- Email: [email protected]
Microstructure and reactivity of Fe2O3-Li2CO3-ZnO ferrite system ball-milled in a planetary mill
Thermochimica Acta 664 (2018) 100–107
Contents lists available at ScienceDirect
Thermochimica Acta journal homepage: www.elsevier.com/locate/tca
Microstructure and reactivity of Fe2O3-Li2CO3-ZnO ferrite system ball-milled in a planetary mill
T
⁎
Elena Lysenko , Evgeniy Nikolaev, Vitaliy Vlasov, Anatoliy Surzhikov Tomsk Polytechnic University, Lenina Avenue 30, 634050 Tomsk, Russia
A R T I C LE I N FO
A B S T R A C T
Keywords: Substituted lithium ferrite LiZn ferrite Mechanical activation Microstructure Reactivity Thermal analysis
In this work, the microstructure of mechanically activated Fe2O3-Li2CO3-ZnO mixture for the lithium-zinc ferrites production was studied using the Brüner, Emmett, Teller and laser diffraction methods as well as X-ray diffraction and scanning electron microscopy analyses. The reactivity of reagent mixture was investigated by thermogravimetric and calorimetric analyses. The ball milling was performed in a AGO-2S high energy planetary ball mill with a vial rotation speed of 2220 rpm using steel grinding balls. The milling times were 0, 5, 15, 30 or 60 min. It was shown that the composition of mixture changes during the ball milling, which consists in decreasing the α-Fe2O3 concentration and increasing the Fe3O4 spinel phase, while the Li2CO3 and ZnO concentrations remain unchanged. It was found that the milling leads to decrease in the average particle size of the reagents and simultaneously formation of large size agglomerates with denser structure and well-interlinked particles. It was established that observed changes in microstructure and phase composition lead to an increase in the reactivity of the Fe2O3-Li2CO3-ZnO system and the acceleration of the chemical reaction between reagents.
1. Introduction The Li2O-Fe2O3 system is of considerable interest for production of a variety of materials such as ferrites with different objective functions, catalysts, solid-oxide electrodes [1–3]. According to the equilibrium phase diagram, the Li2O-Fe2O3 system may form the LiFeO2 lithium orthoferrite (γ – low-temperature high-ordered form and α – hightemperature disordered form; the temperature of γ↔α transition is ca. 670 °C) and the Li0.5Fe2.5O4 lithium ferrite with spinel structure [4,5]. The latter has the structure of an inverse spinel Fe3+[Li0.5+Fe1.53+]O4, and at T < 750 °C, lithium and iron ions are ordered in the Fd-3m (αphase) type in the octahedral sublattice. However, disordered form of lithium ferrite has a random statistical distribution of lithium and iron over all the octahedral positions. The properties of lithium ferrites can be tailor made by substituting them with different metal ions such as Zn2+, Cu2+, Ti4+, Mg2+, Co2+, Ni2+, etc. Zinc is known to play a decisive role in determining the ferrite properties [6–8]. The redistribution of metal ions over the tetrahedral and octahedral sites in the spinel lattice on incorporation of zinc is responsible for the modification of ferrite properties. Thus, substituted lithium ferrites with chemical composition of Li0.5(13+ are substituted by ions of zinc, x)ZnxFe2.5-0.5xO4, in which the ions Fe are characterized by high saturation magnetization and this promotes their wide application in microwave technology [9,10].
⁎
The conventional way to synthesize lithium ferrites involves the use of high-temperature solid-state reaction including the mixing of oxide/ carbonate and then mixtures annealing at high temperatures [11,12]. Such synthesis method has some disadvantages which include chemical inhomogeneity, coarser particle size and volatility of Li2O from ferrites during synthesis [13]. In order to overcome these drawbacks, a number of chemical methods were developed to prepare ferrites at low temperatures. They include the hydrothermal method, the citrate precursor and sol gel methods and etc. [14–19]. However, many of these methods have major drawback which consists in applying the repeated high temperature to form the monophase of lithium ferrite. One of the methods that allow to reduce the synthesis temperature is preliminary mechanical activation of ferrite reagents in high energy ball mills [20–23]. Such powders are highly reactive, which makes it possible to obtain reaction products at lower temperatures and shorter thermal treatment period. According to earlier investigation, mechanical activation can be used to produce lithium [24–26] and substituted lithium [27,28] ferrites at lower temperature than in conventional ceramic processing. In our previous work [29], the influence of mechanical activation of Li2CO3-Fe2O3-ZnO and Li2CO3-TiO2-Fe2O3 initial reagent mixtures on the solid-phase synthesis of Li0.4Fe2.4Zn0.2O4 and Li0.6Fe2.2Ti0.2O4 substituted lithium ferrites was studied. It was established that the ferrites can be obtained from mechanically activated
Corresponding author. E-mail addresses: [email protected] (E. Lysenko), [email protected] (E. Nikolaev), [email protected] (V. Vlasov), [email protected] (A. Surzhikov).
https://doi.org/10.1016/j.tca.2018.04.015 Received 9 February 2018; Received in revised form 27 April 2018; Accepted 28 April 2018 Available online 01 May 2018 0040-6031/ © 2018 Elsevier B.V. All rights reserved.
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
mixtures at temperatures at least 200 °C lower than in the case of using un-milled mixtures. Although a substantial literature has accumulated on the use of mechanical milling for ferrite formation, there is only limited information on the microstructure of ferrite powders. In this work, using the Brüner, Emmett, Teller (BET) and laser diffraction methods, as well as XRD and SEM analysis, the microstructure of un-milled and ballmilled initial reagents for the lithium-zinc ferrites production was studied. The reactivity of reagent mixtures was investigated by thermogravimetric and calorimetric analyses.
2. Experimental The initial reagents were commercially Fe2O3, Li2CO3 and ZnO (99%, Reahim Co., Russia) powders which were pre-dried and mixed in molar ratios of 86.05 mass%, 6.64 mass% and 7.31 mass%, respectively, according to the reaction Li2СO3 + 6Fe2O3 + ZnO→5 Li0.4Fe2.4Zn0.2O4 + СО2↑
(1)
The mixture was divided into few batches; the samples of first batch were simply mixed in an agate mortar and are considered to be simple un-milled mixture. The rest of the powder was subjected to mechanical activation using a AGO-2S (Novic, Russia) high energy planetary ball mill for 5, 15, 30, and 60 min. The steel vial and grinding balls were used, and the weights of the material and balls were in the ratio 1:10. The milling was performed at 2220 rpm rotation speed for vial in the dry state at room temperature. The X-ray diffraction analysis was carried out using ARL X’TRA (Switzerland) diffractometer with a semiconductor Si (Li) Peltier detector and CuKα radiation. XRD patterns were measured in the range 2θ = (10–70)° with scanning rate of 0.02° s−1 and were processed by the full profile analysis using the Powder Cell 2.5 software, where the pseudo-Voigt profile function was used. Phases were identified by the PDF-4+ powder database of the International Center for Diffraction Data (ICDD). The particle size distribution was analysed by laser diffraction using a Fritsch Analysette 22 MicroTec Plus analyzer. From the BrunauerEmmett-Teller (BET), the specific surface area (SSA) for powders was estimated. The microstructure of un-milled and milled samples was examined by Hitachi TM-3000 scanning electron microscope. Simultaneous thermogravimetric (TG) and calorimetric (DSC) analyses were performed using a STA 449C Jupiter (Netzsch-Gerätebau GmbH, Germany) thermal analyzer. Sample with a mass of ca. 10 mg was placed in the alumina crucible and was heated up to 800 °C in air with a heating rate of 5 °C min−1. A part of the equipment including a heating furnace, a sample holder, and thermocouple is schematically shown in Fig. 1. In addition, the samples were analysed by thermomagnetometry method, which is thermogravimetric analysis in magnetic field. For this, the two permanent magnets (Netzsch-Gerätebau GmbH) were attached on the outer side of the measurement cell so that the sample is located in a magnetic field of ca. 5 Oe. The Netzsch Proteus software packages were used for data analysis.
Fig. 2. XRD patterns for initial reagents.
3. Results and discussion 3.1. X-ray diffraction analysis Fig. 2 shows XRD patterns for α-Fe2O3 (PDF No. 40-142), Li2CO3 (PDF No. 66-941) and ZnO (PDF No. 26-170) initial reagents. The Fe2O3 peaks at 2θ ≈ 30.3°, 43° correspond to γ-Fe2O3 spinel phase (PDF No. 79-196), which is added in a small amount (ca. 2 mass%) to the iron oxide powders in the case of ferrites production. Fig. 3a presents the XRD patterns for Li2CO3-Fe2O3-ZnO mixture mechanically activated for various milling times. The values of the lattice parameters obtained by the Powder Cell 2.5 software and crystallite sizes calculated using the Williamson-Hall method, are summarized in Table 1. Measurement
Fig. 3. XRD patterns of Fe2O3–Li2CO3–ZnO (a) mixture and Fe2O3 (b) powder milled for various milling times.
Fig. 1. A schematic of TG/DSC block of thermal analyzer. 101
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
Table 1 XRD and BET data for Li2CO3-Fe2O3-ZnO mixture. Sample
Composition
Lattice parameter (nm)
Crystallite sizes (nm)
Phase concentration (mass%)
Specific surface area (m2 g−1)
Initial reagets
α-Fe2O3 Li2CO3 ZnO Spinel (γ-Fe2O3) α-Fe2O3 Li2CO3 ZnO Spinel (γ-Fe2O3) α-Fe2O3 Li2CO3 ZnO Spinel (Fe3O4) α-Fe2O3 Li2CO3 ZnO Spinel (Fe3O4) α-Fe2O3 Li2CO3 ZnO Spinel (Fe3O4) α-Fe2O3 Spinel (Fe3O4)
а = b = 0.5033; c = 1.3753 а = 0.8259; b = 0.4974; c = 0.6198 а = b = 0.3248; c = 0.5207 a = b = c = 0.8344 а = b = 0.5035; c = 1.3760 а = 0.8443; b = 0.5074; c = 0.6094 а = b = 0.3251; c = 0.5211 a = b = c = 0.8348 а = b = 0.5036; c = 1.3767 а = 0.8368; b = 0.4949; c = 0.6112 а = b = 0.3286; c = 0.5229 a = b = c = 0.8392 а = b = 0.5044; c = 1.3779 а = 0.8346; b = 0.5074; c = 0.6094 а = b = 0.3316; c = 0.5233 a = b = c = 0.8396 а = b = 0.5058; c = 1.3783 а = 0.8338; b = 0.4990; c = 0.6218 а = b = 0.3350; c = 0.5307 a = b = c = 0.8403 а = b = 0.5043; c = 1.3778 a = b = c = 0.8391
89 159 81 42 72 52 41 42 65 32 32 28 43 18 23 27 21 12 23 22 17 20
84.05 6.64 7.31 2.00 84.05 6.64 7.31 2.00 78.49 6.63 7.28 7.60 72.77 6.63 7.30 13.30 66.78 6.63 7.29 19.30 69.90 30.10
8.72 1.53 4.96 – 8.81
Mixture 5 min milling
Mixture 15 min milling
Mixture 30 min milling
Mixture 60 min milling
Fe2O3 120 min milling
9.24
14.01
24.82
27.71
mixture ball milled for 60 min, no Li2CO3 powder morphology was observed indicating that the ball milling produced very fine particles of lithium carbonate. One can observed that the agglomerates of milled mixture are characterized by denser structure compared to un-milled one. Moreover, from SEM images with 200× magnification, which are not presented in the article, the milled mixture exhibits larger agglomerates. These results are very important to explain the subsequent results of the reactivity of investigated mixtures.
errors for lattice parameters are ca. 5%. The data of XRD analysis for un-milled mixture (ball milling for 0 min) indicate the presence of α-Fe2O3, Li2CO3, ZnO and γ-Fe2O3 phases (marked reflections) for Li2CO3-Fe2O3-ZnO composition. According to the XRD data for mechanically activated mixtures, the broadening of the peaks due to decreased crystallite sizes is observed as a result of the ball milling. From Table 1, the α-Fe2O3 concentration reduces while the Li2CO3 and ZnO concentrations remain unchanged. It was also revealed that the spinel phase concentration increases with increasing the milling time. According to lattice parameters of γ-Fe2O3 (a = b = c = 0.8347 nm) and Fe3O4 (a = b = c = 0.8400 nm) from Powder Diffraction Standards cards, this phase in the samples milled for 15 min and more presumably belongs to Fe3O4 (PDF No. 36-314), whose formation during ball milling of α-Fe2O3 was also confirmed in [30]. To support this assumption, we have analysed the α-Fe2O3 powder milled for 120 min in a planetary ball mill in the air. Both the α-Fe2O3 and Fe3O4 phases were identified on the XRD pattern (Fig. 3b), and the amount of Fe3O4 phase is ca. 30.1 mass% (Table 1). The results of the measurement of the magnetic properties of this powder showed an increase in the saturation magnetization due to the formation of Fe3O4 magnetic phase. The crystallite size of the phases presented in the milled mixture is shown in Table 1. The reduction in the crystallite size of all reagents is intensive up to 30 min of milling and changes slightly with prolonged milling time.
3.3. Laser diffraction and BET analyses The particle size distributions of the initial reagents, mixtures unmilled and milled for 60 min were measured by laser diffraction and are shown in Fig. 5. Particle size distribution specifications (D10, D50, and D90 values) based on normalized particle amount analysis are summarized in Table 2. The total measurement error of particle size is ± 10%. For the hematite, the particle size is characterized by bimodal distribution with the small particles in the 0.2 − 3 μm range and a significant fraction of large particles in the 3 − 40 μm range. The average particle size of Fe2O3 powder is 12.1 μm. The ZnO reagent with the average particle size of 3.1 μm has a distribution centred at ca. 2 μm with a greater shoulder towards higher particle dimensions. The Li2CO3 reagent has a wide particle size distribution that consists of two significant fractions of particles in the 3 − 50 μm and 50 − 300 μm ranges. The average particle size of Li2CO3 powder is 47.2 μm. It can be seen that the distribution for un-milled mixture has a behavior similar to the distribution of the Fe2O3. The ball milled mixture is characterized by a wide particle size distribution which differs from the un-milled mixture by the presence of significant fractions of larger particles in the 50 − 100 μm range, indicating the agglomeration effect after mechanical milling the Li2CO3-Fe2O3-ZnO mixture. This assumption is confirmed by an increase in the average particle size of milled mixture (see Table 2). Table 1 shows the specific surface area for initial powders (the first row in the last column) and the average specific surface for all particles in milled mixtures (other rows). The error of SSA is +/− 3%. According to the BET data, the average SSA significantly increases in the mixtures milled for more than 15 min, indicating a strong decrease in their average particle size. Thus, the different behavior of the average particle size obtained from laser diffraction and BET analyses indicates that the ball milling of
3.2. SEM analysis Microstructural analysis was carried out for all samples. However, the article presents only SEM images for initial reagents, mixtures unmilled and milled for 60 min (Fig. 4). The SEM micrographs of samples milled for 5, 15, and 30 min showed similar results. The Fe2O3 image shows microstructure with spherical small particles in addition to the larger particles (Fig. 4). The diameter of Fe2O3 particles ranged from 0.2 to 5 μm. Rod-like and flake-like morphologies were observed in Li2CO3 and ZnO powders, respectively. By means of image processing, the average particle sizes estimated from the geometric analysis are found to be approximately 2, 9 and 3 μm for αFe2O3, Li2CO3 and ZnO, respectively. As shown in Fig. 4, the presence of the Li2CO3 particles is observed on the SEM image for α-Fe2O3-Li2CO3-ZnO un-milled mixture. As for 102
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
Fig. 4. SEM micrographs of initial reagents, mixtures un-milled and milled for 60 min.
transition (α → b) in Li0.5Fe2.5O4 phase that is formed as intermediate phase during the synthesis of LiZn ferrite [29,32]. As for milled Li2CO3-Fe2O3-ZnO mixtures, the mass loss process starts at ca. 100 °C. By increasing the milling time, the end temperature of this process shifts towards lower temperatures more intensively with an increase in the milling time to 15 min (Fig. 6), and then varies slightly with further increasing milling time. The mass loss process can be divided into two ranges. The first range is 100 − 350 °C and the second one is 350 − 600 °C. A slight mass loss process was observed on TG curve in the first range for sample milled 5 min, while the main mass loss occurs in the second temperature range. On the contrary, samples milled for prolonged milling time are characterized by a significant mass loss at low temperatures. The main total mass loss values were higher than calculated value for CO2 evaporation. We have previously shown in [29] using mass spectrometry analysis that only carbon dioxide was released during the reaction from unmilled mixture, while the output of both CO2 and H2O were observed during the reaction from milled mixtures. One can notice that the ball milled mixture have an enhanced ability to adsorb gases from the atmosphere like H2O which evaporate during heating at low temperatures, leading to a higher mass loss than the expected value calculated according to reaction. According to calorimetric analysis, the DSC curves show several peaks. There is a complex peak in the low-temperature range associated with the imposition of the endothermal and exothermal reactions.
mixture in a planetary mill decreases the average particle size of initial reagents and simultaneously increases the free energy of the system resulting in formation of agglomerates. 3.4. Thermogravimetric and calorimetric analyses Fig. 6 shows the TG, derivative TG (DTG) and DSC curves for the Li2CO3-Fe2O3-ZnO reagent mixture mechanically activated for various milling times. For the un-milled mixture (0 min milling), the thermal behavior of sample is typical in lithium-zinc ferrite formation during non-isothermal synthesis [31]. The mass loss process starts at a temperature of 500 °C and, as shown by DTG curve, comprises two steps, which correspond to the interaction between reagents with the Li2CO3 decomposition in the temperature range 500 − 650 °C and the melting of lithium carbonate residues at ca. 724 °C. The shape of the DTG curve is similar to the DSC curve, which shows that the mass change in the temperature range 500 − 730 °C can be attributed to the observed endothermic effects (↑exo in Fig. 6a indicates the direction of DSC peaks observed; the peak above the baseline means exothermic effect and below the baseline − endothermic one). The total mass loss was 4.02% and this value is close to the expected value for CO2 releasing calculated from Eq. (1) (3.95%). As was shown in [29] by using mass spectrometry analysis of gases, only CO2 is released in this temperature range. A small DSC peak at ca. 750 °C is related to order − disorder phase 103
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
Furthermore, the area of exothermal DSC peak in the temperature range 200–300 °C, which is most likely associated with the release of H2O, increases with increasing the milling time. Another DSC peak is observed in the temperature range 400 − 600 °C for the 5 min milled mixture and it shifts toward lower temperatures for samples milled for prolonged milling time. Since the area of this DSC peak and mass loss corresponding to this range decrease by increasing the milling time, the significant part of the reaction (Eq. (1)) occurs prior to the onset of the second range. It can be seen from Fig. 7, the end temperature of Li2CO3 decomposition, which was estimated from TG/DSC analysis, greatly decreases with increasing the milling time up to 15 min and then changes slightly at prolonged milling times. Consequently, mechanical activation of the reagent mixture considerably enhances the reactivity of the Li2CO3Fe2O3-ZnO system and shifts the reaction toward lower temperatures. 3.5. Thermomagnetometric analysis Fig. 8 shows the schematic of TG/DTG curves behavior during heating of magnetic sample in thermal analyzer. It is evident that if no change in a mass of sample occur in a certain temperature range (for example, due to chemical reactions), then the TG curve does not change without a magnetic field (Fig. 8, dashed TG curve). The behavior of TG curve under magnetic field, TG(M), demonstrates the presence of mass jump at temperature of ferrimagnet-paramagnet transition (Curie temperature) for the magnetic phase (Fig. 8, solid TG curve). The DTG curve as a derivative TG(M) curve, DTG(M), allows to determine the Curie temperature of this transition and hence to characterize magnetic phase in the sample [33]. First of all, we analysed the un-milled and milled for 60 min Fe2O3 reagent by thermomagnetometry method to examine the formation of new phases after mechanical milling of the main component of mixture (Fig. 9). During the experiment, the magnets were attached and removed on the outer side of the measurement cell at regular intervals, which enables simultaneous analysis of both TG and TG(M) curves. Fig. 9a presents the thermal behavior of initial iron oxide. TG curve indicates the small mass jumps corresponding to low magnetization of α-Fe2O3, which disappear after magnetic phase transition at ca. 678 °C that is Curie temperature of hematite. The thermal analysis result for the iron oxide milled for 60 min is shown in Fig. 9b. TG curve presents a significant increase in the mass of the sample in the temperature range 100–200 °C. It can be assumed that if Fe3O4 forms during the milling of α-Fe2O3, then a change in mass can be observed due to the oxidation of Fe3O4 to Fe2O3 when the sample is heated [34]. According to the increase in mass by 0.96% (Fig. 9b, TG curve) due to oxidation by the reaction 4Fe3O4 + O2 → 6Fe2O3, the calculated value of Fe3O4 content in the total iron oxide powder was 28 mass%, and this value is close to the XRD analysis (Table 1) where about 30.1 mass% of spinel Fe3O4 was estimated. In this temperature range the DSC curve shows an exothermal peak, reflecting the process of oxidation of the sample. It is known that the oxidation of Fe3O4 to α-Fe2O3 proceeds through the formation of intermediate γ-Fe2O3 phase under the scheme: Fe3O4 → γ-Fe2O3 → α-Fe2O3 [35,36]. As well known that both Fe3O4 and γ-Fe2O3 have a high magnetization [37,38], so we can observe high mass jumps on TG curve from room temperature to 460 °C (Fig. 9b, TG curve). After this, TG and DSC curves show γ-Fe2O3→α-Fe2O3 transition at ca. 458 °C, which is accompanied by a change in mass of sample only due to applying a magnetic field. Thus, the above results confirm the formation of the Fe3O4 phase in α-Fe2O3 reagent during the mechanical activation of the Li2CO3-Fe2O3ZnO mixture. Fig. 6b shows the thermomagnetometric analysis results of mixtures, which are obtained at cooling regime where the magnets were attached immediately after the heating mode. No mass changes were observed for samples milled for 0 and 5 min. TG(M) and DTG(M) for the
Fig. 5. Particle size distribution of initial reagents, un-milled and milled for 60 min mixtures. Table 2 Particle size distribution from laser diffraction analysis. Composition
D10 (μm)
D50 (μm)
D90 (μm)
α-Fe2O3 Li2CO3 ZnO Mixture un-milled Mixture milled
0.5 5.7 0.9 0.4 0.9
12.1 47.2 3.1 9.6 12.8
25.6 212.9 14.6 24.6 68.3
104
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
Fig. 6. Thermal analysis of Fe2O3–Li2CO3–ZnO mixture milled for various milling times: (a) – heating regime, (b) – cooling regime.
Fig. 7. End temperature of Li2CO3 decomposition vs. milling time.
Fig. 9. Thermal analysis of Fe2O3 powder un-milled (a) and milled for 120 min (b).
analyzer. Comparing the thermomagnetometric analysis for un-milled and milled mixtures (see Fig. 6b), it can be deduced that the reactivity of the system increases strongly when we use the activation time of 15 min and more. In general, the reactivity of the solid-phase system can increase due to several factors including the reduction in particle size and the formation of particle structure defects during mechanical activation. On the other hand, the γ-Fe2O3 formation in the temperature range 200–400 °C, which is due to the oxidation of Fe3O4 under the scheme Fe3O4 → γ-Fe2O3 → α-Fe2O3, can also increase the reactivity of the
Fig. 8. A schematic of thermal behavior of magnetic sample in and without magnetic field.
remaining samples show mass changes at ca. 493 °C, and according to data in [29], this temperature value correspond to magnetic phase transition at the Curie temperature in Li0.4Fe2.4Zn0.2O4 ferrite phase. One can observe that the height of the mass jump increases with increasing the milling time and this is most likely due to increasing the ferrite phase formation in reaction mixture during its heating in thermal 105
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
and more at 2220 rpm rotation speed for vial.
• It can be assumed that the observed agglomerates consist of well-
interlinked small size particles of reagents, which leads to an increase in contact area between reagents and the acceleration of the chemical reaction.
Acknowledgments This work was supported by The Ministry of Education and Science of the Russian Federation in part of the Science program (project 11.980.2017/4.6). The measurements of X-ray diffraction analysis data were funded from Tomsk Polytechnic University Competitiveness Enhancement Program grant.
Fig. 10. Model of the un-milled and milled particles.
References Li2CO3-Fe2O3-ZnO mixture and shift the reaction towards lower temperatures. Additional experiments showed that the solid-state reaction of the mixture, consisting of powders milled separately, does not show any significant difference as compared to those of the as-milled samples. Only ball milling of the whole mixture leads to reactivity increase. Since, in conventional solid-phase synthesis, contacts of three or more particles of different powders in mixture are unlikely, reactions in such systems proceed through intermediate stages in which pairs of particles in contact take part. As a result of the plastic deformation of solid particles during ball milling of powders in a planetary mill, the point contacts are transformed into contacts along the surface, and thus the formation of triple contacts becomes possible. This leads to a decrease in the number of intermediate stages, acceleration of the synthesis process and homogeneity of the final product. Thus, for milled Li2CO3-Fe2O3-ZnO mixture, it can be assumed that the agglomerates observed from SEM images consist of well-interlinked small size particles of reagents (see Fig. 10), which leads to an increase in contact area between reagents and the acceleration of the chemical reaction.
[1] P.D. Baba, G.M. Argentina, W.E. Courtney, G.F. Dionne, D.H. Temme, Fabrication and properties of microwave lithium ferrite, IEEE Trans. Magn. 8 (1972) 83–94. [2] Parveen Kumar Dipti, J.K. Juneja, Sangeeta Singh, K.K. Raina, Chandra Prakash, Improved dielectric and magnetic properties in modified lithium–ferrites, Ceram. Int. 41 (2015) 3293–3297. [3] N. Rezlescu, C. Doroftei, E. Rezlescu, P.D. Popa, Lithium ferrite for gas sensing applications, J. Sens. Actuators 133 (2008) 420. [4] W. Cook, M. Manley, Raman characterization of α- and β-LiFe5O8 prepared through a solid-state reaction pathway, J. Solid State Chem. 183 (2010) 322–326. [5] P.B. Brown, A superstructure in spinels, Nature 170 (1952) 1123. [6] J. Smit, H.P.J. Wijn, Ferrites Physical Properties of Ferrimagnetic Oxides in Relation to Their Technical Applications, Philips Technical Library, Eindhoven, 1959. [7] C.E. Patton, C.A. Edmondson, Y.H. Liu, Magnetic properties of lithium zinc ferrite, J. Appl. Phys. 53 (1982) 2431–2433. [8] A. Gruskova, V. Jancarik, J. Slama, R. Dosoudil, Effect of Zn-Ti substitution on electromagnetic properties of Li ferrites, J. Magn. Magn. Mater. 304 (2006) 762–765. [9] G.M. Argentina, P.D. Baba, Microwave lithium ferrites: an overview: IEEE trans, Microwave Theory Tech. 22 (1974) 652–658. [10] V. Verma, S.P. Gairola, V. Pandey, J.S. Tawale, H. Su, R.K. Kotanala, High permeability and low power loss of Ti and Zn substitution lithium ferrite in high frequency range, J. Magn. Magn. Mater. 321 (2009) 3808–3812. [11] B. Edelio, Tle. Mercier, M. Quarton, Microstructure and physicochemical studies of pure and zinc-substituted lithium ferrites sinteredabove 1000 °C, J. Am. Ceram. Soc. 78 (1995) 365–368. [12] S.S. Teixeira, M.P.F. Graça, L.C. Costa, Dielectric, morphological and structural properties of lithium ferrite powders prepared by solid state method, J. NonCrystalline Solids 358 (2012) 1924–1929. [13] D.H. Ridgley, H. Lessoff, J.V. Childress, Effects of lithium and oxygen losses on magnetic and crystallographic properties of spinel lithium ferrite, J. Am. Ceram. Soc. 53 (1971) 304–311. [14] M. Tabuchi, K. Ado, H. Sakaebe, C. Masquelier, H. Kageyama, O. Nakamura, Preparation of AFeO2 (A = Li Na) by hydrothermal method, Solid State Ion. 79 (1995) 220. [15] V.K. Sankaranarayanan, Q.A. Pankhurst, D.P.E. Dickson, C.E. Johnson, Synthesis and characterization of ultrafine lithium ferrite from a citrate precursor, J. Magn. Magn. Mater. 130 (1994) 288–292. [16] P. Vijaya Bhasker Reddy, B. Ramesh, Ch. Gopal Reddy, Electrical conductivity and dielectric properties of zinc substituted lithium ferrites prepared by sol-gel method, Physica B 405 (2010) 1852–1856. [17] H. Anwar, A. Maqsood, Microwave magnetic and absorption properties of Li0.5Mnx/ 2Zn0.75-x/2Fe2O4 soft nano ferrites prepared by sol-Gel auto combustion method, Electron. Mater. Lett. 9 (2013) 641–647. [18] S.N. Hosseini, M.H. Enayati, F. Karimzadeh, N.M. Sammes, Formation mechanism, crystallite growth and electrical conductivity of nano-crystalline CuxFe3-xO4 (0.75 ≤ x ≤ 1.25) spinels prepared by glycine-nitrate process, Thermochim. Acta 639 (2016) 91–97. [19] M. Stoia, E. Muntean, C. Păcurariu, C. Mihali, Thermal behavior of MnFe2O4 and MnFe2O4/C nanocomposite, Thermochim. Acta 652 (2017) 1–8. [20] A.K. Giri, Nanocrystalline materials prepared through crystallization by ball milling, Adv. Mater. 9 (1997) 163–166. [21] V.V. Boldyrev, Mechanochemistry and mechanical activation of solids, Russ. Chem. Rev. 75 (2006) 177–189. [22] V. Berbenni, C. Milanese, G. Bruni, A. Marini, I. Pallecchi, Synthesis and magnetic properties of ZnFe2O4 obtained by mechanochemically assisted low-temperature annealing of mixtures of Zn and Fe oxalates, Thermochim. Acta. 447 (2006) 184–189. [23] V. Berbenni, G. Bruni, C. Milanese, A. Girella, A. Marini, Synthesis and characterization of LaFeO3 powders prepared by a mixed mechanical/thermal processing route, J. Therm. Anal. Calorim. (2017) 1–7, http://dx.doi.org/10.1007/s10973017-6878-z First Online: 11 December 2017. [24] V. Berbenni, A. Marini, P. Matteazzi, R. Ricceri, N. Welham, Solid-state formation of lithium ferrites from mechanically activated Li2CO3-Fe2O3 mixtures, J. Eur. Ceram. Soc. 23 (2003) 527–536. [25] H.M. Widatallah, C. Johnson, F.J. Berry, The influence of ball milling and
4. Conclusions As a result of studies of microstructure and reactivity of Li2CO3Fe2O3-ZnO mixture ball milled in a planetary mill, the following conclusions were made.
• It was established from XRD analysis that the composition of reagent
• •
• •
mixture changes during mechanical milling. At increased milling time, the α-Fe2O3 concentration reduces while the Li2CO3 and ZnO concentration remains unchanged. It was also revealed that the Fe3O4 spinel phase is formed during mechanical milling of mixture in a planetary mill and the spinel phase concentration increases with increasing the milling time. The thermomagnetometric analysis of α-Fe2O3, which was subjected to a long milling, confirm the formation of the Fe3O4 phase in it. Consequently, this phase is formed in the Li2CO3-Fe2O3-ZnO mixture during mechanical activation. The crystallite sizes of initial reagents were found to decrease with increasing the milling time. Moreover, BET analysis indicates a significant increase in specific surface area of mechanically activated mixtures resulting in a high ability to adsorb water from atmosphere, which evaporates during sample heating at low temperatures. From SEM and laser diffraction analyses, an increase in size of agglomerates in the mechanically activated mixture is observed. Thus, a ball milling of mixture in a planetary mill decreases the average particle size of reagents and simultaneously increases the free energy of the system resulting in formation of agglomerates. It was found that the reactivity of the mechanically activated mixture increases strongly when we use the activation time of 15 min 106
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
[26]
[27]
[28]
[29]
[30]
[31]
subcequent calcination on the formation of LiFeO2, J. Mater. Sci. 37 (2002) 4621–4625. A.P. Surzhikov, E.N. Lysenko, A.V. Malyshev, A.M. Pritulov, O.G. Kazakovskaya, Influence of mechanical activation of initial reagents on synthesis of lithium ferrite, Russ. Phys. J. 6 (2012) 627–677. M. Kavanlooee, B. Hashemi, H. Maleki-Ghaleh, J. Kavanlooee, Effect of annealing on phase evolution, microstructure, and magnetic properties of nanocrystalline ballmilled LiZnTi ferrite, J. Electron. Mater. 41 (2012) 3082–3086. H.M. Widatallah, X.L. Ren, I.A. Al-Omari, The influence of TiO2 polymorph, mechanical milling and subsequent sintering on the formation of Ti-substituted spinelrelated Li0.5Fe2.5O4, J. Mater. Sci. 41 (2006) 6333–6338. E.N. Lysenko, A.P. Surzhikov, V.A. Vlasov, A.V. Malyshev, E.V. Nikolaev, Thermal analysis study of solid-phase synthesis of zinc- and titanium- substituted lithium ferrites from mechanically activated reagents, Therm. Anal. Calorim. 122 (2015) 1347–1353. M. Zdujić, C. Jovalekić, Lj. Karanović, M. Mitrić, The ball milling induced transformation of α-Fe2O3 powder in air and oxygen atmosphere, Mater. Sci. Eng. A 262 (1999) 204–213. A.P. Surzhikov, A.M. Pritulov, E.N. Lysenko, A.N. Sokolovskii, V.A. Vlasov, E.A. Vasendina, Dependence of lithium–zinc ferrospinel phase composition on the
[32]
[33] [34]
[35] [36] [37]
[38]
107
duration of synthesis in an accelerated electron beam, J. Therm. Anal. Calorim. 110 (2012) 733–738. A.P. Surzhikov, T.S. Frangulyan, S.A. Ghyngazov, E.N. Lysenko, Investigation of structural states and oxidation processes in Li0.5Fe2.5O4-δ using TG analysis, J. Therm. Anal. Calorim. 108 (2012) 1207–1212. P.K. Gallagher, Thermomagnetometry, J. Therm. Anal. Calorim. 49 (1997) 33–44. E.N. Lysenko, A.P. Surzhikov, S.P. Zhuravkov, V.A. Vlasov, A.V. Pustovalov, N.A. Yavorovskii, The oxidation kinetics study of ultrafine iron powders by thermogravimetric analysis, J. Therm. Anal. Calorim. 115 (2014) 1447–1452. J.P. Sanders, P.K. Gallagher, Kinetics of the oxidation of magnetite using simultaneous TG/DSC, J. Therm. Anal. Calorim. 72 (2003) 777–789. J.P. Sanders, P.K. Gallagher, Thermomagnetometric evidence of γ-Fe2O3 as an intermediate in the oxidation of magnetite, Thermochim. Acta 406 (2003) 241–243. T. Osaka, T. Matsunaga, T. Nakanishi, A. Arakaki, D. Niwa, H. Iida, Synthesis of magnetic nanoparticles and their application to bioassays, Anal. Bioanal. Chem. 384 (2006) 593–600. A.C. Grigorie, C. Muntean, M. Stefanescu, Obtaining of γ-Fe2O3 nanoparticles by thermal decomposition of polyethyleneglycol–iron nitrate mixtures, Thermochim. Acta 621 (2015) 61–67.
Contents lists available at ScienceDirect
Thermochimica Acta journal homepage: www.elsevier.com/locate/tca
Microstructure and reactivity of Fe2O3-Li2CO3-ZnO ferrite system ball-milled in a planetary mill
T
⁎
Elena Lysenko , Evgeniy Nikolaev, Vitaliy Vlasov, Anatoliy Surzhikov Tomsk Polytechnic University, Lenina Avenue 30, 634050 Tomsk, Russia
A R T I C LE I N FO
A B S T R A C T
Keywords: Substituted lithium ferrite LiZn ferrite Mechanical activation Microstructure Reactivity Thermal analysis
In this work, the microstructure of mechanically activated Fe2O3-Li2CO3-ZnO mixture for the lithium-zinc ferrites production was studied using the Brüner, Emmett, Teller and laser diffraction methods as well as X-ray diffraction and scanning electron microscopy analyses. The reactivity of reagent mixture was investigated by thermogravimetric and calorimetric analyses. The ball milling was performed in a AGO-2S high energy planetary ball mill with a vial rotation speed of 2220 rpm using steel grinding balls. The milling times were 0, 5, 15, 30 or 60 min. It was shown that the composition of mixture changes during the ball milling, which consists in decreasing the α-Fe2O3 concentration and increasing the Fe3O4 spinel phase, while the Li2CO3 and ZnO concentrations remain unchanged. It was found that the milling leads to decrease in the average particle size of the reagents and simultaneously formation of large size agglomerates with denser structure and well-interlinked particles. It was established that observed changes in microstructure and phase composition lead to an increase in the reactivity of the Fe2O3-Li2CO3-ZnO system and the acceleration of the chemical reaction between reagents.
1. Introduction The Li2O-Fe2O3 system is of considerable interest for production of a variety of materials such as ferrites with different objective functions, catalysts, solid-oxide electrodes [1–3]. According to the equilibrium phase diagram, the Li2O-Fe2O3 system may form the LiFeO2 lithium orthoferrite (γ – low-temperature high-ordered form and α – hightemperature disordered form; the temperature of γ↔α transition is ca. 670 °C) and the Li0.5Fe2.5O4 lithium ferrite with spinel structure [4,5]. The latter has the structure of an inverse spinel Fe3+[Li0.5+Fe1.53+]O4, and at T < 750 °C, lithium and iron ions are ordered in the Fd-3m (αphase) type in the octahedral sublattice. However, disordered form of lithium ferrite has a random statistical distribution of lithium and iron over all the octahedral positions. The properties of lithium ferrites can be tailor made by substituting them with different metal ions such as Zn2+, Cu2+, Ti4+, Mg2+, Co2+, Ni2+, etc. Zinc is known to play a decisive role in determining the ferrite properties [6–8]. The redistribution of metal ions over the tetrahedral and octahedral sites in the spinel lattice on incorporation of zinc is responsible for the modification of ferrite properties. Thus, substituted lithium ferrites with chemical composition of Li0.5(13+ are substituted by ions of zinc, x)ZnxFe2.5-0.5xO4, in which the ions Fe are characterized by high saturation magnetization and this promotes their wide application in microwave technology [9,10].
⁎
The conventional way to synthesize lithium ferrites involves the use of high-temperature solid-state reaction including the mixing of oxide/ carbonate and then mixtures annealing at high temperatures [11,12]. Such synthesis method has some disadvantages which include chemical inhomogeneity, coarser particle size and volatility of Li2O from ferrites during synthesis [13]. In order to overcome these drawbacks, a number of chemical methods were developed to prepare ferrites at low temperatures. They include the hydrothermal method, the citrate precursor and sol gel methods and etc. [14–19]. However, many of these methods have major drawback which consists in applying the repeated high temperature to form the monophase of lithium ferrite. One of the methods that allow to reduce the synthesis temperature is preliminary mechanical activation of ferrite reagents in high energy ball mills [20–23]. Such powders are highly reactive, which makes it possible to obtain reaction products at lower temperatures and shorter thermal treatment period. According to earlier investigation, mechanical activation can be used to produce lithium [24–26] and substituted lithium [27,28] ferrites at lower temperature than in conventional ceramic processing. In our previous work [29], the influence of mechanical activation of Li2CO3-Fe2O3-ZnO and Li2CO3-TiO2-Fe2O3 initial reagent mixtures on the solid-phase synthesis of Li0.4Fe2.4Zn0.2O4 and Li0.6Fe2.2Ti0.2O4 substituted lithium ferrites was studied. It was established that the ferrites can be obtained from mechanically activated
Corresponding author. E-mail addresses: [email protected] (E. Lysenko), [email protected] (E. Nikolaev), [email protected] (V. Vlasov), [email protected] (A. Surzhikov).
https://doi.org/10.1016/j.tca.2018.04.015 Received 9 February 2018; Received in revised form 27 April 2018; Accepted 28 April 2018 Available online 01 May 2018 0040-6031/ © 2018 Elsevier B.V. All rights reserved.
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
mixtures at temperatures at least 200 °C lower than in the case of using un-milled mixtures. Although a substantial literature has accumulated on the use of mechanical milling for ferrite formation, there is only limited information on the microstructure of ferrite powders. In this work, using the Brüner, Emmett, Teller (BET) and laser diffraction methods, as well as XRD and SEM analysis, the microstructure of un-milled and ballmilled initial reagents for the lithium-zinc ferrites production was studied. The reactivity of reagent mixtures was investigated by thermogravimetric and calorimetric analyses.
2. Experimental The initial reagents were commercially Fe2O3, Li2CO3 and ZnO (99%, Reahim Co., Russia) powders which were pre-dried and mixed in molar ratios of 86.05 mass%, 6.64 mass% and 7.31 mass%, respectively, according to the reaction Li2СO3 + 6Fe2O3 + ZnO→5 Li0.4Fe2.4Zn0.2O4 + СО2↑
(1)
The mixture was divided into few batches; the samples of first batch were simply mixed in an agate mortar and are considered to be simple un-milled mixture. The rest of the powder was subjected to mechanical activation using a AGO-2S (Novic, Russia) high energy planetary ball mill for 5, 15, 30, and 60 min. The steel vial and grinding balls were used, and the weights of the material and balls were in the ratio 1:10. The milling was performed at 2220 rpm rotation speed for vial in the dry state at room temperature. The X-ray diffraction analysis was carried out using ARL X’TRA (Switzerland) diffractometer with a semiconductor Si (Li) Peltier detector and CuKα radiation. XRD patterns were measured in the range 2θ = (10–70)° with scanning rate of 0.02° s−1 and were processed by the full profile analysis using the Powder Cell 2.5 software, where the pseudo-Voigt profile function was used. Phases were identified by the PDF-4+ powder database of the International Center for Diffraction Data (ICDD). The particle size distribution was analysed by laser diffraction using a Fritsch Analysette 22 MicroTec Plus analyzer. From the BrunauerEmmett-Teller (BET), the specific surface area (SSA) for powders was estimated. The microstructure of un-milled and milled samples was examined by Hitachi TM-3000 scanning electron microscope. Simultaneous thermogravimetric (TG) and calorimetric (DSC) analyses were performed using a STA 449C Jupiter (Netzsch-Gerätebau GmbH, Germany) thermal analyzer. Sample with a mass of ca. 10 mg was placed in the alumina crucible and was heated up to 800 °C in air with a heating rate of 5 °C min−1. A part of the equipment including a heating furnace, a sample holder, and thermocouple is schematically shown in Fig. 1. In addition, the samples were analysed by thermomagnetometry method, which is thermogravimetric analysis in magnetic field. For this, the two permanent magnets (Netzsch-Gerätebau GmbH) were attached on the outer side of the measurement cell so that the sample is located in a magnetic field of ca. 5 Oe. The Netzsch Proteus software packages were used for data analysis.
Fig. 2. XRD patterns for initial reagents.
3. Results and discussion 3.1. X-ray diffraction analysis Fig. 2 shows XRD patterns for α-Fe2O3 (PDF No. 40-142), Li2CO3 (PDF No. 66-941) and ZnO (PDF No. 26-170) initial reagents. The Fe2O3 peaks at 2θ ≈ 30.3°, 43° correspond to γ-Fe2O3 spinel phase (PDF No. 79-196), which is added in a small amount (ca. 2 mass%) to the iron oxide powders in the case of ferrites production. Fig. 3a presents the XRD patterns for Li2CO3-Fe2O3-ZnO mixture mechanically activated for various milling times. The values of the lattice parameters obtained by the Powder Cell 2.5 software and crystallite sizes calculated using the Williamson-Hall method, are summarized in Table 1. Measurement
Fig. 3. XRD patterns of Fe2O3–Li2CO3–ZnO (a) mixture and Fe2O3 (b) powder milled for various milling times.
Fig. 1. A schematic of TG/DSC block of thermal analyzer. 101
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
Table 1 XRD and BET data for Li2CO3-Fe2O3-ZnO mixture. Sample
Composition
Lattice parameter (nm)
Crystallite sizes (nm)
Phase concentration (mass%)
Specific surface area (m2 g−1)
Initial reagets
α-Fe2O3 Li2CO3 ZnO Spinel (γ-Fe2O3) α-Fe2O3 Li2CO3 ZnO Spinel (γ-Fe2O3) α-Fe2O3 Li2CO3 ZnO Spinel (Fe3O4) α-Fe2O3 Li2CO3 ZnO Spinel (Fe3O4) α-Fe2O3 Li2CO3 ZnO Spinel (Fe3O4) α-Fe2O3 Spinel (Fe3O4)
а = b = 0.5033; c = 1.3753 а = 0.8259; b = 0.4974; c = 0.6198 а = b = 0.3248; c = 0.5207 a = b = c = 0.8344 а = b = 0.5035; c = 1.3760 а = 0.8443; b = 0.5074; c = 0.6094 а = b = 0.3251; c = 0.5211 a = b = c = 0.8348 а = b = 0.5036; c = 1.3767 а = 0.8368; b = 0.4949; c = 0.6112 а = b = 0.3286; c = 0.5229 a = b = c = 0.8392 а = b = 0.5044; c = 1.3779 а = 0.8346; b = 0.5074; c = 0.6094 а = b = 0.3316; c = 0.5233 a = b = c = 0.8396 а = b = 0.5058; c = 1.3783 а = 0.8338; b = 0.4990; c = 0.6218 а = b = 0.3350; c = 0.5307 a = b = c = 0.8403 а = b = 0.5043; c = 1.3778 a = b = c = 0.8391
89 159 81 42 72 52 41 42 65 32 32 28 43 18 23 27 21 12 23 22 17 20
84.05 6.64 7.31 2.00 84.05 6.64 7.31 2.00 78.49 6.63 7.28 7.60 72.77 6.63 7.30 13.30 66.78 6.63 7.29 19.30 69.90 30.10
8.72 1.53 4.96 – 8.81
Mixture 5 min milling
Mixture 15 min milling
Mixture 30 min milling
Mixture 60 min milling
Fe2O3 120 min milling
9.24
14.01
24.82
27.71
mixture ball milled for 60 min, no Li2CO3 powder morphology was observed indicating that the ball milling produced very fine particles of lithium carbonate. One can observed that the agglomerates of milled mixture are characterized by denser structure compared to un-milled one. Moreover, from SEM images with 200× magnification, which are not presented in the article, the milled mixture exhibits larger agglomerates. These results are very important to explain the subsequent results of the reactivity of investigated mixtures.
errors for lattice parameters are ca. 5%. The data of XRD analysis for un-milled mixture (ball milling for 0 min) indicate the presence of α-Fe2O3, Li2CO3, ZnO and γ-Fe2O3 phases (marked reflections) for Li2CO3-Fe2O3-ZnO composition. According to the XRD data for mechanically activated mixtures, the broadening of the peaks due to decreased crystallite sizes is observed as a result of the ball milling. From Table 1, the α-Fe2O3 concentration reduces while the Li2CO3 and ZnO concentrations remain unchanged. It was also revealed that the spinel phase concentration increases with increasing the milling time. According to lattice parameters of γ-Fe2O3 (a = b = c = 0.8347 nm) and Fe3O4 (a = b = c = 0.8400 nm) from Powder Diffraction Standards cards, this phase in the samples milled for 15 min and more presumably belongs to Fe3O4 (PDF No. 36-314), whose formation during ball milling of α-Fe2O3 was also confirmed in [30]. To support this assumption, we have analysed the α-Fe2O3 powder milled for 120 min in a planetary ball mill in the air. Both the α-Fe2O3 and Fe3O4 phases were identified on the XRD pattern (Fig. 3b), and the amount of Fe3O4 phase is ca. 30.1 mass% (Table 1). The results of the measurement of the magnetic properties of this powder showed an increase in the saturation magnetization due to the formation of Fe3O4 magnetic phase. The crystallite size of the phases presented in the milled mixture is shown in Table 1. The reduction in the crystallite size of all reagents is intensive up to 30 min of milling and changes slightly with prolonged milling time.
3.3. Laser diffraction and BET analyses The particle size distributions of the initial reagents, mixtures unmilled and milled for 60 min were measured by laser diffraction and are shown in Fig. 5. Particle size distribution specifications (D10, D50, and D90 values) based on normalized particle amount analysis are summarized in Table 2. The total measurement error of particle size is ± 10%. For the hematite, the particle size is characterized by bimodal distribution with the small particles in the 0.2 − 3 μm range and a significant fraction of large particles in the 3 − 40 μm range. The average particle size of Fe2O3 powder is 12.1 μm. The ZnO reagent with the average particle size of 3.1 μm has a distribution centred at ca. 2 μm with a greater shoulder towards higher particle dimensions. The Li2CO3 reagent has a wide particle size distribution that consists of two significant fractions of particles in the 3 − 50 μm and 50 − 300 μm ranges. The average particle size of Li2CO3 powder is 47.2 μm. It can be seen that the distribution for un-milled mixture has a behavior similar to the distribution of the Fe2O3. The ball milled mixture is characterized by a wide particle size distribution which differs from the un-milled mixture by the presence of significant fractions of larger particles in the 50 − 100 μm range, indicating the agglomeration effect after mechanical milling the Li2CO3-Fe2O3-ZnO mixture. This assumption is confirmed by an increase in the average particle size of milled mixture (see Table 2). Table 1 shows the specific surface area for initial powders (the first row in the last column) and the average specific surface for all particles in milled mixtures (other rows). The error of SSA is +/− 3%. According to the BET data, the average SSA significantly increases in the mixtures milled for more than 15 min, indicating a strong decrease in their average particle size. Thus, the different behavior of the average particle size obtained from laser diffraction and BET analyses indicates that the ball milling of
3.2. SEM analysis Microstructural analysis was carried out for all samples. However, the article presents only SEM images for initial reagents, mixtures unmilled and milled for 60 min (Fig. 4). The SEM micrographs of samples milled for 5, 15, and 30 min showed similar results. The Fe2O3 image shows microstructure with spherical small particles in addition to the larger particles (Fig. 4). The diameter of Fe2O3 particles ranged from 0.2 to 5 μm. Rod-like and flake-like morphologies were observed in Li2CO3 and ZnO powders, respectively. By means of image processing, the average particle sizes estimated from the geometric analysis are found to be approximately 2, 9 and 3 μm for αFe2O3, Li2CO3 and ZnO, respectively. As shown in Fig. 4, the presence of the Li2CO3 particles is observed on the SEM image for α-Fe2O3-Li2CO3-ZnO un-milled mixture. As for 102
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
Fig. 4. SEM micrographs of initial reagents, mixtures un-milled and milled for 60 min.
transition (α → b) in Li0.5Fe2.5O4 phase that is formed as intermediate phase during the synthesis of LiZn ferrite [29,32]. As for milled Li2CO3-Fe2O3-ZnO mixtures, the mass loss process starts at ca. 100 °C. By increasing the milling time, the end temperature of this process shifts towards lower temperatures more intensively with an increase in the milling time to 15 min (Fig. 6), and then varies slightly with further increasing milling time. The mass loss process can be divided into two ranges. The first range is 100 − 350 °C and the second one is 350 − 600 °C. A slight mass loss process was observed on TG curve in the first range for sample milled 5 min, while the main mass loss occurs in the second temperature range. On the contrary, samples milled for prolonged milling time are characterized by a significant mass loss at low temperatures. The main total mass loss values were higher than calculated value for CO2 evaporation. We have previously shown in [29] using mass spectrometry analysis that only carbon dioxide was released during the reaction from unmilled mixture, while the output of both CO2 and H2O were observed during the reaction from milled mixtures. One can notice that the ball milled mixture have an enhanced ability to adsorb gases from the atmosphere like H2O which evaporate during heating at low temperatures, leading to a higher mass loss than the expected value calculated according to reaction. According to calorimetric analysis, the DSC curves show several peaks. There is a complex peak in the low-temperature range associated with the imposition of the endothermal and exothermal reactions.
mixture in a planetary mill decreases the average particle size of initial reagents and simultaneously increases the free energy of the system resulting in formation of agglomerates. 3.4. Thermogravimetric and calorimetric analyses Fig. 6 shows the TG, derivative TG (DTG) and DSC curves for the Li2CO3-Fe2O3-ZnO reagent mixture mechanically activated for various milling times. For the un-milled mixture (0 min milling), the thermal behavior of sample is typical in lithium-zinc ferrite formation during non-isothermal synthesis [31]. The mass loss process starts at a temperature of 500 °C and, as shown by DTG curve, comprises two steps, which correspond to the interaction between reagents with the Li2CO3 decomposition in the temperature range 500 − 650 °C and the melting of lithium carbonate residues at ca. 724 °C. The shape of the DTG curve is similar to the DSC curve, which shows that the mass change in the temperature range 500 − 730 °C can be attributed to the observed endothermic effects (↑exo in Fig. 6a indicates the direction of DSC peaks observed; the peak above the baseline means exothermic effect and below the baseline − endothermic one). The total mass loss was 4.02% and this value is close to the expected value for CO2 releasing calculated from Eq. (1) (3.95%). As was shown in [29] by using mass spectrometry analysis of gases, only CO2 is released in this temperature range. A small DSC peak at ca. 750 °C is related to order − disorder phase 103
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
Furthermore, the area of exothermal DSC peak in the temperature range 200–300 °C, which is most likely associated with the release of H2O, increases with increasing the milling time. Another DSC peak is observed in the temperature range 400 − 600 °C for the 5 min milled mixture and it shifts toward lower temperatures for samples milled for prolonged milling time. Since the area of this DSC peak and mass loss corresponding to this range decrease by increasing the milling time, the significant part of the reaction (Eq. (1)) occurs prior to the onset of the second range. It can be seen from Fig. 7, the end temperature of Li2CO3 decomposition, which was estimated from TG/DSC analysis, greatly decreases with increasing the milling time up to 15 min and then changes slightly at prolonged milling times. Consequently, mechanical activation of the reagent mixture considerably enhances the reactivity of the Li2CO3Fe2O3-ZnO system and shifts the reaction toward lower temperatures. 3.5. Thermomagnetometric analysis Fig. 8 shows the schematic of TG/DTG curves behavior during heating of magnetic sample in thermal analyzer. It is evident that if no change in a mass of sample occur in a certain temperature range (for example, due to chemical reactions), then the TG curve does not change without a magnetic field (Fig. 8, dashed TG curve). The behavior of TG curve under magnetic field, TG(M), demonstrates the presence of mass jump at temperature of ferrimagnet-paramagnet transition (Curie temperature) for the magnetic phase (Fig. 8, solid TG curve). The DTG curve as a derivative TG(M) curve, DTG(M), allows to determine the Curie temperature of this transition and hence to characterize magnetic phase in the sample [33]. First of all, we analysed the un-milled and milled for 60 min Fe2O3 reagent by thermomagnetometry method to examine the formation of new phases after mechanical milling of the main component of mixture (Fig. 9). During the experiment, the magnets were attached and removed on the outer side of the measurement cell at regular intervals, which enables simultaneous analysis of both TG and TG(M) curves. Fig. 9a presents the thermal behavior of initial iron oxide. TG curve indicates the small mass jumps corresponding to low magnetization of α-Fe2O3, which disappear after magnetic phase transition at ca. 678 °C that is Curie temperature of hematite. The thermal analysis result for the iron oxide milled for 60 min is shown in Fig. 9b. TG curve presents a significant increase in the mass of the sample in the temperature range 100–200 °C. It can be assumed that if Fe3O4 forms during the milling of α-Fe2O3, then a change in mass can be observed due to the oxidation of Fe3O4 to Fe2O3 when the sample is heated [34]. According to the increase in mass by 0.96% (Fig. 9b, TG curve) due to oxidation by the reaction 4Fe3O4 + O2 → 6Fe2O3, the calculated value of Fe3O4 content in the total iron oxide powder was 28 mass%, and this value is close to the XRD analysis (Table 1) where about 30.1 mass% of spinel Fe3O4 was estimated. In this temperature range the DSC curve shows an exothermal peak, reflecting the process of oxidation of the sample. It is known that the oxidation of Fe3O4 to α-Fe2O3 proceeds through the formation of intermediate γ-Fe2O3 phase under the scheme: Fe3O4 → γ-Fe2O3 → α-Fe2O3 [35,36]. As well known that both Fe3O4 and γ-Fe2O3 have a high magnetization [37,38], so we can observe high mass jumps on TG curve from room temperature to 460 °C (Fig. 9b, TG curve). After this, TG and DSC curves show γ-Fe2O3→α-Fe2O3 transition at ca. 458 °C, which is accompanied by a change in mass of sample only due to applying a magnetic field. Thus, the above results confirm the formation of the Fe3O4 phase in α-Fe2O3 reagent during the mechanical activation of the Li2CO3-Fe2O3ZnO mixture. Fig. 6b shows the thermomagnetometric analysis results of mixtures, which are obtained at cooling regime where the magnets were attached immediately after the heating mode. No mass changes were observed for samples milled for 0 and 5 min. TG(M) and DTG(M) for the
Fig. 5. Particle size distribution of initial reagents, un-milled and milled for 60 min mixtures. Table 2 Particle size distribution from laser diffraction analysis. Composition
D10 (μm)
D50 (μm)
D90 (μm)
α-Fe2O3 Li2CO3 ZnO Mixture un-milled Mixture milled
0.5 5.7 0.9 0.4 0.9
12.1 47.2 3.1 9.6 12.8
25.6 212.9 14.6 24.6 68.3
104
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
Fig. 6. Thermal analysis of Fe2O3–Li2CO3–ZnO mixture milled for various milling times: (a) – heating regime, (b) – cooling regime.
Fig. 7. End temperature of Li2CO3 decomposition vs. milling time.
Fig. 9. Thermal analysis of Fe2O3 powder un-milled (a) and milled for 120 min (b).
analyzer. Comparing the thermomagnetometric analysis for un-milled and milled mixtures (see Fig. 6b), it can be deduced that the reactivity of the system increases strongly when we use the activation time of 15 min and more. In general, the reactivity of the solid-phase system can increase due to several factors including the reduction in particle size and the formation of particle structure defects during mechanical activation. On the other hand, the γ-Fe2O3 formation in the temperature range 200–400 °C, which is due to the oxidation of Fe3O4 under the scheme Fe3O4 → γ-Fe2O3 → α-Fe2O3, can also increase the reactivity of the
Fig. 8. A schematic of thermal behavior of magnetic sample in and without magnetic field.
remaining samples show mass changes at ca. 493 °C, and according to data in [29], this temperature value correspond to magnetic phase transition at the Curie temperature in Li0.4Fe2.4Zn0.2O4 ferrite phase. One can observe that the height of the mass jump increases with increasing the milling time and this is most likely due to increasing the ferrite phase formation in reaction mixture during its heating in thermal 105
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
and more at 2220 rpm rotation speed for vial.
• It can be assumed that the observed agglomerates consist of well-
interlinked small size particles of reagents, which leads to an increase in contact area between reagents and the acceleration of the chemical reaction.
Acknowledgments This work was supported by The Ministry of Education and Science of the Russian Federation in part of the Science program (project 11.980.2017/4.6). The measurements of X-ray diffraction analysis data were funded from Tomsk Polytechnic University Competitiveness Enhancement Program grant.
Fig. 10. Model of the un-milled and milled particles.
References Li2CO3-Fe2O3-ZnO mixture and shift the reaction towards lower temperatures. Additional experiments showed that the solid-state reaction of the mixture, consisting of powders milled separately, does not show any significant difference as compared to those of the as-milled samples. Only ball milling of the whole mixture leads to reactivity increase. Since, in conventional solid-phase synthesis, contacts of three or more particles of different powders in mixture are unlikely, reactions in such systems proceed through intermediate stages in which pairs of particles in contact take part. As a result of the plastic deformation of solid particles during ball milling of powders in a planetary mill, the point contacts are transformed into contacts along the surface, and thus the formation of triple contacts becomes possible. This leads to a decrease in the number of intermediate stages, acceleration of the synthesis process and homogeneity of the final product. Thus, for milled Li2CO3-Fe2O3-ZnO mixture, it can be assumed that the agglomerates observed from SEM images consist of well-interlinked small size particles of reagents (see Fig. 10), which leads to an increase in contact area between reagents and the acceleration of the chemical reaction.
[1] P.D. Baba, G.M. Argentina, W.E. Courtney, G.F. Dionne, D.H. Temme, Fabrication and properties of microwave lithium ferrite, IEEE Trans. Magn. 8 (1972) 83–94. [2] Parveen Kumar Dipti, J.K. Juneja, Sangeeta Singh, K.K. Raina, Chandra Prakash, Improved dielectric and magnetic properties in modified lithium–ferrites, Ceram. Int. 41 (2015) 3293–3297. [3] N. Rezlescu, C. Doroftei, E. Rezlescu, P.D. Popa, Lithium ferrite for gas sensing applications, J. Sens. Actuators 133 (2008) 420. [4] W. Cook, M. Manley, Raman characterization of α- and β-LiFe5O8 prepared through a solid-state reaction pathway, J. Solid State Chem. 183 (2010) 322–326. [5] P.B. Brown, A superstructure in spinels, Nature 170 (1952) 1123. [6] J. Smit, H.P.J. Wijn, Ferrites Physical Properties of Ferrimagnetic Oxides in Relation to Their Technical Applications, Philips Technical Library, Eindhoven, 1959. [7] C.E. Patton, C.A. Edmondson, Y.H. Liu, Magnetic properties of lithium zinc ferrite, J. Appl. Phys. 53 (1982) 2431–2433. [8] A. Gruskova, V. Jancarik, J. Slama, R. Dosoudil, Effect of Zn-Ti substitution on electromagnetic properties of Li ferrites, J. Magn. Magn. Mater. 304 (2006) 762–765. [9] G.M. Argentina, P.D. Baba, Microwave lithium ferrites: an overview: IEEE trans, Microwave Theory Tech. 22 (1974) 652–658. [10] V. Verma, S.P. Gairola, V. Pandey, J.S. Tawale, H. Su, R.K. Kotanala, High permeability and low power loss of Ti and Zn substitution lithium ferrite in high frequency range, J. Magn. Magn. Mater. 321 (2009) 3808–3812. [11] B. Edelio, Tle. Mercier, M. Quarton, Microstructure and physicochemical studies of pure and zinc-substituted lithium ferrites sinteredabove 1000 °C, J. Am. Ceram. Soc. 78 (1995) 365–368. [12] S.S. Teixeira, M.P.F. Graça, L.C. Costa, Dielectric, morphological and structural properties of lithium ferrite powders prepared by solid state method, J. NonCrystalline Solids 358 (2012) 1924–1929. [13] D.H. Ridgley, H. Lessoff, J.V. Childress, Effects of lithium and oxygen losses on magnetic and crystallographic properties of spinel lithium ferrite, J. Am. Ceram. Soc. 53 (1971) 304–311. [14] M. Tabuchi, K. Ado, H. Sakaebe, C. Masquelier, H. Kageyama, O. Nakamura, Preparation of AFeO2 (A = Li Na) by hydrothermal method, Solid State Ion. 79 (1995) 220. [15] V.K. Sankaranarayanan, Q.A. Pankhurst, D.P.E. Dickson, C.E. Johnson, Synthesis and characterization of ultrafine lithium ferrite from a citrate precursor, J. Magn. Magn. Mater. 130 (1994) 288–292. [16] P. Vijaya Bhasker Reddy, B. Ramesh, Ch. Gopal Reddy, Electrical conductivity and dielectric properties of zinc substituted lithium ferrites prepared by sol-gel method, Physica B 405 (2010) 1852–1856. [17] H. Anwar, A. Maqsood, Microwave magnetic and absorption properties of Li0.5Mnx/ 2Zn0.75-x/2Fe2O4 soft nano ferrites prepared by sol-Gel auto combustion method, Electron. Mater. Lett. 9 (2013) 641–647. [18] S.N. Hosseini, M.H. Enayati, F. Karimzadeh, N.M. Sammes, Formation mechanism, crystallite growth and electrical conductivity of nano-crystalline CuxFe3-xO4 (0.75 ≤ x ≤ 1.25) spinels prepared by glycine-nitrate process, Thermochim. Acta 639 (2016) 91–97. [19] M. Stoia, E. Muntean, C. Păcurariu, C. Mihali, Thermal behavior of MnFe2O4 and MnFe2O4/C nanocomposite, Thermochim. Acta 652 (2017) 1–8. [20] A.K. Giri, Nanocrystalline materials prepared through crystallization by ball milling, Adv. Mater. 9 (1997) 163–166. [21] V.V. Boldyrev, Mechanochemistry and mechanical activation of solids, Russ. Chem. Rev. 75 (2006) 177–189. [22] V. Berbenni, C. Milanese, G. Bruni, A. Marini, I. Pallecchi, Synthesis and magnetic properties of ZnFe2O4 obtained by mechanochemically assisted low-temperature annealing of mixtures of Zn and Fe oxalates, Thermochim. Acta. 447 (2006) 184–189. [23] V. Berbenni, G. Bruni, C. Milanese, A. Girella, A. Marini, Synthesis and characterization of LaFeO3 powders prepared by a mixed mechanical/thermal processing route, J. Therm. Anal. Calorim. (2017) 1–7, http://dx.doi.org/10.1007/s10973017-6878-z First Online: 11 December 2017. [24] V. Berbenni, A. Marini, P. Matteazzi, R. Ricceri, N. Welham, Solid-state formation of lithium ferrites from mechanically activated Li2CO3-Fe2O3 mixtures, J. Eur. Ceram. Soc. 23 (2003) 527–536. [25] H.M. Widatallah, C. Johnson, F.J. Berry, The influence of ball milling and
4. Conclusions As a result of studies of microstructure and reactivity of Li2CO3Fe2O3-ZnO mixture ball milled in a planetary mill, the following conclusions were made.
• It was established from XRD analysis that the composition of reagent
• •
• •
mixture changes during mechanical milling. At increased milling time, the α-Fe2O3 concentration reduces while the Li2CO3 and ZnO concentration remains unchanged. It was also revealed that the Fe3O4 spinel phase is formed during mechanical milling of mixture in a planetary mill and the spinel phase concentration increases with increasing the milling time. The thermomagnetometric analysis of α-Fe2O3, which was subjected to a long milling, confirm the formation of the Fe3O4 phase in it. Consequently, this phase is formed in the Li2CO3-Fe2O3-ZnO mixture during mechanical activation. The crystallite sizes of initial reagents were found to decrease with increasing the milling time. Moreover, BET analysis indicates a significant increase in specific surface area of mechanically activated mixtures resulting in a high ability to adsorb water from atmosphere, which evaporates during sample heating at low temperatures. From SEM and laser diffraction analyses, an increase in size of agglomerates in the mechanically activated mixture is observed. Thus, a ball milling of mixture in a planetary mill decreases the average particle size of reagents and simultaneously increases the free energy of the system resulting in formation of agglomerates. It was found that the reactivity of the mechanically activated mixture increases strongly when we use the activation time of 15 min 106
Thermochimica Acta 664 (2018) 100–107
E. Lysenko et al.
[26]
[27]
[28]
[29]
[30]
[31]
subcequent calcination on the formation of LiFeO2, J. Mater. Sci. 37 (2002) 4621–4625. A.P. Surzhikov, E.N. Lysenko, A.V. Malyshev, A.M. Pritulov, O.G. Kazakovskaya, Influence of mechanical activation of initial reagents on synthesis of lithium ferrite, Russ. Phys. J. 6 (2012) 627–677. M. Kavanlooee, B. Hashemi, H. Maleki-Ghaleh, J. Kavanlooee, Effect of annealing on phase evolution, microstructure, and magnetic properties of nanocrystalline ballmilled LiZnTi ferrite, J. Electron. Mater. 41 (2012) 3082–3086. H.M. Widatallah, X.L. Ren, I.A. Al-Omari, The influence of TiO2 polymorph, mechanical milling and subsequent sintering on the formation of Ti-substituted spinelrelated Li0.5Fe2.5O4, J. Mater. Sci. 41 (2006) 6333–6338. E.N. Lysenko, A.P. Surzhikov, V.A. Vlasov, A.V. Malyshev, E.V. Nikolaev, Thermal analysis study of solid-phase synthesis of zinc- and titanium- substituted lithium ferrites from mechanically activated reagents, Therm. Anal. Calorim. 122 (2015) 1347–1353. M. Zdujić, C. Jovalekić, Lj. Karanović, M. Mitrić, The ball milling induced transformation of α-Fe2O3 powder in air and oxygen atmosphere, Mater. Sci. Eng. A 262 (1999) 204–213. A.P. Surzhikov, A.M. Pritulov, E.N. Lysenko, A.N. Sokolovskii, V.A. Vlasov, E.A. Vasendina, Dependence of lithium–zinc ferrospinel phase composition on the
[32]
[33] [34]
[35] [36] [37]
[38]
107
duration of synthesis in an accelerated electron beam, J. Therm. Anal. Calorim. 110 (2012) 733–738. A.P. Surzhikov, T.S. Frangulyan, S.A. Ghyngazov, E.N. Lysenko, Investigation of structural states and oxidation processes in Li0.5Fe2.5O4-δ using TG analysis, J. Therm. Anal. Calorim. 108 (2012) 1207–1212. P.K. Gallagher, Thermomagnetometry, J. Therm. Anal. Calorim. 49 (1997) 33–44. E.N. Lysenko, A.P. Surzhikov, S.P. Zhuravkov, V.A. Vlasov, A.V. Pustovalov, N.A. Yavorovskii, The oxidation kinetics study of ultrafine iron powders by thermogravimetric analysis, J. Therm. Anal. Calorim. 115 (2014) 1447–1452. J.P. Sanders, P.K. Gallagher, Kinetics of the oxidation of magnetite using simultaneous TG/DSC, J. Therm. Anal. Calorim. 72 (2003) 777–789. J.P. Sanders, P.K. Gallagher, Thermomagnetometric evidence of γ-Fe2O3 as an intermediate in the oxidation of magnetite, Thermochim. Acta 406 (2003) 241–243. T. Osaka, T. Matsunaga, T. Nakanishi, A. Arakaki, D. Niwa, H. Iida, Synthesis of magnetic nanoparticles and their application to bioassays, Anal. Bioanal. Chem. 384 (2006) 593–600. A.C. Grigorie, C. Muntean, M. Stefanescu, Obtaining of γ-Fe2O3 nanoparticles by thermal decomposition of polyethyleneglycol–iron nitrate mixtures, Thermochim. Acta 621 (2015) 61–67.