TRANSFORMATIONS IN SOME AIR COOLED Fe-Mn-Cr-Cu CORROSION RESISTANT WHITE IRONS

Abstract

Of the three varieties of corrosion resistant alloy cast irons in use, the high Si irons have useful applications only in strongly oxidizing conditions. They however, suffer from poor mechanical strength and shock resistance. The high nickel irons, although extensively used in a number of aqueous environments, have a low strength, suffer from graphitic corrosion and are unsuitable at operating temperatures ?800°C. The high chromium irons exhibit relatively higher strength and can be employed upto higher service temperatures. Their shock resistance is improved by lowering carbon content. A critical- analysis revealed that little information is available on the structure-property interrelations in alloy cast irons in general. Furthermore, there is a lack of systematic information on the electra-chemical and on the deformation behaviour of microstructures commonly encountered in alloy white irons namely, 'martensite + carbide'(M + C), 'austenite carbide'(A + C), and their allied counterparts. Detailed information on these aspects is likely to prove useful in ascertaining whether microstructures exhibiting good resistance to aqueous corrosion and useful mechanical properties could be attained through the 'white iron' route. A major advantage foreseen is that the limitations encountered in alloyed gray irons would stand eliminated. It would be equally pertinent to investigate whether these microstructures could be generated by utilizing low cost alloying elements (1.1ri, Cu etc.) in preference to the conventionally employed costlier alloying elements Ni and Mo. The present investigation, therefore, essentially comprised investigating in detail, certain newly designed Fe-Mn-Cr-Cu white iron compositions, in the air cooled condition. Investigations were mainly devoted to assessing their heat-treatment response aimed at establishing interrelations between structure and properties. A study of this kind would require a detailed insight into the transformation characteristics of the alloys. This aspect has received maximum attention in the present study. The alloys which were air induction melted and sand cast (25mm round and 120x20x8mm rectangular strips), were investigated for the transformation behaviour by employing hardness measurements, optical and scanning metallography, quantitative metallography, X-ray diffractometry, electron probe micro analysis and differential thermal analysis. The electro-chemical characterization of the alloys was carried out by employing the potentiostatic method. Compression testing was also carried out to a limited extent to assess the deformation behaviour of the experimental alloys. Computational techniques were extensively employed for data analysis using DEC-2050 and IBM compatible PC-XT and PC-AT systems. Necessary software packages were also developed in FORTRAN IV as and when required. The experimental work involved subjecting disc specimens of the four alloys, containing t6% and 48% Mn, z'4% Cr and t1.5 and t3% Cu, to heat-treatments comprising holding for 2, 4, 6, 8, and 10 hours at 800, 850, 900, 950, 1000 and 1050°C followed by air cooling. This treatment was preferred over oil quenching because B1:4Cr-6Mn-1.5Cu; B2:4Cr-8Mn-l.5Cu; 133:4Cr-6Mn-3Cu; B4:4Cr-8Mn-3Cu ix it can be directly utilized for industrial applications. Optical metallography was extensively used to assess how the alloy content and heat-treating schedule influenced the microstructure which comprised : (i) P/B + M + MC with and without RA in the as-cast state, (ii) M + MC + DC with and without retained austenite(RA) on heat-treating from upto 900°C, (iii) A + MC + DC or A + MC with and without M (in traces) on heat-treating from upto 1000°C, and (iv) A + MC + New phase (eutectic of austenite + carbide called 'anomalous eutectic') on heat-treating from upto 1050°C; volume fraction of the eutectic reaching very low levels at higher soaking periods. The volume fraction of massive carbides(MC) decreased with temperature or with soaking period at a given heat-treating temperature, the effect being marked at temperatures 950°C. Simultaneously, massive carbides were rendered discontinuous from the 'early' stages of heat-treatment. The 'rounding-off' tendency set in at 1000°C. Dispersed carbides(DC) formed at 800°C, 10 hrs. heat-treatment directly from austenite. They underwent coarsening with an increase in temperature and or soaking period. The extent of coarsening which was marked at 900°C and 950°C, has been represented by the 'coarsening indexI(CI). The dispersed carbides dissolved on heat-treating from 1000°C. Hardness measurements provided a quick yet reliable indication of the mechanical properties. A mathematical model was developed based on the effect of heat-treating temperature and time on hardness. This can be represented as: H = Cl ec2' + (C3 + C4.T).t where, H = hardness, VHN30 T = temperature, °K t = time in seconds Cl, C2, C3 and C4 are constants and are different for different alloys. Through intensive calculations it has been possible to demonstrate that the first term of this model represents the matrix related transformations and the second term represents the 'carbide' transformations. The model is thus physically consistent. The predicted hardness values are within t5% of the experimentally determined values. 3D plots amongst the hardness-heat treating temperature & time were also constructed to study the overall transformation behaviour at a glance. The plots revealed that the abovesaid relationship can be represented by a surface with opposite slopes on the two sides of the temperature axis. X-ray diffractometry proved extremely helpful in identifying the different microconstituents observed in the experimental alloys (both in the as-cast and in the heat-treated conditions). It proved helpful in identifying the matrix microstructure in 'marginal' cases e.g. in confirming the presence of P/B & M in the as-cast condition• and of martensite even upto z950°C heat-treated condition. It also established that MDC, M23C6, M5C2 and M7C2 carbides formed in differently identified temperature orm regimes. Additionally, presence of Cu in the elementalAand of xi Fe-Si-carbide(FeaSi2C) was also established. Even after such a detailed analysis, carried out with the help of developed software packages, certain reflections remained unidentified whose indexing was possible&on the likely formation of CrMn2 and Cu2S. This aspect needs further investigation. EPMA studies (point, line and area analysis) carried out on specimens heat-treated at 950°C,(10 hrs.,AC) and 1050°C, (4hrs., AC), besides confirming the deductions arrived at on the basis of X-ray diffractometry and optical metallography, helped in establishing the partitioning behaviour of the different alloying elements -e.g. Mn, Cr and Cu into.the matrix and carbide phases. Chemical compositions of the carbides were also determined. Differential thermal analysis (DTA) of the experimental alloys in the as-cast condition revealed that all the alloys underwent transformations at 2720-735°C (matrix transformation) 890-95 and -=94.5.--9-60-.°C (carbide transformation). Additionally, the alloys 82 and B4 underwent a third transformation at 1050-1075°C representing another carbide transformation. The same study also proved useful in predicting the suitability of the experimental alloys for high temperature applications through an analysis of thermogravimetric (weight gain) data. The as-cast microstructures were found to be suitable upto a service temperature of 600°C only. However, on giving the 950°C, 10 hrs., AC heat-treatment, the usefulness of the alloys was extended upto <800°C. Similarly, on imparting the 1050°C, 10 hrs., AC heat-treatment, the temperature upto which the alloys are useful was further extended at least upto 800°C. This beneficially reflects upon attaining a microstructure, normally xii observed at high temperatures, down to room temperature for improving the high temperature performance of the alloys. A further analysis also explains the relative merits of different microstructures which are a function of the composition and heat-treatment employed for high temperature applications. A mathematical model, developed to interrelate the weight gain with temperature, is of the form TG = Al.e'-''''' where, TG = weight gain T = temperature Al and A2 are constants. Potentiostatic studies, carried out in the Tafel region in 5% NaCI solution, were helpful in characterizing the alloys/ selected microstructures to assess their suitability in resisting corrosion. Two Ni-resist compositions were also studied under similar conditions for the purpose of a comparison. The study showed that in most instances single step polarization curves were obtained signifying corrosion to be a unitary process. In some instances two-step polarization curves, revealing corrosion to occur in two stages, have been obtained. Factors leading to these differences have been identified and the reasons for the occurrence explained. The effect of heat-treatment on the E„,„ & Icon values, via the medium of the microstructure, revealed that the larger the stability and volume fraction of the austenite matrix the more noble (less -ve) the E,„,„ and smaller the I„,„. The effect of the second phase(MC '4- DC) was a function of the morphology and volume fraction of the MC and the size, shape and distribution of the DC. Microstructure, formed on heat-treating at 1050°C (4 hrs., AC}, illustrates the adverse effect of plate like morphology and a large volume fraction of the MC in spite of the austenite matrix being favourably disposed (in improving corrosion resistance). Similarly the microstructure at 950°C (10 hrs., AC) clearly brings out that the corrosion resistance is adversely affected by the coarsened DC. An analysis of the data obtained has proved extremely useful in interrelating the test parameters E„„ & I„,„ with the microstructure and the electro-chemical events constituting corrosion. This has enabled laying down of guide lines for developing corrosion resistant microstructures in terms of E„„, and I,. In the study involving modelling of the corrosion behaviour (interrelating corrosion rate with the microstructure), the models developed in a recent study were critically examined. The first model interrelating corrosion rate with the total volume fraction of MC+DC and the number of particles (NOP) was of the form : CR = 1C1 t C2 (VCLO C3 (.VCb)21 (NOP)e4 where, VCb . volume of carbides (MC+DC) CR = corrosion rate in mdd Noe Cl, C2, C3 and C4 are constants which were different for different alloys. The above model did not predict the corrosion rates very satisfactorily as the contribution of the DC was included in the VCb as well as NOP. Therefore, the volume fraction of DC was excluded from the VCb (only the volume fraction of MC was xiv included) and DC represented by the NOP. The resulting expression is represented as CR = [C1' C2' (VMC) C3' (VMC)2] (NON". This was justified on the basis that the mean diameter (or the surface area) of a DC particle is much smaller as compared to the surface area of massive carbide and since DC were expected to enhance corrosion through the formation of a number of electro-chemical cells, NOP was a more representative measure of this tendency rather than the surface area.