Kinetics of Fe2O3-Al reaction prior to mechanochemical synthesis of Fe3Al-Al2O3 nanocomposite powder using thermal analysis
Mehdi Khodaei
1
(
Faculty of Materials Science and Engineering, K.N. Toosi University of Technology, Tehran, Iran
)
Keywords: Kinetics, nanocomposite, Differential Scanning Calorimetry (DSC), Mechanochemical, Fe3Al-Al2O3,
Abstract :
The effect of ball milling on kinetics of the thermite reaction of 3Fe2O3 + 8Al powder mixture to synthesize Fe3Al-Al2O3 nanocomposite was investigated using differential thermal analysis. A model-free method was applied to the non-isothermal differential calorimetry (DSC) data to evaluate the reaction kinetics according to the Starink method. The activation energy of the thermite reaction in the Fe2O3-Al system in ball milled 3Fe2O3 + 8Al powder mixture was determined to be 97 kJ/mole, which is smaller than that for non-milled powder mixture indicating the change of reaction mechanism. The change in the reaction mechanism could be resulted from the formation of short-circuit diffusion paths occurring in the precursors during milling. The change in the reaction mechanism of such nanostructured 3Fe2O3 + 8Al powder mixture could be reason of the formation of desired phases (Fe3Al and Al2O3), which such stoichiometric phases cannot be achieved by conventional molten state thermite reaction.
Kinetics of Fe2O3-Al reaction prior to mechanochemical synthesis of Fe3Al-Al2O3 nanocomposite powder using thermal analysis
Abstract
The effect of ball milling on kinetics of the thermite reaction of 3Fe2O3 + 8Al powder mixture to synthesize Fe3Al-Al2O3 nanocomposite was investigated using differential thermal analysis. A model-free method was applied to the non-isothermal differential calorimetry (DSC) data to evaluate the reaction kinetics according to the Starink method. The activation energy of the thermite reaction in the Fe2O3-Al system in ball milled 3Fe2O3 + 8Al powder mixture was determined to be 97 kJ/mole, which is smaller than that for non-milled powder mixture indicating the change of reaction mechanism. The change in the reaction mechanism could be resulted from the formation of short-circuit diffusion paths occurring in the precursors during milling. The change in the reaction mechanism of such nanostructured 3Fe2O3 + 8Al powder mixture could be reason of the formation of desired phases (Fe3Al and Al2O3), which such stoichiometric phases cannot be achieved by conventional molten state thermite reaction.
Keyword: Kinetics; differential scanning calorimetry (DSC); mechanochemical; nanocomposite; Fe3Al-Al2O3.
1. Introduction
Various techniques have been introduced for fabrication of metal/ceramic composites, which are based on ex-situ processes (addition of reinforcement to the matrix) such as convectional powder metallurgy and casting techniques or in-situ processes (direct fabrication of reinforcement in the matrix) such as direct metal oxidation, reactive melt penetration, reactive sintering, and self-propagating high-temperature synthesis (SHS) [1, 2].
The in-situ processes can economically fabricate a significant wider range of interpenetrating phase composites. In order to synthesis the metal(intermetallic)/alumina composite, the in-situ reaction between metal oxide and elemental aluminum could be utilized using various methods such as aluminothermic reaction [3] and mechanochemical reaction [4]. Synthesis of Fe(Al)/Al2O3 composite by alominothermic reaction [5] and mechanochemical reaction [6,7] have been investigated. The kinetics of thermite reaction of 3Fe2O3 + 8Al powder mixture in molten state to synthesize Fe3Al/Al2O3 composite was investigated by Fan et al. [5]. The theoretical as well as experimental investigations revealed that FeAl2O4 phase was obtained instead of thermodynamically predicted Fe3Al phase during reaction of molten Al with Fe2O3 [5].
In our previous works [7, 8], Fe3Al/Al2O3 nanocomposite was successfully synthesized by mechanochemical reaction of 3Fe2O3 + 8Al powder mixture. In this paper, the kinetics aspect of mechanically activated 3Fe2O3 + 8Al powder mixture was investigated by differential scanning calorimetry (DSC) in order to reveal the capability of mechanochemical process rather than thermite reaction to synthesize Fe3Al/Al2O3 composite.
2. Experimental
2.1. Experimental methods
Fe2O3 (99.9% purity) and Al (99.5% purity) were mixed according to the following reaction.
3Fe2O3 + 8Al = 2Fe3Al + 3Al2O3 (1)
Ball milling of powder mixture was carried out in a Spex8000 type ball mill. The vial rotation speed and its vibration frequency were 700 rpm and 11.7 Hz, respectively. The milling media were hardened chromium steel consisted of five 12 mm diameter balls, confined in a 75 ml volume vial. A total powder of 7 g without any process control agent was milled under argon atmosphere. The ball-to-powder weight ratio was 5:1. The phase composition of samples was investigated by X-ray diffractometery (XRD) using a Philips X'PERT MPD diffractometer with Cu Kα radiation (λ=0.15418 nm). Differential scanning calorimetry (DSC) was carried out in a Q100 V9/4 Build 287 thermal analyzer under flowing argon atmosphere in alumina crucibles within a temperature range of 30-1200 °C, applying different heating rate from 5 to 20 °C/min.
2.2. Theoretical methods
Kinetics of thermally activated solid state reaction has been widely investigated by following formula:
dα/dt = k(T) . f(α) = A exp (-Ea / RT) . f(α) (2)
Where α is the extent of reaction (conversion degree); t, the time; R, universal gas constant; T, temperature; f(α), kinetic model function; A and Ea are Arrhenius parameters, apparent pre-exponential factor and apparent activation energy, respectively.
Two approaches that used in kinetic evolutions are model-fitting and model-free methods. In model-fitting methods, a kinetic model can be assumed to describe the reaction behavior for both isothermal and non-isothermal kinetic evolutions. By fitting the experimental data to different assumed model function f(α), the overall reaction is described by the combination of formal reaction steps with constant Arrhenius parameters, which are determined by choosing the form of f(α). However, in a non-isothermal experiments both T and α vary simultaneously, the model-fitting methods generally fail to achieve accurate separation between k(T) and f(α) functions. As a result, any f(α) can satisfactorily fit data at a cost of drastic variations in Arrhenius parameters, which compensate for the different between the assumed form of f(α) and the true but unknown reaction model. Consequently, the model-fitting method leads to achieving the uncertain values of Arrhenius parameters [9].
In model-free method, the Arrhenius parameters could be determined without choosing the reaction model. The well known approaches, if several measurements with different heating rates and/or with different temperatures are performed, are the isoconversional method according to the Friedman and the integral isoconversional method according to the Kissinger and Ozawa [9]. For non-isothermal analysis at constant heating rate, mean activation energy can be derived using Kissinger-type isoconversional methods. However, these methods are subject to approximations which can introduce inaccuracies in the determination of Ea. A new method for the derivation of activation energy is proposed by Starink [10]. It is shown that this method is more accurate than the Kissinger and Ozawa methods.
A comparison of the Kissinger-type methods shows that they all comply with the following equation:
ln (Tps / φ ) = A (Ea / RTp) + constant (3)
where Tp is the peak temperature of DSC curve; φ, is the heating rate; s, is a constant, and A is a constant which depends on a choice of s. In the case of Kissinger method, s = 2 and A = 1, and for Ozawa method s = 0 and A = 1.0518, while for the Starink method s = 1.8 and A = 1.007-1.2 x 10-5 Ea (Ea in kJ/mol). The activation energy, Ea, can be determined from a plot of ln (Tps / φ ) versus of 1/Tp.
3. Results and discussion
The 3Fe2O3 + 8Al reaction is thermodynamically favorable giving Fe3Al and Al2O3 phases. Fan and et al. [5] reported that during reaction of Fe2O3 and molten aluminum, instead of thermodynamically predicted Fe3Al intermetallic compound, FeAl2O4 is predominately formed and thermite reaction in 3Fe2O3 + 8Al powder mixture resulted in formation of Al2O3 and FeAl2O4. It has been revealed [5] that the activation energy for the thermite reaction in 3Fe2O3 + 8Al powder mixture is near to that for the diffusion of Al into FeAl2O4 phase. Hence, the formation of FeAl2O4 phase at Fe2O3/Al interfaces controls the rate of reaction. Khodaei and et al [7, 8] reported that the Fe3Al and Al2O3 phases were successfully synthesized by mechanochemical reaction of 3Fe2O3 + 8Al powder mixture. The calculated adiabatic temperature (Tad) [7] as well as experimental observations [8] revealed that the mechanochemical reaction of 3Fe2O3 + 8Al powder mixture occurs with a combustion mode in which all expected phase, Fe3Al and Al2O3, are achieved.
The XRD patterns of 3Fe2O3 + 8Al powder mixture as-received and after different ball milling times are shown in Fig. 1. XRD pattern of powder mixture after 2 h of milling time (prior to combustion reaction, Fig. 1(b)) was identified as a mixture of Fe2O3 and Al indicting that ball milling up to 2 h had no effect on as-received powder mixture except broadening of Bragg peaks resulting from nanocrystallization as well as the enhancement of lattice strain. XRD pattern taken immediately after combustion (by recording the vial temperature during ball milling [8]), Fig. 1(c), showed no Fe2O3 and Al peaks. The diffraction peaks of reaction products were identified as α-Al2O3 and disordered Fe3Al intermetallic compound. Although, the trace of Fe peak can be observed on XRD pattern, the peaks related to the FeAl2O4 phase were not detected. These results are in contrast with the results of the thermite reaction of 3Fe2O3 + 8Al [5] where the molten Al reacts with Fe2O3 and the undesired FeAl2O4-Al2O3 phases were obtained.
In order to evaluate the effect of mechanical activation on kinetics of Fe2O3-Al reaction, the 3Fe2O3 + 8Al powder mixture after 2 h of ball milling (before starting the combustion reaction during milling) was heated in DSC. The activation energy of reaction was obtained according to the Starink method [10].
Fig.1 XRD patterns of 3Fe2O3 + 8Al powder mixture as-received and after different milling time.
Fig. 2 shows the DSC curves of 2 h ball milled 3Fe2O3 + 8Al powder mixture at different heating rates. As can be seen, all curves have similar peaks including two exothermic peaks and one endothermic peak. The endothermic peak at about 660 ˚C is related to the melting of Al. In order to determine the first exothermic peak, the 2 h ball milled 3Fe2O3 + 8Al powder mixture was heated up to 600 ˚C with a heating rate of 10 ˚C/min. The XRD pattern of the 2 h ball milled 3Fe2O3 + 8Al powder mixture before and after heating up to 600 ˚C are shown in Fig. 3. Fig. 3(b) consisted of the Al2O3, Fe3Al, Fe2O3, and Al peaks. The first exothermic peak is therefore related to the solid state reaction of Fe2O3 and Al at about 600 ˚C. On DSC curve of 3Fe2O3 + 8Al powder mixture without ball milling, this exothermic peak occurred after melting of Al at about 900 ˚C [5]. The second exothermic peaks at around 800 ˚C on DSC curves may be due to the reaction of remaining Al and Fe2O3.
Fig. 2 DSC curves of 2 h ball milled 3Fe2O3 + 8Al powder mixture at different heating rates.
As mentioned in section 2-2, the activation energy (Ea) could be determined from the slope of the Starink plot. Fig. 4 shows the Starink plot of ln(Tp1.8/φ) versus 1/Tp for the first exothermic peak. The activation energy of 2 h ball milled 3Fe2O3 + 8Al powder mixture was measured to be about 97 kJ/mole.
Fig. 3 XRD pattern of 2 h ball milled 3Fe2O3 + 8Al powder mixture (a) as-milled, and (b) after subsequent heating up to 600 ˚C.
The first exothermic peak on DSC curves of 3Fe2O3 + 8Al powder mixture decreased from about 900 ˚C for unmilled powder [5] to about 600 ˚C for ball milled powder mixture. Also Ea of unmilled 3Fe2O3 + 8Al powder mixture, calculated by Starink method, was reported to be about 145 kJ/mole [5], whereas in ball milled powder mixture, the Ea reduced to the 97 kJ/mole. These results indicate that the reaction mechanism of 3Fe2O3 + 8Al powder mixture is changed as a result of ball milling. Similar results are found in case of CuO-Fe reaction, which has been reported [11] that Ea of unmilled 5CuO + 4Fe powder mixture was about 575 kJ/mole which reduced gradually during ball milling to about 200 kJ/mole prior to combustion. Investigation of microstructural changes during ball milling of 3Fe2O3 + 8Al powder mixture revealed [8] that the crystallite size and mean lattice strain of Fe2O3 and Al after 2 h of milling time became to 77nm, 0.8% and 81nm, 0.41%, respectively. Reducing the crystallite size to nanometer range and increasing the defect densities increase the reaction rate by providing short-circuit diffusion paths [12]. This nanostructured composite of Fe2O3 and Al layer before ignition may be caused to the formation of predicted Fe3Al phase. The results of molecular dynamics (MD) simulations of the planar interfacial contact of Al/Fe2O3 nanolaminate thermite systems reported by Lin et. al. [13] revealed this hypothesis. They defined that the reaction will start when the initial distance among reactants is shorter than a certain distance after passing through the ignition delay. Moreover, Sui et. al. [14] revealed experimentally that the ignition of Al and Fe2O3 layers is closely associated with oxygen which is produced from thermal decomposition of Fe2O3 and the thickness of the reaction zone is restricted by the diffusion length of oxygen across the layer-to-layer interface of Al and Fe2O3, suggesting the benefit of nano-scale structure in such combustion system.
Fig. 4 Starink plot for the first exothermic peak of DSC curves
It was proposed by Forrester and Schaffer [11] that in unmilled powder mixture, the bulk ionic diffusion in oxide phase is rate-controlling of the reaction, subsequently, diffusion along grain boundaries and ionic short-circuit diffusion paths control the reaction in ball milled powder mixture, which is supported by dynamic molecular simulation [13] and experimental analysis [14]. It seems that the introducing the short-circuit diffusion paths by increasing the structural defect densities plays an important role in changing the reaction mechanism. In nanostructured powder mixture in which reactants are uniformly mixed at atomic scale, the diffusion mode has been changed. For ball milled powder mixture, the Fe2O3-Al reaction takes place before melting of Al in the solid state whereas, for unmilled 3Fe2O3 + 8Al powder mixture this reaction occurs after melting of Al [5]. These observations along with suddenly combustional synthesis of Fe3Al/Al2O3 nanocomposite during mechanochemical reaction of 3Fe2O3 + 8Al powder mixture [7, 8] suggest that during ball milling process, the attainment of the critical combustion condition depends on the developments of an excess short-circuit diffusion paths and reducing the activation energy to occurring the sudden reaction. Hence, for ball milled 3Fe2O3 + 8Al powder mixture the metal ionic diffusion (diffusion of Al in FeAl2O4) that proposed by Fan and et al. [5] does not occur and the change of ionic diffusion mode result in formation of desired phases.
4. Conclusion
The effect of ball milling on kinetics of Fe2O3-Al reaction for mechanically activated 3Fe2O3 + 8Al powder mixture was investigated in order to reveal the influence of mechanical activation on reaction mechanism. The activation energy of 2 h ball milled 3Fe2O3 + 8Al powder mixture (prior to combustion during mechanochemical synthesizing the Fe3Al/Al2O3 nanocomposite) was determined by non-isothermal differential scanning calorymetry performed at different heating rate using Starink method. The activation energy of Fe2O3-Al reaction for ball milled 3Fe2O3 + 8Al powder mixture is measured to be 97 kJ/mole, which is smaller than that for unmilled powder mixture reported elsewhere, indicating the change in reaction mechanism. It seems that ball milling cause to nanocrystallization of precursor powders providing the short-circuit diffusion paths to enhance their reaction ability resulted in formation of desired phases, Fe3Al and Al2O3. Such desired and stoichiometric products could be resulted from the nanostructured 3Fe2O3 + 8Al powder mixture, whereas the conventional molten state thermite reaction has not such capability.
References
[1] C. Suryanarayana, Mechanical alloying and milling, Prog. Mater. Sci. 46 (2001) 1–184.
[2] R. Casati, M. Vedani, Metal Matrix Composites Reinforced by Nano-Particles—A Review, Metals, 4 (2014) 65-83.
[3] H.X. Zhu, R. Abbaschian, In-situ processing of NiAl–alumina composites by thermite reaction Mater. Sci. & Eng. A 282 (2000) 1-7.
[4] B.S.B. Reddy, K. Rajasekhar, M. Venu, J.J.S. Dilip, S. Das, K. Das, Mechanical activation-assisted solid-state combustion synthesis of in situ aluminum matrix hybrid (Al3Ni/Al2O3) nanocomposites, J. Alloys & Comp. 465 (2008) 97-105.
[5] R.H. Fan, H.L. Lu, K.N. Sun, W.X. Wang, X.B. Yi, Kinetics of Thermite Reaction in Fe-Al2O3 System, Thermochim. Acta 440 (2006) 129-131.
[6] D. Oleszak, M. Krasnowski, Mechanically alloyed nanocrystalline intermetallic matrix composite reinforced with alumina, Mater. Sci. Forum 360-362 (2001) 235-240.
[7] M. Khodaei , M.H. Enayti, F. Karimzadeh, Mechanochemical behavior of Fe2O3–Al–Fe powder mixtures to produce Fe3Al–Al2O3 nanocomposite powder, J Mater. Sci. 43 (2008) 132-138.
[8] M. Khodaei , M.H. Enayti, F. Karimzadeh, Mechanochemically synthesized Fe3Al-Al2O3 nanocomposite, J. Alloys & Comp. 467 (2009) 159-162.
[9] S. Vyazovkin, C.A. Wight, Model-free and model-fitting approaches to kinetic analysis of isothermal and nonisothermal data, Thermochim. Acta 340–341 (1999) 53-68.
[10] M.J. Starink, A new method for the derivation of activation energies from experiments performed at constant heating rate, Thermochim. Acta 288 (1996) 97-104.
[11] J.S. Forrester, G.B. Schaffer, The chemical kinetics of mechanical alloying, Metall. Trans. A 26 (1995) 725-730.
[12] G.B. Schaffer, P.G. McCormick, Displacement Reactions during Mechanical Alloying, Metall. Trans. A 21 (1990) 2789-2794.
[13] L.Z. Lin, X.L. Cheng, B. Ma, Reaction characteristics and iron aluminides products analysis of planar interfacial Al/α-Fe2O3 nanolaminate, Computational Materials Science 127 (2017) 29-41.
[14] H. Sui, S. Atashin. J.Z. Wen, Thermo-chemical and energetic properties of layered nano-thermite composites, Thermochimica Acta 642 (2016) 17-24.