COMBUSTION SYNTHESIS OF FERRITE MAGNETS

 

Quentin A. PANKHURST,† Louise AFFLECK,†‡ Marco D. AGUAS,‡ Maxim V. KUZNETSOV,† Ivan P. PARKIN,‡ Luis FERNÁNDEZ BARQUÍN, Marius I. BOAMFA,* and Jos A. A. PERENBOOM.*

† Dept. of Physics and Astronomy, University College London, Gower Street, London WC1E 6BT, UK; ‡ Dept. of Chemistry, University College London, 20 Gordon Street, London WC1H 0AJ, UK; ¶ Dept. CITIMAC, Universidad de Cantabria, Santander 39005, Spain; * Research Institute for Materials and High Field Magnet Laboratory, University of Nijmegen, 6525 ED, The Netherlands.

Recent studies of self-propagating high-temperature synthesis (SHS) routes to ferrite magnets, involving the application of a steady state magnetic field during combustion, are reviewed and discussed. Particular attention is paid to the use of large fields (up to 15 T) to accentuate the field effects, and time-resolved X-ray diffraction (at timescales down to 100 ms) to monitor short-lived intermediate phase formation during the passage of the SHS wave.

  

  1. INTRODUCTION

    2. Self-propagating high-temperature synthesis (SHS) routes to the formation of ferrite magnets involve highly exothermic solid state combustion processes. SHS reactions are self-energetic and require no heating from an external source. The best known example is the thermite reaction which is used to join railway tracks (Fe2O3 + Al ® Al2O3 + Fe). Commonly, SHS reactions use the elements or element oxides as the starting materials, with M + E transforming to ME, where M is a metal or metal oxide and E is a non metal. The reactions can be started by point source initiation, where a hot wire or filament is brought into contact with the green mixture, igniting the powder. A synthesis wave or thermal flash moves through the solid, promoting the reaction of successive layers of powder. The synthesis wave is extremely hot (often more than 1200 °C), is highly directional (away from the source of ignition), and has a uniform velocity. This is illustrated in Figure 1, which shows the heat transfer and thermal characteristics of a typical SHS reaction for the formation of a magnetically soft ferrite, Mg0.5Zn0.5Fe2O4. The figure illustrates both the high temperatures attained during synthesis, and the very rapid heating and cooling rates that are involved.

    Time from ignition

    Minimum temperature (red)

    Maximum temperature (black)

     

    4 s

    500 K

    1485 K

    10 s

    500 K

    1560 K

    19 s

    500 K

    1045 K

    27 s

    500 K

    865 K

    FIGURE 1

    Infra-red camera images of a self-propagating high-temperature synthesis (SHS) reaction of a mixture of MgO, ZnO, Fe, Fe2O3 and NaClO4 to form a mixture of Mg0.5Zn0.5Fe2O4 and NaCl, in zero field. The images are artificially colour enhanced, ranging from black for the hottest regions on each image, through blue, green and yellow, to red for the coolest regions. Approximately 3 g of reactants were used, placed in a 65 mm by 10 mm rectangular area on a heat resistant ceramic tile.

     

    An important feature of these reactions, and one which has only recently been recognised and quantified, is that the passage of the wave induces an electrical pulse and a small magnetic field, both of which are thought to be caused by the movement of ions and electrons at the molten reactant front. The latter ‘chemomagnetic fields’ have been observed in reactions involving both non-magnetic reactants and products, for example BaO2 + 0.4 Ti + 0.6 TiO2 ® BaTiO2, (maximum field 6.5 nT measured by a SQUID magnetometer 10 mm from the reaction front) as well as for reactions involving magnetic products, such as Fe2O3 + 2 Al ® 2 Fe + Al2O3 (maximum field 4.6 nT). In light of such observations, it is now timely to consider possible charge transport mechanisms that may play a role in influencing both the nature of the combustion process, and of the resultant products, for SHS reactions.

    We report here on recent studies of both hard and soft ferrite magnets undertaken at University College London. These build on a series of experiments conducted over the last five years in which the application of a magnetic field during the SHS reaction has been found to significantly modify the microstructure, magnetic figures of merit and lattice constants of the product. This work began in 1996 with the first synthesis and characterisation of the M-type ferrites (Ba,Sr)Fe12O19 by SHS in air, and continued in 1997 with the first preparation, by SHS, of the soft ferrite, LiFe5O8. Investigations of the effect of applied fields on these reactions also started in 1997. It was found that for both the hard and soft ferrites, SHS reactions in a field proceeded faster and had higher combustion temperatures than the corresponding zero field reactions. Changes were found in the bulk magnetic properties of the ground and sintered end-products, compared to those made either by zero field SHS or ceramic methods. For a field of 1.1 T the largest changes seen were a 20% decrease in coercivity in BaFe12O19, and magnetisation increases of 15% in MgFe2O4 and 35% in Mg0.5Zn0.5Fe2O4. Incorporating other elements into the samples modified these results, with the coercivity change in BaFe12O19 rising to almost 100% for BaFe10Cr2O19. Changes in unit cell volumes in the sintered end-products were also noted, with a 0.14% reduction in MgFe2O4, a 0.75% reduction in LiFe5O8, and a 0.22% expansion in BaFe12O19, although these results were not based on the most rigorous Rietveld analyses.

    In this paper we highlight new experiments involving: (1) the application of large steady state magnetic fields (up to 15 T), and the subsequent characterisation of the products obtained as a function of field strength, including Rietveld refinements; and (2) time-resolved X-ray diffraction (at intervals down to 0.10 sec) in which the SHS reactions themselves were observed in both zero field and in a field of 1.1 T, and evidence was found for intermediate phase formation in the presence of an applied field. We comment on the possible relationships between these macroscopic effects and the influence of the applied field on the charge transport pathways at the combustion wave front. We also discuss how these effects might be used to controllably influence product microstructure and other technologically important properties, such as the magnetic figures of merit, in ferrite magnets.

     

  2. SHS IN LARGE MAGNETIC FIELDS

    3. We have recently used the Bitter magnet at the Nijmegen High Field Magnet Laboratory to conduct SHS reactions on mixtures of BaO2 (or SrO2), Fe and Fe2O3 in steady state magnetic fields of up to 15 T. A brief communication on this work has appeared elsewhere. A significant new result was that we observed two macroscopically distinct parts in the post-SHS product, one of which had a shiny, metallic appearance, the other having a dull, matt appearance. The formation of the metallic part was promoted by the application of higher fields, as shown in Figure 2.

    3 T

    6 T

    15 T

    20 T

    FIGURE 2

    Photographs of the products of SHS reactions of SrO2, Fe and Fe2O3 in large steady state applied fields. In each case the pellets were 17 mm in diameter.

     

    The shiny and matt parts of the post-SHS products were found to have different phase compositions, microstructures and magnetic properties. This observation may be illustrated through consideration of 57Fe Mössbauer spectra, such as those shown in Figure 3 for the products of the reaction of BaO2, Fe and Fe2O3. It is clear on inspection that the spectra of the matt aliquots are dominated by a central doublet signal, while the spectra of the shiny aliquots have a characteristic sharp right-hand-most absorption line due to one of the component sextet subspectra. Detailed least-squares analyses of these spectra allow the iron-containing different phases to be identified. Table 1 summarises the results of such fits.

    FIGURE 3

    Room temperature 57Fe Mössbauer spectra of the post-SHS products of the reaction of BaO2, Fe, and a -Fe2O3. Sextets due to unreacted Fe and a -Fe2O3, and a doublet from partially reacted Fe1-xO, are evident in the ‘matt’ aliquots. A pair of sextets due to magnetite Fe3O4 accounts for the remainder of the signal in the matt aliquots, and dominates the spectra of the ‘shiny’ aliquots.

    TABLE 1

    Relative site populations determined by least squares fitting the room temperature 57Fe Mössbauer spectra of Figure 3 using combinations of Lorentzian sextets and doublets.

    Applied field

    a -Fe

    Hematite, a-Fe2O3

    Wüstite, Fe1-xO

    Magnetite, Fe3O4

    Barium monoferrite, BaFe2O4

    Matt parts of post-SHS products

     

     

     

    0 T

    1.3 (2)

    -

    25.6 (2)

    75.1 (24)

    -

    5 T

    5.5 (6)

    5.1 (9)

    35.4 (4)

    45.4 (33)

    8.6 (16)

    10 T

    12.5 (5)

    14.9 (9)

    27.3 (3)

    45.3 (29)

    -

    15 T

    13.9 (1)

    7.2 (2)

    46.8 (9)

    32.1 (22)

    -

    Shiny parts of post-SHS products

     

     

     

    0 T

    -

    -

    1.4 (5)

    78.9 (41)

    19.7 (2)

    5 T

    -

    -

    5.6 (3)

    71.3 (34)

    23.1 (2)

    10 T

    4.2 (2)

    -

    26.0 (2)

    61.1 (12)

    8.7 (6)

    15 T

    -

    -

    11.9 (2)

    76.4 (15)

    11.7 (7)

     

    The various phases identified all have distinctive Mössbauer spectra. Any unreacted starting material is evident in the form of sextets: a -Fe has a hyperfine field Bhf of 33.0 T and zero values for both its isomer shift d and its quadrupole shift 2e ; hematite a -Fe2O3 has Bhf ≈ 51.8 T, d ≈ 0.37 mms-1 and 2e ≈ -0.20 mms-1. Among the reaction products, wüstite Fe1-xO is identified as a doublet with d  ≈ 0.85 mms-1 and a quadrupole splitting D ≈ 0.70 mms-1, while magnetite Fe3O4 comprises two sextets, both with near-zero quadrupole shifts, and d ≈ 0.26 and 0.67 mms-1 and Bhf ≈ 49.2 and 46.1 T respectively. A fifth phase, clearly seen in some of the spectra, has d ≈ 0.37 mms-1, 2e ≈ 0.28 mms-1 and Bhf ≈ 42.0 T. On the basis of independent X-ray data we ascribe this phase to barium monoferrite, BaFe2O4, although it should be noted that the Mössbauer parameters are not typical of those seen previously for this material. It may be that this is a disordered or perhaps oxygen deficient variant on the BaFe2O4 structure.

    On inspection of the data in Table 1 it is clear that in general the shiny parts of the product are more fully combusted than the matt parts, with the latter having a greater tendency to contain a portion of unreacted starting mixture. The proportion of wüstite in the product is higher for the matt parts than for the shiny parts, while the shiny parts have more barium monoferrite than do the matt parts, suggesting that part of the combustion process comprises a transformation between these phases.

    Subsequently, the post-SHS samples were all ground and sintered at 1200 ° C for 2 hours, resulting in the formation of monophase BaFe12O19 products. These hexaferrites were found to have significantly different magnetic properties, as well as subtly different structural properties, as detailed in Table 2. Detailed Rietveld analyses were performed to obtain the measured values for the changes in lattice parameter, the largest being an 0.5% reduction in cell volume for the metallic part of the 15 T reaction.

    TABLE 2

    Room temperature magnetic properties (coercivity Hc and maximum magnetisation s max) measured in fields up to 5 kOe, and lattice parameters a and c, for monophase barium hexaferrite BaFe12O19 obtained via an SHS production route.

    Applied field

    Hc (kOe)

    s max (emug-1)

    a (Å)

    c (Å)

    Matt parts of post-SHS products, subsequently sintered at 1200 ° C for 2 hours

    0 T

    2.42 (1)

    50.0 (1)

    5.8886 (3)

    23.2111 (8)

    5 T

    1.47 (1)

    48.9 (1)

    5.8882 (3)

    23.2131 (7)

    10 T

    2.32 (1)

    50.6 (1)

    5.8881 (4)

    23.2125 (9)

    15 T

    2.43 (1)

    46.3 (1)

    5.8865 (5)

    23.1992 (12)

    Shiny parts of post-SHS products, subsequently sintered at 1200 ° C for 2 hours

    0 T

    1.43 (1)

    47.3 (1)

    5.8873 (4)

    23.2026 (9)

    5 T

    1.27 (1)

    53.1 (1)

    5.8876 (5)

    23.2049 (11)

    10 T

    1.52 (1)

    50.6 (1)

    5.8869 (5)

    23.2069 (11)

    15 T

    1.19 (1)

    50.9 (1)

    5.8769 (6)

    23.1583 (16)

     

    The magnetic data in Table 2 show that the coercivities of the BaFe12O19 that originated from the shiny parts of the post-SHS product are consistently less than those derived from the matt parts of the post-SHS product, while the corresponding magnetisations are roughly similar. This is likely to be an indication of a microstructural effect, with the better combusted shiny phase material giving rise to a sintered product with fewer pinning centres, either at grain boundaries or sites of impurities. The variations in lattice parameter, although small, vary smoothly with increasing applied field. The samples derived from the shiny parts of the post-SHS product all have smaller a and c parameters than their matt product counterparts, with the largest differences appearing in the 15 T applied field samples.

    These high field experiments present us with some intriguing implications with regard to applied field SHS routes to ferrite magnets. One is that product microstructure and coercivity may be controlled by the application of a field during combustion. This could be of interest as an alternative means of controlling coercivity compared to the current practice of adding dopant atoms such as cobalt and titanium to the lattice to disrupt the internal exchange interactions in the hexaferrite lattice. The observation of a field dependent lattice contraction effect is also significant, since it implies that changes are being evidenced on an atomic nearest neighbour level, rather than simply microstructural or macroscopic. Another aspect of this is the concept that the applied field may be influencing the combustion process itself, at a local level. Such considerations are the basis of the work described in the next section, where indeed it is found that the combustion processes are different in zero field and in an applied field.

     

  3. TIME-RESOLVED X-RAY DIFFRACTION STUDIES

    4. High quality, relevant data is needed to enable us to characterise and understand the effect of the applied field on the combustion process. Research in the last five years in France and the US has shown that an excellent method for such a purpose is time-resolved X-ray diffraction (TRXRD) using synchrotron radiation. Experiments at LURE in Orsay on the formation of Al-Ni-Ti intermetallic compounds, and at Brookhaven on the reaction Ta + C ® TaC, have shown kinetic features and intermediate phases that are only revealed given the 50-100 ms timing resolution possible with a synchrotron source. However, these TRXRD experiments were performed on bulk compacted samples in large evacuated or helium gas filled chambers, in zero field.

    In 1999 we performed the first TRXRD experiments on SHS reactions in the presence of an applied field – in this case a 1.1 T field from a Nd-Fe-B Halbach cylinder permanent magnet. These experiments were undertaken on Station 16.4 at Daresbury Laboratory, UK. We monitored structure and phase changes in powder mixtures before, during and after the passage of an SHS combustion wavefront. We looked at both hard and soft ferrites, in zero field and in an applied field of 1.1 T. Very rapid changes were observed, and at the fastest time resolution attempted, spectra were recorded consecutively for 100 ms each. In the applied field reactions, there were clear signatures of intermediate phases which formed as the combustion wave passed, then within 0.2-0.5 seconds were transformed to the final product. For example, magnetite Fe3O4 was seen in 4 out of 5 cases to be a dominant transient intermediary in the applied field SHS reaction of BaO2, Fe and Fe2O3, while no such intermediate phase was observed in any of 8 different scans of the zero field reaction. Figure 4 shows one of the applied field SHS runs.

     

     

    FIGURE 4

    Time resolved X-ray powder diffraction patterns of the SHS reaction of BaO2, Fe and Fe2O3 under an oxygen atmosphere and in a magnetic field of 1.1T. Scans were each recorded for 0.25 s; scan numbers are as noted on the figure. An energy dispersive detector system was used; in this case the detector was mounted at 4.67° to the direction of the incident X-ray beam.

    The TRXRD scans in Figure 4 show that at the onset of the SHS wave the diffraction pattern immediately becomes diffuse (scan 4). Shortly thereafter a strong peak due to the (111) reflection for magnetite dominates the pattern (scans 7-10). At longer times, after the passage of the combustion wavefront, the product cools and the X-ray scans take on the appearance of those that are normally observed for the post-SHS product (scan 20 and beyond). In zero field, the observed TRXRD patterns are similar, but without the presence of any intermediate phase.

     

  4. DISCUSSION

The experiments described in this paper show that applying a magnetic field during an SHS reaction can significantly influence the progress of the reaction, and the nature of the products obtained. Multiphase products, such as the shiny and matt parts seen for the reactions of (Ba,Sr)O2, Fe and Fe2O3, are possible. There is an atomic scale effect, as evidenced by changes in lattice parameter in the post-sintering samples. There is also an influence on the combustion wavefront itself, with the field promoting the formation of short-lived intermediate phases.

It is now known that the flow of ions and electrons during the passage of the SHS synthesis wave can constitute an electric current. This leads naturally to the speculation that this flow of charges could be influenced by an applied magnetic field. Recent studies have linked similar ideas to observations of unusual mass transport and fractal morphology effects in the electrochemical deposition of copper and the changed morphologies and reduced yields of carbon nanotubes and fullerenes formed by arc discharge. Although the solid state reactions embodied in the SHS combustion systems are clearly quite different from these liquid and gas phase results, it is reasonable to conceive that, given the very high temperature at the combustion wavefront itself, there could be substantial mobility imparted to the reactant ions and electrons, albeit fleetingly.

We conclude that recent studies of applied field SHS reactions have revealed a wealth of intriguing phenomena, and that much work lies ahead before we will be able to fully comprehend and model these phenomena. Nevertheless, progress in this area is needed both to add to the wider debate about the effect of electromagnetic fields on chemical reactions, as well as to underpin the potential that exists for future exploitation of applied field combustion routes to ferrite magnet synthesis.

 

ACKNOWLEDGEMENTS

We are grateful to Simon Clark, Dave Taylor and Lauchlan Cranswick of Daresbury Laboratory for their help with the TRXRD experiments. We also thank EPSRC (for X-ray beam time and a PhD studentship), the EU (for enabling access to the high field laboratory), and the British Council and Spanish CICYT (for an Acciones Integradas grant), for their support. The Mössbauer data were obtained courtesy of the ULIRS scheme. FLIR Systems International Ltd are thanked for the loan of a ThermaCAM SC2000 infra-red camera.

 

REFERENCES

1) H.C. YI and J.J. MOORE, J. Mat. Sci. 25(1990)1159.

2) Y.G. MOROZOV, M.V. KUZNETSOV, M.D. NERSESYAN, et al., Doklady Akademii Nauk SSSR 351(1997)780.

3) M.D. NERSESYAN, J.R. CLAYCOMB, Q. MING, et al., Appl. Phys. Lett. 75(1999)1170.

4) I.P. PARKIN, G. ELWIN, L.FERNANDEZ BARQUIN, et al., Adv. Mater. 9(1997)643.

5) M.V. KUZNETSOV, Q.A. PANKHURST, and I.P. PARKIN, J. Phys. D 31(1998)2886.

6) I.P. PARKIN, M.V. KUZNETSOV, and Q.A. PANKHURST, J. Mat. Chem. 9(1999)273.

7) M.V. KUZNETSOV, Q.A. PANKHURST, and I.P. PARKIN, J. Mat. Chem. 8(1998)2701.

8) M.D. AGUAS, L. AFFLECK, I.P. PARKIN, et al., J. Mat. Chem. Commun. 10(2000)235.

9) W.B. CROSS, L. AFFLECK, M.V. KUZNETSOV, et al., J. Mat. Chem. 9(1999)2545.

10) J.F. JAVEL, M. DIRAND, F.Z. NASSIK, et al., J. de Physique IV 6(1996)C2-229.

11) E.M. LARSON, J. WONG, J.B. HOLT, et al., J. Mater. Res. 8(1993)1533.

12) G. HINDS, J.M.D. COEY, and M.E.G. LYONS, J. Appl. Phys. 83(1998)6447.

13) J.M.D. COEY, G. HINDS, and M.E.G. LYONS, Europhysics Letters 47(1999)267.

14) H. YOKOMICHI, H. SAKIMA, M. ICHIHARA, et al., Appl. Phys. Lett. 74(1999)1827.