GENERAL DISCUSSION Mr. D. G. Fraser (University of Oxford) said With reference to the thermo- dynamics of phosphate melts the subject of the paper by Jeffes et al. it might be of interest to outline a theoretical model which Prof. Anderson and I have been working 0n.l The basis of current theories concerning the reaction of basic oxides with oxy-acid melts is the Lux2-Flood expression (I) O”+O” +20’ ; K = (0’)2/(0’’)(00) (1) where 0” = free oxide 0” = 0x0-bridge and 0’ = unshared oxygen. In this the equilibrium between unreacted oxide 0x0-bridges and unshared oxygen is considered and is represented here using the notation of Toop and Sami~.~ One of the assumptioiis of this model which was developed more fully by the latter authors and has been expressed more recently by Masson and his co-workers 5-6 in terms of polymer theory is that the equilibrium constant be independent of bulk composition.This assumption is equivalent to that of the equal reactivity of func-tional groups in co-condensing organic polymers proposed by Flory.’ However in oxy-acid polymers the reactivity of 0x0-bridges may well be dependent on their structural environment and this is especially true of group 5 oxy-acids owing to the differing possibilities for the resonance stabilization of the X-0 double bond in different structural positions.* To account for this we have developed a new model in which reactions are written between oxide ion and the fundamental structural units or structons of which the poly-anions are composed. Using the notation ‘jX where i = no.of unshared oxygens j = no. of oxo- bridges and X = central atom equilibrium in oxy-acid melts may be described by the set of structon equilibria (2) where Slj = structon fraction = nij/Znij+ng. We have included reaction (2.5) in the light of the results of Sridhar and Jeffes.” The overall equilibrium is the resultant of the set of simultaneous structon equilibria and the oxide activity in a binary oxy-acid melt is determined by the set of equilibrium constants and the mass and charge-balance conditions. The chemistry of oxy-acid melts is thus seen to be similar to that of polyprotic acids in aqueous solution com- pleting Lux’s original analogy. The use of the structon model to account for the observed variation in oxide activity with bulk composition for a number of binary oxy-acid melts will be published elsewhere.D. G. Fraser and J. S. Anderson to be published. H. Lux,2.Elektrochem. 1939,45 303. H. Flood and T. Forland Acta Chem. Scand. 1947 1 592. G. W. Toop and C. S. Samis Trans. A.I.M.E. Met. Soc. 1962 224 878. C. R. Masson Proc. Roy.SOC.A 1965 287 201. C. R.Masson I. B. Smith and S. G. Whiteway Can. J. Chem. 1970 48 33 1456. ’P. J. Flory Principles ofpolymer Chemistry (Cornell University Press,Ithaca New York 1953). J. R. Van Wazer Phosphorus and its Compounds vol. 1 (Interscience New York 1958). M. L. Huggins J. Phys. Chem. 1954,58 1141. R. Sridhar and J. H. E. Jeffes Trans. Znst. Min. Met. C 1969 78 14. 64 GENERAL DISCUSSION Dr. J. H.E. Jeffes (Imperial College) (communicated) Fraser’s approach to the problem of polyanionic equilibria in terms of structons appears to be an interesting one and the authors look forward to the elaboration of this treatment. The problem of deviations from ideality which favour formation of oxygen ions and ring-chain equilibria will apparently still create problems in this interpretation of the system. Dr. H. A. Skinner (University of Manchester) said Jeffes refers to “ a calibration by dropping in a suitable platinum sphere from room temperature ”. Would he elaborate on this method and its advantages? Presumably a sphere gains less heat during free fall into the hot zone than would the same mass in the form of a rod or wire; what percentage of (HT-H298)for a Pt sphere was gained during free fall? Dr.J. H. E. Jeffes (Imperial College) (communicated) The heat gained by the platinum spheres during their fall into the calorimeter was determined by the method of Kleppa namely by dropping in spheres of varying mass and extrapolating the resultant effects to an infinitely large mass. This correction was about 2 %. Dr. R. H. Moore (University of Shefield) said In the paper of Mills et al. it is implied that the combined area under the “ 820 K anomaly ” and the a’ -+ a dis-ordering peak was constant for the different heating rates employed. Could the authors state how good the agreement is and compare this agreement with the estimated uncertainty in choosing their base line BHE on fig. 1 ? Dr. K. C. Mills (Nat. Physical Lab.Teddington) said We have measured only the temperature of the CPpeak of the “ 820 K anomaly” as a function of heating rate; all the heat capacity measurements were carried out with a heating rate of 0.333 K s-l. The total enthalpy of the a’ -+a transformation will be unaffected by the heating rate as this represents the total enthalpy required to transform the ordered phase at equilibrium into the disordered phase at equilibrium. Dr. R. H. Moore (University of SheBeld) said I notice that Mills et al. use the data in Kaufmann and Nesor’s paper to calculate the shape of the ordered region in the Fe-Co system. Kaufmann and Nesor’s compilation is a self-consistent set which uses Belton and Fruehan’s liquid data as part of its data-base. The paper by Argent et aL3 suggests that the f.c.c.phase is nearly ideal and that the deviations for the liquid phase are not as strong as suggested by Belton and Fruehaa2 Do Mills et al. think that this discrepancy is sufficiently important to affect significantly the values that they derive for the a’phase? Dr. P. J. Spencer (Nat. Physical Lab. Teddington) said In making calculations of the shape of the ordered region in the Fe-Co system we first attempted to make use of all the available experimental thermodynamic data for Fe-Co alloys. Un-fortunately in contrast to calorimetric studies of the system which in general provide consistent sets of AH data for the liquid f.c.c. and b.c.c. phases there is poor agree- ment between the available experimental Gibbs energy values.The latter include data obtained by Lyubimov et aL4 and Bell in addition to the values given in the L. Kaufmann and H. Nesor 2.Metallkunde 1973 64,249. G. R. Belton and R. J. Fruehan J. Phys. Chem. 1967,71 1403. B. B. Argent P. E. Bloomfield R. H. Moore and D. Robinson this Symposium. A. P. Lyubimov V. Ya. Zobens and V. I. Rakhovski Zhur. Fiz. Khim. 1958 32 1804. H. B. Bell (University of Strathclyde) private communication. SS-3 GENERAL DISCUSSION paper by Argent et al.' Consequently the thermodynamic analysis of the Fe-Co system carried out by Kaufman and Nesor,2 which is consistent both with the pub- lished calorimetric and phase-diagram information was used for the present calcula- tions. In this way the calculated order/disorder boundaries will also be consistent with the data given by Kaufman and Nesor for the different phases of the Fe-Co system.The experimental data of Argent et al. at 1500 K and the analytical expression given by Kaufman and Nesor in fact provide values of AG; for f.c.c. Fe-Co alloys which are in agreement to well within the 15 % accuracy limit quoted by Argent et al. for their experimental activity values. Dr. A. S. Normanton * and P. E. Bloomfield t (University of Shefield) said In principle the Cpvalues obtained using the adiabatic calorimeter at Sheffield Univer- sity agree well with those reported by Mills. The preliminary values reported by Sale were part of the overall set of data obtained on the Fe-Co system and are also consistent with the present data and not 5 % lower as stated in the paper.Heating rates in the order-disorder region were between 7-10 K h-I in an attempt to ensure equilibrium conditions and avoid blurring of the transformation due to fast heating rates. It is thought that further consideration should be given to the interpretation of the order of the order-disorder transformation. Mills et al. state that the phase boundaries that they have determined would still be applicable even if the trans- formation was second order. It is difficult to see how this can be so as second-order transformations are characterized by a single phase for a particular temperature i.e. phase separation does not occur. The use of free energy-temperature curves to define the transformation limits necessitates recognition of the co-existence of ordered and disordered phases.How therefore can the phase boundaries so determined define transformation limits if the transformation is second order ? First-order transformations are accompanied by the evolution of latent heat and with the Fe-Co system using the adiabatic calorimeter we have found no evidence which points conclusively to a latent heat contribution. In our work on the a -+y transformation of pure iron using similar heating rates a latent heat of 930 J mol-' was evolved in 3000 s this time being the length of the plateau on a temperature-time heating curve. Thus if a latent heat of -50 J mole-' had been evolved during the disordering transformation it would have resulted in a plateau of -3 min which would easily have been noticed if present.The behaviour of the Cpmeasurements when passing through the critical temperature was analogous to that occurring when passing through the ferromagnetic to paramagnetic change in dilute Fe-Co alloys. Hence we assume that the latent heat evolved was very small or that the trans- formation occurred entirely by a second-order reaction. Dr. I(. C. Mills (Nat. Physical Lab. Teddington) said In the paper of Mills et al. we have pointed out that the order-disorder transformation in Fe-Co alloys cannot be assigned unequivocally to a " first-order " or " second-order " transition on the basis of extant experimental information. If this transformation should eventually be proved to be " second-order " then the two phase boundaries shown in fig.8 would be more conventionally replaced by a single line lying within these two boun- daries this line representing the critical temperature as a function of composition. B. B. Argent P. E. Bloomfield R. H. Moore and D. Robinson this Symposium. L. Kaufman and H. Nesor 2.Metallkunde 1973 64 249. * now at British Steel Corporation Advanced Process Laboratory Hoyle Street Sheffield. t now at Myers Grove Comprehensive School Wood Lane Sheffield 6. GENERAL DISCUSSION The evidence presented by Normanton and Bloomfield gives some support that the order-disorder transformation in Fe-Co is “ second-order ”. However the evidence is not conclusive and we believe that ‘‘order ” can only be conclusively determined by means of metallographic analyses of the samples annealed for long times at temperatures below the critical temperature so as to attain equilibrium conditions.We are carrying out such experiments but we have no results available at the present time. Dr. Irani has pointed out that heat-capacity measurements are unsuitable for determining the “ order ” of an order-disorder transformation unless they can be carried out at such a slow heating rate that true equilibrium is attained at all times. Irani has also pointed out that CJT) curves obtained for known “first order ” order-disorder transformations do not have the form of classical “first-order ” transformations ; the order-disorder transformation gives rise to a gradual increase in Cpover a large temperature range whereas the “ classical transformation ” gives rise to an infinite increase in Cpat the transformation temperature. R. S. Irani ref. (23) and private communication to National Physical Laboratory Dec. 1973.