Introduction

Iron based alloys are important materials, however, still a lack of experimental material data describing thermo-physical and thermo-dynamical properties of these alloys exist. That is why these alloys are still subject of extensive research. This paper deals with the study of phase transformations temperatures of Fe–C–O based metallic alloys in the eutectoid and alpha–gamma region. Insufficient quantity of thermo-physical and thermo-dynamical data about this system is available at present. For example the ASM international database of phase diagrams [1] contains only two isothermal sections through the ternary phase diagram of the system Fe–C–O, namely for the temperatures of 400 and 600 °C. (The same database contains more than 100 ternary diagrams for the Fe–C–Cr system) Important data are for example temperatures and latent heats of phase transformations, heat capacities [2], surface tension. It is possible to find in available literature values of some of the above physical quantities for the Fe–C [3] and Fe–O [4] systems, however even these available data differ [5]. One of the possibilities of obtaining the necessary data is use of computational relations [6] or simulation (computational) programmes [7], such as IDS software for calculation of phase transformation temperatures and other material properties of steels and Fe–C based alloys [8, 9].

Differential thermal analysis (DTA) [10, 11] was used for obtaining the temperatures of phase transformations for six Fe–C–O based alloys in low-temperature region. Experimental measurements were carried out using the Setaram SETSYS 18TM experimental equipment.

The newly obtained data will serve for completion of the existing thermo-physical data on the system Fe–C–O, and they can subsequently contribute also to a clearer description and easier understanding of more complex poly-component systems, like for example steels [5].

Theoretical background

At present there is insufficient quantity of thermo-physical and thermo-dynamical data on the system Fe–C–O. That is why the binary systems Fe–C and Fe–O, for which substantially more data are contained in literature, were in this study taken as a basis for investigation of the Fe–C–O system.

Fe–C System

Two important phase transformations take place in Fe–C system in the low-temperature region [3]. When the critical temperature, characterised by the GS curve (Fig. 1) is exceeded (cooling), alpha ferrite starts to precipitate from austenite. Alpha ferrite is paramagnetic above the temperature of 760 °C (temperature of Curie point), below this temperature it is getting ferromagnetic properties. At the temperature of 723 °C precipitation of ferrite is terminated and austenite with eutectoid concentration decomposes to a mixture of ferrite and cementite (P), see the straight line OS (Fig. 1). At the temperature, characterised by the OQ curve (Fig. 1), tertiary cementite (Fe3CIII) starts to precipitate from alpha ferrite.

Fig. 1
figure 1

Fe–C metastable equilibrium diagram

Full size image

Fe–O System

Several important phase transformations run in the low-temperature area of the Fe–O system (Fig. 2) [4]. During cooling first occurs precipitation of Fe0.92O from Feγ in conformity with the curve AB. At the temperature of 912 °C three phase transformations run in dependence on oxygen concentration. Feγ(O) is in the concentration range from 0 to 0.0008 at % O transformed to Feα(O), see the straight line CB. In the concentration range from 0.0008 to 0.0017 at % O a transformation of Feγ(O) to Feα(O) takes place, as well as dissolution of Fe0.92O, see the straight line BD. Feγ(O) is in the range from 0.0017 to ~50 at % O transformed to Feα(O) at presence of Fe0.92O, see the straight line DF. Fe0.92O precipitates from Feα(O) according to the straight line DE.

Fig. 2
figure 2

Fe–O equilibrium phase diagram

Full size image

More information can be found for example in [12, 13]. The temperatures of phase transformations and concentration ranges, in which individual transformations run, differ from each other.

Experiment

Samples characterisation

Fe based alloys with graded carbon and oxygen content, prepared in laboratory, were studied. Samples of Fe–C–O alloys were prepared by vacuum melting of ingots made of plasma prepared iron with addition of graphitic carbon. Chemical composition of the Fe plasma is given in Table 1 (Fepl). During the melting the added carbon reacted with oxygen, that was present in the plasma prepared iron, creating oxide of carbon (CO with high probability). The more carbon was added during vacuum melting, the more oxygen reacted with carbon, creating thus oxide of carbon (it was manifested also by an increase of pressure in vacuum furnace). Chemical composition of the analysed samples is given in Table 1. Chemical composition of samples was determined using spectrometer with spark discharge. Carbon and oxygen content in samples was determined using combustion analysers. The samples for DTA analysis were processed into the form of cylinders with diameter of 3.5 mm and height of ~3 mm. Mass of the cylinders was ~150 mg.

Table 1 Chemical composition of analysed samples
Full size table

Experimental conditions

Setaram SETSYS 18TM (TG/DTA/DSC/TMA) laboratory system for thermal analysis and DTA ‘S’ type (Pt/PtRh 10%) measuring rod were used for obtaining of phase transformation temperatures. The samples were analysed in corundum crucibles with volume of 100 μl. An empty corundum crucible served as reference. Dynamic atmosphere of Ar (purity > 99.9999%) was maintained in furnace during analysis to protect the sample against oxidation. Each alloy was analysed at the controlled heating rates of 2, 4, 7, 10 15, 20 °C min−1.

Reference measurements were performed with samples 2, 4 and 6 with use of the DSC method [14] and Netzsch STA 449 F3 Jupiter equipment at the same experimental conditions. Reference measurements were performed at only one heating rate (4 °C min−1). Small thermal effects became less distinct. Comparison of temperatures of phase transformations (Table 2) obtained by DTA and DSC methods showed very good agreement (differences no more than 2 °C).

Table 2 Comparison of characteristic temperatures obtained by DTA and DSC, 4 °C min−1
Full size table

IDS

The IDS software (Solidification Analysis Package) was also used for calculation of phase transformation temperatures. This software enables calculation of phase transformation temperatures of Fe–C based systems depending not only on the chemical composition, but also on the cooling rate and the dendrite arm diameter. In the interval below 1000 °C the software also takes into account the influence of the austenitic grain size and the influence of austenitization after solidification. Calculation of phase transformation temperatures assumes some simplifications, such as e.g.: regular dendritic structure, perfect solubility in liquid phase, thermodynamic equilibrium at interfacial boundaries and some other simplifications and limitations are also applied [8]. Certain suitable concentration interval for every element is recommended by authors. Concentration interval for oxygen is 0.0–0.001 wt % and for carbon is 0.05–1.2 wt %. If the element content is entered outside the given range, the calculated values of temperatures may (need not) be unrealistic. This is also the case of some calculated temperatures of samples analysed in this study. Calculation of temperatures comprised concentrations of all elements present in the analysed samples (Table 1). The maximal possible value of oxygen content enabling to start the calculation for sample 2 was 0.044 wt % (0.044 wt % is the maximum possible concentration for starting the calculation). The cooling rate used for calculation was 0.06 °C min−1. (In order to achieve the best possible approximation to the equilibrium temperatures.) Although eutectoid phase transformation runs at one temperature, for the reasons of non-equilibrium conditions the IDS programme calculated for this transformation a temperature range.

Results and discussion

Temperatures of phase transformations were obtained on the basis of DTA curves (Fig. 3) evaluation. Figure 3 shows DTA curves obtained for the analysed Fe–C–O samples at the heating rate of 2 °C min−1. Evolution of individual DTA curves differs in dependence on carbon and oxygen content in the samples.

Fig. 3
figure 3

DTA curves of the analysed samples, heating rate 2 °C min−1

Full size image

Eutectoid phase transformation (Feα(C) + Fe3C → Feγ(C)) takes place in the samples 4, 5, 6 (Fig. 3). The samples 1, 2, 3 have low carbon content, and for this reason no eutectoid phase transformation took place in them (Fig. 3). Temperature of eutectoid phase transformation is marked as T EUT. Thermal effect of eutectoid phase transformation increases with the increasing carbon content. Thermal effect in the sample 4 is very small. At low heating rates (2 and 4 °C min−1) the thermal effect in the sample 4 was not detected. (Temperature marked as T EUT for sample 4 (Fig. 3) is temperature calculated for the zero heating rate.)

Magnetic transition was detected in all samples. Temperature of Curie point T C was obtained.

Alpha–gamma phase transformation (Feα(C) → Feγ(C)) takes place in a wide temperature range (731–917 °C; temperature range was taken from temperature values calculated for the zero heating rate) for samples 4–6 and in a narrow temperature range for samples 1–3 (893–919 °C). Start of this transformation is marked as T α–γ, S. Temperature range marked as ΔT α–γ, is the range, in which transformation runs, and end of transformation is marked as T α–γ, E. T α–γ, S is identical to the temperature T EUT for samples 4–6. T α–γ, E is identical to the temperature T Feα–Feγ(o) for samples 1–3.

Transformation Feα(O) + Fe0.92O → Feγ(O) + Fe0.92O runs in the samples 1–4. Temperature of this transformation is marked as T Feα–Feγ(o).

Probably, the Fe0.92O in the samples 5 and 6 dissolves and solid solution of Feα(O) forms (T D) and afterwards the transformation Feα(O) → Feγ(O) + Fe0.92O (T Feα–Feγ) takes place. Marking of characteristic temperatures is for clarity summarised in Table 3.

Table 3 Labeling of characteristic temperatures
Full size table

On the basis of Fe–O and Fe–C system knowledge the possible phase transformations were assigned to the observed thermal effects. In the samples 1, 2, 3, 4 two thermal effects occur on DTA curves in the narrow temperature range (of ~890–930 °C). The phase transformations Feα(C) → Feγ(C) and Feα(O) + Fe0.92O → Feγ(O) + Fe0.92O were assigned to thermal effects, on the basis of interpretation of Fe–C and Fe–O diagrams [3, 4] and also on the basis of magnitude of thermal effect of the relevant transformation and carbon or oxygen content in the sample. Assignment of these transformations to individual thermal effects assumes the presence of two phases-solid solution of oxygen in iron and solid solution of carbon in iron. The same assumption was applied for the thermal effects assignment in the samples 5 and 6 also. Succession of individual transformations is schematically presented in Table 4.

Table 4 Scheme of succession of transformations of samples
Full size table

Characteristic transformation temperatures are shown in Tables 5, 6, 7, 8, 9, 10. Small thermal effects (for example thermal effects corresponding to the transformation Feα(C) → Feγ(C) in the samples 1–3) were observed only at lower heating rates (2 and 4 °C min−1). At lower heating rates the overlapping thermal effects are separated. At higher heating rates so small thermal effects weren’t detected, that is why characteristic temperatures of these transformations for higher heating rates are not given in Tables 5, 6, 7, 8, 9, 10.

Table 5 Experimental temperatures of phase transformations and temperatures calculated for the zero heating rate for sample 1
Full size table
Table 6 Experimental temperatures of phase transformations and temperatures calculated for the zero heating rate for sample 2
Full size table
Table 7 Experimental temperatures of phase transformations and temperatures calculated for the zero heating rate for sample 3
Full size table
Table 8 Experimental temperatures of phase transformations and temperatures calculated for the zero heating rate for sample 4
Full size table
Table 9 Experimental temperatures of phase transformations and temperatures calculated for the zero heating rate for sample 5
Full size table
Table 10 Experimental temperatures of phase transformations and temperatures calculated for the zero heating rate for sample 6
Full size table

It is evident from the results given in Tables 5, 6, 7, 8, 9, 10, that heating rate influences the shift of temperatures of phase transformations. Temperatures of phase transformations are shifted with an increasing heating rate towards higher temperatures. The shift varies within an order of units of °C. This can be explained particularly by process dynamics and by detection capability of the device. Different rates of the process of heating/cooling could have also substantial influence on kinetics of the phase transformations. In general, phase transformation mechanism could change with changing heating/cooling rate. Phase transformations may be for this reason at various rates detected at various temperatures (in various temperature intervals) [15].

Extrapolation of temperatures to the zero heating rate

In order to minimize the influence of experimental conditions (heating rate), the temperatures of phase transformations, obtained at heating, were extrapolated to the zero heating rate. The temperatures for the zero heating rate approach the temperatures of phase transformations at equilibrium conditions. More detailed explanation related to the equilibrium temperatures of phase transformations is given for example in [16].

Comparison with accessible data

Temperatures of phase transformations obtained for the zero heating rate (in some cases for 2 °C min−1) were compared with the equilibrium metastable Fe–C diagram (Fig. 4), with the equilibrium Fe–O diagram (Fig. 5), and with the temperatures calculated by the programme IDS (Table 11).

Fig. 4
figure 4

Comparison of zero heating rate temperatures with the metastable equilibrium diagram Fe–C for samples 1, 4–6

Full size image
Fig. 5
figure 5

Comparison of zero heating rate temperatures with the equilibrium diagram Fe–O for a samples 1–3 with higher oxygen content b for samples 4–6 with lower oxygen content

Full size image
Table 11 Comparison of the phase transformations temperatures for the zero heating rate with the temperatures calculated by IDS software
Full size table

Figure 4 presents a comparison of experimentally obtained temperatures of phase transformations with the equilibrium metastable Fe–C diagram. Experimentally obtained temperatures are in this diagram plotted in dependence on the carbon content (wt %). Carbon content in the samples 1–3 is very similar (low), for this reason only the temperatures for sample 1 are given in the diagram for clarity. In the case of temperatures (T EUT, T C, T α–γ, S, T α–γ, E) it was a very good agreement achieved. Differences between experimentally obtained temperatures and temperatures given in the Fe–C diagram are of the order of °C. Higher difference between the temperatures was found only for T α–γ, E of the sample 4 (~15 °C).

The fact that experimentally obtained temperatures of eutectoid phase transformation and Curie temperature are higher in comparison with Fe–C diagram, is probably caused by content of other elements in the samples (Table 1), particularly by content of oxygen.

Figure 5 shows comparison of experimentally obtained temperatures of phase transformations with an equilibrium Fe–O diagram. The experimentally obtained temperatures are in this diagram plotted in dependence on oxygen content in the samples. Oxygen content in the samples varies within the range from 0.0004 to 0.088 wt % O (0.0014–0.3065 at %). Due to different oxygen content in the samples (difference of an order), two cuts of Fe–O phase diagram were selected. Comparison of experimentally obtained temperatures of phase transformations with the phase diagram Fe–O showed a good agreement in the case of temperatures T Feα–Feγ(O), T Feα–Feγ, TD. The differences between experimentally obtained temperatures and temperatures presented in the Fe–O diagram were of the order of centigrades. Higher difference between temperatures was observed only at the transformation Feα(O) + Fe0.92O → Feα(O) in the sample 5, namely ~20 °C.

Experimental temperatures were obtained at the heating mode, contrary to that the temperatures calculated using IDS software for cooling mode. The experimental temperatures of eutectoid transformation lay in the temperature range calculated by the programme IDS (Table 11 ). Only the temperature of eutectoid transformation of the sample 4 is lower (~ by 20 °C) than the lower limit of the temperature range, in which the eutectoid transformation runs according to the IDS programme. Temperatures of the end of alpha-gamma phase transformation obtained in this study are higher than temperatures Ae3 (equilibrium temperature of formation of pro-eutectoid ferrite) calculated by the programme IDS (by more than 10 °C).

It follows from obtained results that the prepared Fe–C–O samples exhibit characteristic features of both Fe–C and Fe–O system. This is evident from the comparison of experimental values of phase transformation temperatures with presented equilibrium phase diagrams (Figs. 4, 5).

Conclusions

Six model alloys with graded carbon and oxygen content were investigated by the DTA and Setaram Setsys 18TM laboratory system. Low-temperature region (below 1,000 °C) of the Fe–C–O system was the investigated area. Phase transformation temperatures were obtained. New original data (phase transformation temperatures) of the Fe–C–O system in the low-temperature region were obtained. Succession of phase transformation was suggested for all the samples.

In order to minimize the influence of experimental conditions (heating rate), the temperatures of phase transformations were extrapolated to the zero heating rate. The temperatures obtained for the zero heating rate were compared with the equilibrium Fe–C diagram, with the equilibrium Fe–O diagram and with the temperatures calculated by IDS software. In majority of transformations temperature differences of the order of units of  °C were observed. Influence of carbon and oxygen content (chemical composition) on temperatures of phase transformations was assessed also.

It follows from literature search that there is a lack of thermo-physical and thermo-dynamical data on Fe–C–O system.

The Fe–C–O system will be still subject of extensive research at our working site.