such as CaF2. O2–).15 1 ---. respectively. and electron density have been employed besides Pauling’s electronegativity. in addition to oxygen.g. Hence. even though there may not be any CaO in sample. υCaO and υsample are frequencies at peak for free Pb2+. from a thermodynamic viewpoint.. there is no method available for experimental determination of ( a O2 – ) . MnO is controversial. and this in turn produces a shift in frequency in UV spectral band. Differing findings and opinions have been published in the literature.g. ©2001 CRC Press LLC .74) where αi is Pauling electronegativity of the cation in a single oxide i. (2. The estimated Λi is known as theoretical optical basicity (Λth.14. since Eq.16 Optical basicity of individual oxides was estimated from Eq. For a multicomponent system such as slag or glass. Λ is to be estimated for slags. υ free – υ sample Λ = ---------------------------υ free – υ CaO (2. However. Experimental measurements of optical basicity in transparent media such as transparent glass and aqueous solutions were carried out employing Pb2+ as the probe ion.73) where υfree. CaO) tends to dissociate into a cation and oxygen ion (e. 2.i (2. lesser-known methods also have been employed. Therefore. for their contribution toward basicity. It was applied to metallurgical slags first by Duffy. Ingram.75) where X i ′ = equivalent cation fraction of oxide i. The same is true of glasses containing oxides of transition metals. irrespective of medium and temperature. and Somerville. Based on experimental measurements. Moreover. Here. electron donation by oxygen brings about a reduction in the 6s–6p energy gap. Λ = 1 for pure CaO by definition. the activity of oxygen ion ( a O2 – ) may be taken as an appropriate measure of the basicity of slag.

414 Optical basicity of various species (considering data of Nakamura et al. Λ CaO = 1.76) where p O2 ⋅ p S2 are partial pressures of O2 and S2 in the gas at equilibrium with slag. for Example 2.= 0.091 X′ Si4 + = ---------------------. TiO2. especially for transition metal oxides such as FeO.291 2.0384 X′ Mn2 + = ------------------------.091 + 2 × 0.21 were estimated from average electron density.414 3 × 0.0. Solution X′ i = 2 × 0. and others.2 SLAG CAPACITIES Along with basicity. Example 2. % S 2– ) ⋅ p O2 ⁄ p S2 1⁄2 1⁄2 (2.4): Λ MnO = 0. Λ Al2 O3 = 0. (2.82 (Ans.2339 + 4 × 0. It may be noted that there is both agreement as well as disagreement among various investigators.47.4. has evolved.151 2.21.6 Calculate the optical basicity of a slag of composition same as in Example 2. MnO.414 2 × 0. Richardson and Fincham10 in 1954 defined sulfide capacity (Cs) as the potential capacity of a melt to hold sulfur as sulfide.= 0. Noting that the reaction is 1/2 S2 (g) + (O2–) = 1/2 O2 (g) + (S2–) it can be shown that K 77 ⋅ a O2 – C s = ----------------------φ S2 – ©2001 CRC Press LLC (2. electronic polarizability.504 2. and its use has become quite widespread.414. C s = ( wt. Λth.3. The values of optical basicity recommended by Duffy and coworkers18.6058 X′ Cn2 + = ------------------------.95.66.77) (2. Mathematically.3.031 2.78) . It is not presently possible to recommend one set over another. The present status is shown in Appendix 2.0384 + 2 × 0.) 2 × 0.023 2.414 4 × 0. Appendix 2.20 were on the basis of electronegativity.028 X′ Fe2 + = ---------------------.0284 = 2.2339 X′ Al3 + = ------------------------.= 0.= 0. Λ SiO2 = 0. whereas those by Nakamura et al. Λ FeO = 0.slag = 0.6058 + 3 × 0. and refractive index. Equivalent cation fractions of species are: 2 × 0. namely that of slag capacity.94 Substituting the values of X i ′ and Λi in Eq. another concept.75).One important problem facing application of the optical basicity concept has been differing values of Λi proposed by different investigators.= 0.8.414 2.

. carbonate capacity. etc. (2. Capacities have been correlated with Λ. The combination of Eqs. (2.O 2 ( g ) + -. [ a O2 – ] C p = K 79 ⋅ --------------------φ PO3– 4 3⁄2 (2. He has discussed interrelationships among capacities.79) (2. (2. At a constant temperature.75) forms the basis of estimation of optical basicity in slags. Figure 2.83) where m and n are empirical constants. Optical basicity concept is being utilized by industries as well. 3⁄2 in the same slag system (e.80). % PO 4 ) C p = -------------------------------------1⁄2 5⁄4 ( P p2 ) ( p O2 ) 3– (2.22 Hence.78) and (2. Na2O – SiO2 system). Many workers have shown that log Cs = m Λ + n (2.17.g. Figure 2. increasing basicity) and hence ought to have a relation to optical basicity. phosphate capacity. Young et al. temperature. Many papers have been published on measurements and application of these capacities.( O ) = ( PO 4 ) 2 4 2 for which the phosphate capacity of slag may be defined as ( wt. Λ2. Cs increases with increasing a O2– (i.6 is based on review by Sano et al. Wagner22 has critically discussed the concept of basicity and various capacity parameters such as sulfide capacity. ( φ S2– ) ⁄ φPO3– parameter is expected to be 4 constant.24 It shows agreement with the above expectation.77) and φ S2 – is the activity coefficient of S2– in slag in an appropriate scale.e.. He also suggested use of carbonate capacity as a method of measurement of basicity.81) leads to the following relation: φ S2 – 3 log C p = -.81) where K79 is the equilibrium constant for the reaction of Eq.24 The reaction of phosphorus under oxidizing condition may be written as 1 5 3 2– 3– -.log C s + log K 82 + log ---------φ PO3– 2 4 3⁄2 (2.79) and (2.18 have recently questioned applicability of a simple linear dependence of log Cs and log Cp on Λ.723 shows such an attempt for various slags at 1500°C. a corollary to this conclusion is that there is no universal correlation between log Cp and log Cs that will be applicable to all kinds of slag systems.82) K82 is an equilibrium constant term and depends only on temperature.79) and φ PO3– is the activity coefficient 3– 4 of PO 4 in slag in an appropriate scale.where K77 is the equilibrium constant for the reaction of Eq. (2.80) From Eqs. Of course. a single straight line with slope of 3/2 is expected. Similar conclusions can be drawn about interrelationships of other capacities. and relationships among these. as well as some ©2001 CRC Press LLC . Equation (2.P 2 ( g ) + -.

(2.. M. 1980.. 6. in secondary steelmaking. Metallurgical Thermochemistry. a linear relationship as in Eq.. and Grjotheim.. Reading. and Gurry. 4. 7. 1963. Forland. Thermochemistry for Steelmaking. Wicks. N. J.. Cambridge. 1960. Halides. Engrs.S. 1963. 1981. Since.W. E. 9. United States Government Printing Office.01. 4th Ed. 1973. E. ISS..E.. Slag Atlas. Pergamon Press.. Committee for Fundamental Metallurgy.. 10. Vol. T. R.. Physical Chemistry of High Temperature Technology.. Evans. reprinted by permission of the Iron & Steel Society. it has also been shown that. Physical Chemistry of Metals. Mass. USA. and Nitrides. Warrendale. 2. USA. Reading. composition parameters.. J. London. New York.. Elliott. Washington. 1953. Carbides. 1967.. AddisonWesley Publishing Co. C. 59 1922–23. ©2001 CRC Press LLC . 5. Elliott. U. in The Physical Chemistry of Melts. in Proceedings of the Elliott Symposium.L.R.. Mass. and Gleiser.F. Kubaschewski.B. Bureau of Mines. O. 36. Verlag Stahleisen M. Academic Press. Introduction to Metallurgical Thermodynamics. Cs lies in this range. and Alcock. McGraw-Hill Book Co. 1953.E. at values of Cs less than 0. Whiteley. 3. Addison-Wesley Publishing Co.H.T.FIGURE 2. Darken... Gaskell.. Eq. V.B. Thermodynamic Properties of 65 Elements—Their Oxides.723 is also all right. Source: Sano.. D. 2. F. and Block.S. Vol. p. Mass. McGraw-Hill Book Co. Cleveland Inst.F. 8. Flood. M. New York. Thermochemistry for Steelmaking.6 Relationship between sulfide capacities and phosphate capacities for various fluxes. (2. C. K. New York. Proc. and Ramakrishna. REFERENCES 1. H. 2nd Ed.. Gleiser. Turkdogan. Mining and Metallurgy. L. J.. Oxford. Inst. 1.H. However. Dusseldorf. PA...83) as employed in Figure 2.83) would be adequate for industrial uses.

D. 405.. A. ( . 3rd Int. Cambridge... 373. I. p.. in Proc. J. p. p. 21. on Molten Slags and Fluxes.Al203Si02 4. 17. 1410. ISS. J. 13. 1989. in Proc.A. M..Si02 Ca0 . 6B.0 5. 2.Si02 . 1990. B. of Metals. on Molten Slags and Fluxes.. 14.80 0. and Toguri.Mg0 . Ironmaking and Steelmaking. Wagner.Si02 Ca0 . T. Sano.55 0. 1986. 18. Japan Inst.. 1976. 16. D. Sommerville. Ban Ya. 24.C.Mg0 . Inst..0 3. S. Cambridge. Proc. R. N. 1978.65 0.LOG CS Ca0 . 697. of Non-Crystalline Solids. Duffy. of Metals. and McLean.60 0. 1986.. 201.75 0. Washington D. Y. 1993. S. London.. p.23 11. I.B203 Ca0 . J.. p. ISS.. 242. p. 23. 74.. 19.. and Toguri. 19. 1968.. Trans AIME. 3rd Int. USA.Mg0 . Soc. Iguchi.D.M. Metall. 33.M. in Proc. p. optical basicity at 1823 K. and Ingram... p. 50. M.. 22. Conf. USA.0 . Inst. Chem.A..J.Al203 Ca0 . J. 163. 410.J.0 T = 1500 C o 0. Mass. 12. Yokoyama. Duffy.. Z. 1993. and Chipman. 21. p.. 17.70 0. 150. The Elliott Symposium. Ban Ya. Ueda. ISS. Mass. Table 2. in Proc. and Sommerville.1992. T. 1989.. p. and Gustafsson. G. Hassall. and Xu. J.AI203 . 204. Nakamura. 132.D. London. Trans. Conf. J. ©2001 CRC Press LLC .Si02 Ca0 . C. ISIJ Int. J. Ironmaking and Steelmaking. Duffy. Faraday Trans.W.A. Sosinsky. p. Bergman. 6th PTD Conference. ISIJ Int. p. 1990. A. J. p. T. The Elliott Symposium. Trans. Young. 15.7 Logarithm of sulfide capacity vs.D. Duffy. Nakamura.A. A. Y.. 456.) FIGURE 2. Met.. 940.85 0PTICAL BASICITY (Λ).. J.1: Specific heats and enthalpies of transformation for iron.Al203 Ca0 . 1990. 20.. 1. 33..2. 1975. Ingram.

Sign up to vote on this title
UsefulNot useful