- Email: [email protected]
Novel nanocrystalline mixed conductors based on LiFeBO3 glass
ARTICLE IN PRESS SOSI-14141; No of Pages 5 Solid State Ionics xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Solid State Ionics journal homepage: www.elsevier.com/locate/ssi
Novel nanocrystalline mixed conductors based on LiFeBO3 glass ´ Przemysław P. Michalski* , Tomasz K. Pietrzak, Jan L. Nowinski, Marek Wasiucionek, Jerzy E. Garbarczyk Faculty of Physics, Warsaw University of Technology, Koszykowa 75, Warsaw 00-662, Poland
A R T I C L E
I N F O
Article history: Received 29 July 2016 Received in revised form 30 November 2016 Accepted 6 December 2016 Available online xxxx Keywords: Lithium–borate glass Thermal nanocrystallization Mixed conductor Li-ion batteries Cathodes
A B S T R A C T Novel nanocrystalline mixed conductors were obtained by thermal nanocrystallisation of lithium– iron–borate glasses (nominally LiFeBO3 ) prepared via melt-quenching method. After heat treatment, a glass–ceramics consisting of FeBO3 and LiFeBO3 phases was obtained. Grain sizes, estimated using Scherrer formula, were 45–50 nm and 70–80 nm, respectively. Value of electrical conductivity of the initial glass was estimated as 5.8·10−12 S cm−1 . After annealing at 475 ◦ C, a growth by a factor 2.4·106 was observed, leading to final conductivity value 1.4·10–5 S cm−1 at room temperature. The activation energy of electronic conductivity was lowered from 0.81 eV for glass to 0.18 eV for thermally treated material. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Due to their high gravimetric energy and power, lithium ion (Li-ion) batteries are preferred energy sources for mobile devices and electric cars. Cathode performance and its production costs are the biggest bottlenecks in Li-ion technology development. Therefore, there is a high need for novel and inexpensive cathode materials. Lithium iron borate (LiFeBO3 ) is one of the potential candidates. Due to low molar mass of the borate group, its gravimetric capacity reaches 220 mAhg−1 with one lithium ion exchanged. Also, thanks to high abundance of the substrate elements in Earth’s crust, the material can be produced at low cost; it is also environmentally benign. On the other hand, LiFeBO3 exhibits low electronic conductivity – when prepared in polycrystalline form, the value is 3.9 • 10–7 S cm −1 [1], which is too low to obtain high current rates in Li-ion battery. It is also known that electrochemical performance of LiFeBO3 is strongly reduced for grain sizes bigger than 100 nm [1]. Here, we would like to propose a novel method of materials’ preparation, based on thermal nanocrystallisation of initially amorphous precursors. This method allows to obtain materials with highly enhanced conductivity and nanometric-sized grains. Recently, we have successfully applied this approach in case of cathode-like V2 O5 –P2 O5 [2], LiF–V2 O3 –P2 O5 [3] and Li2 O–FeO– V2 O5 –P2 O5 [4] systems. In the latter case, the material with final
* Corresponding author. E-mail address: [email protected] (P. Michalski).
conductivity value of 10 −3 S cm−1 and microstructure consisting of 3–5 nm grains of LiFePO4 olivine was obtained. Inspired by previous results, we are determined to study other glass systems. A proposed borate system seems to be an ideal candidate for investigations because of a few reasons. First of all, boron (III) oxide is a good glass-former 5]. Also, boron-related materials are good Li+ conductors, which allows their use as electrolytes (e.g. LIBOB salt [6]), glassy coatings on cathode materials providing better electrochemical efficiency [7] or as glassy cathodes exhibiting high performance itself [8]. 2. Experimental Glasses of nominal composition 0.25Li2 O • 0.5FeO • 0.25 B2 O3 (LiFeBO3 ) were prepared. Starting reagents: Li2 CO3 (Aldrich, 99.9%), FeC2 O4 • 2H2 O (Aldrich, 99.9%) and H3 BO3 (Aldrich, 99.5%) were mixed and homogenised in Retsch planetary mill (400 rpm, 20 min). The obtained batches were melted for 10 min in furnace preheated to temperatures in 1000–1250 ◦ C range. Double crucible method [9] was used to prevent iron oxidation. The melted batches were rapidly poured on a metal plate and quickly pressed with another metal plate (melt-quenching method. X-ray diffractometry (Phillips X’Pert Pro with Bragg–Brentano configuration, using CuKa line with k = 1.5406 Å) was used to verify the amorphousness of obtained samples and for temperaturedependent investigations. In the latter case, Anton Paar HTK1200 oven was used and the measurements were conducted in nitrogen flow. The temperature was stabilised for 40 min before each measurement.
http://dx.doi.org/10.1016/j.ssi.2016.12.002 0167-2738/© 2016 Elsevier B.V. All rights reserved.
Please cite this article as: P. Michalski et al., Novel nanocrystalline mixed conductors based on LiFeBO3 glass, Solid State Ionics (2016), http://dx.doi.org/10.1016/j.ssi.2016.12.002
ARTICLE IN PRESS 2
P. Michalski et al. / Solid State Ionics xxx (2016) xxx–xxx
Temperatures of glass transition and crystallisation were obtained using differential thermal analysis method. TA Instruments SDT Q600 instrument was used. Measurements were performed with 1C min−1 and 10C min−1 heating rates, in argon flow. Electrical properties were studied in situ during temperature ramps, using Solartron 1260 Gain Phase/Impedance Analyzer integrated with a furnace and a temperature programming module (Eurotherm 2404). Before each measurement, the temperature was stabilised for 40 min. Maximum temperatures were chosen in 400–500 ◦ C range. Impedance spectra were collected in the 10 mHz– 10 MHz frequency range for each temperature in the ramp. The applied voltage amplitude was equal to 0.1 V.
3. Results and discussion 3.1. XRD of as-prepared samples For all melting temperatures in the investigated range, the materials obtained after quenching were amorphous with small traces of non-stoichiometric Fe3 O4 phase (Fig. 1). The best match was obtained for Fe2.897 O4 phase (PDF-2 reference code: 01-086-1337, later referred as Fe3 O4 ). The amount of this phase decreased with increasing melting temperature. This was the reason why the material obtained by melting at the highest temperature (1250 ◦ C) was chosen for further studies. The existence of non-stoichiometric Fe3 O4 phase is also an indication of presence of Fe2+ and Fe3+ ions in glass, which are relevant for electronic hopping conductivity.
3.2. DTA of as-prepared samples DTA curves obtained for 1 ◦ C min−1 (blue) and 10 ◦ C min−1 (red) are typical for glassy materials (Fig. 2) and consist of glass transition step which is interpreted as a change of heat capacity during devitrification, followed by a few exothermic peaks which may be attributed to crystallisation of different phases. The characteristic temperatures were collected in Table 1. For lower rates, the transitions were less pronounced and the temperatures could be only estimated. This fact can be explained by the nature of DTA experiment, where the heat difference between analyzed and reference sample is measured. This difference can be written as follows: dQ dQ dT dT = = cp dt dT dt dt
Fig. 2. DTA curves collected with heating rates 10 ◦ C min−1 (red) and 1 ◦ C min−1 (blue). Characteristic thermal events are marked by arrows. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
where cp is the heat capacity of the material. Assuming cp constant in current temperature, one can see that the heat difference dQ/dt is proportional to heating rate dT/dt. 3.3. Temperature XRD investigations The results of XRD temperature-dependent measurements are presented in Fig. 3. Due to existence of small Fe3 O4 inclusions in initial glass, this phase is the first to crystallise at temperatures in the 425–450 ◦ C range. These are in proximity of glass transition and first crystallisation peak temperatures for low heating rate. As soon as the glass devitrifies (Tg ), atoms movement is facilitated and they may form crystal structure based on initially present iron oxide units. The other phases – delithiated, hexagonal iron borate phase and monoclinic lithium iron borate phase – crystallise at ca. 525–550 ◦ C. This temperature may be ascribed to second crystallisation peak (Tc2 ), while the third one could be correlated with recrystallisation of lithium boron oxide (Li2 B4 O7 , reference code: 00-011-0408). The peak matched by this pattern firstly occurs at 525 ◦ C at 2H ≈ 22◦ . Up to 600 ◦ C, its intensity increases and then, for higher temperatures, decreases, to almost vanish at 650 ◦ C. Drawing conclusions based on DTA data collected for 1 ◦ C min−1 is reasonable, because similar heating rate occurred during stabilisation of temperature in XRD measurements. The FeBO3 /LiFeBO3 patterns were obtained from Crystallography Open Database [10] (COD IDs 1511131 and 1511129, respectively). After cooling down the sample to room temperature, the grain sizes were estimated using Scherrer formula, taking into calculations the most intense peaks of FeBO3 (the peaks of lithiated phase were hardly visible). The obtained result is 20–45 nm. The intensities of peaks of FeBO3 /LiFeBO3 phases in XRD data acquired during temperature-dependent measurements were relatively low, indicating slight content of these phases in obtained glass-ceramics. Therefore, in the next step, the as-prepared glass was isothermally annealed at 550 ◦ C for 12 h (purple line in Fig. 4). The temperature was chosen as the one, at which FeBO3 /LiFeBO3 peaks firstly occurred. During temperature-dependent measurements (green line), the sample was kept in one temperature for ca. Table 1 Temperatures of characteristic thermal events.
Fig. 1. X-ray diffractograms of glasses prepared with different temperatures of melting (shown on the right).
Heating rate
Tg [◦ C]
Tc1 [◦ C]
Tc2 [◦ C]
Tc3 [◦ C]
1 ◦ C min−1 10 ◦ C min−1
421 435
454 476
502 546
588 639
Please cite this article as: P. Michalski et al., Novel nanocrystalline mixed conductors based on LiFeBO3 glass, Solid State Ionics (2016), http://dx.doi.org/10.1016/j.ssi.2016.12.002
ARTICLE IN PRESS P. Michalski et al. / Solid State Ionics xxx (2016) xxx–xxx
3
One can roughly estimate the FeBO3 /LiFeBO3 phase ratio, by comparing the area of the most intense peaks of these phases (2H = 25.18◦ and 35.37◦ , respectively). The calculated ratio is 75/25. Similarly to the previous sample, grain sizes were estimated using Scherrer formula. The values of 45–50 nm and 70–80 nm were obtained for FeBO3 and LiFeBO3 , respectively. 3.4. Electrical investigations The room temperature conductivity value of the initial glass was too low to be measured. After extrapolation, it was calculated as 5.8 • 10−12 Scm−1 . The activation energy was equal to 0.81 eV (Fig. 5). Heat treatment at different temperatures between Tg and Tc2 led to irreversible conductivity growth. The best results were obtained for 475 ◦ C — after annealing, the conductivity value at room temperature increased by a factor 2.4 • 106 leading to the final value 1.4 • 10−5 S cm−1 , while the activation energy was lowered to 0.18 eV. One can also see that cooling ramps for annealing temperatures 425 ◦ C and 450 ◦ C consist of two parts with different slopes. The higher slope is present in higher temperatures. This phenomenon may be explained as follows: the activation energy value for ion is higher than for electrons. Due to that, at low temperatures, the
Fig. 3. XRD patterns collected during temperature-dependent measurements. FeBO3 peaks positions are marked by blue circles and LiFeBO3 – by red diamonds. Peak ascribed to Li2 B4 O7 is marked by asterisk. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
1 h (stabilisation and measurement time). It can be seen that longer annealing time led to further growth of FeBO3 and LiFeBO3 phases, as peaks related to these phases became more pronounced. What is more, the most intense peak matched by Fe3 O4 pattern present at 2H ≈ 35◦ in data collected during temperature ramp has split into two, matched by LiFeBO3 pattern. It means that present iron oxide phases possibly recrystallise to desirable lithium iron borate phases.
◦
Fig. 4. Comparison of XRD patterns collected at 550 C: during isothermal annealing for 12 h (purple); during temperature ramp (green). Dashed lines indicate peaks positions of FeBO3 (blue) and LiFeBO3 (red) peaks. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 5. Temperature dependence of electric conductivity presented in Arrhenius coordinates. Black points represent the data obtained during heating ramp, points in different colours – data collected during cooling ramps after heating to various temperatures. DTA curve for 1 ◦ C min−1 and the conductivity value for crystalline LiFeBO3 [1] (brown line) are shown for comparison. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Please cite this article as: P. Michalski et al., Novel nanocrystalline mixed conductors based on LiFeBO3 glass, Solid State Ionics (2016), http://dx.doi.org/10.1016/j.ssi.2016.12.002
ARTICLE IN PRESS 4
P. Michalski et al. / Solid State Ionics xxx (2016) xxx–xxx
electronic conductivity is predominant. The situation is different at higher temperatures, where ionic conductivity becomes more pronounced. This leads to rise of activation energy. In case of samples annealed at higher temperatures (475 ◦ C and 500 ◦ C), due to nanocrystallisation, the electronic conductivity dominates in the whole temperature region. On the other hand, the sample annealed at 400 ◦ C, which is lower than Tg , still represents the same linear behavior as glass (black curve). Impedance spectra collected at 273 ◦ C and 396 ◦ C are presented in Figs. 6 and 7, respectively. As it can be seen, regardless of temperature, both in heating and cooling ramps, spectra consist of two arcs. We postulate that they can be attributed to electronic and ionic parts of the conductivity, respectively. In heating ramps, both spectra were collected at temperatures lower than Tg . Therefore, in these temperatures, internal structure of a glass exhibit only short range order and is bereft of grains and intergrain regions, which can usually explain the presence of arcs in the impedance spectra. At higher temperatures, the ratio of diameters of high and low frequency arcs decreases, meaning that thermally activated motion of ions is becoming more effective. In cooling ramps, after nanocrystallisation, the situation is slightly different — the low frequency arc is becoming less visible as the temperature decreases, meaning that electronic transport is predominant. This can be explained based on Mott’s theory of electronic conductivity via electron hopping in disordered systems [11]:
s(T) = me c(1 − c)
e2 −Ea exp (−2aR) exp RkB T kB T
(1)
where R is the average distance between hopping centres (in this case: Fe2+ and Fe3+ ions), me = /me R2 , a is the inverse localisation length of the electron wave function, c is the fraction of occupied hopping sites for electrons and Ea is the activation energy of electronic conduction. During nanocrystallisation, a large number of nanometer-sized grains precipitate in glassy matrix. This creates an area of connected, highly defected grain–glass interfaces with the high concentration of Fe2+ and Fe3+ ions, between which the electron hopping takes place, creating easy-conduction paths (the c(1 − c) and R factors in
Fig. 7. Impedance spectra collected for sample temperature of 396 ◦ C, taken during heating ramp (red) and cooling ramp, after nanocrystallisation (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Eq. (1) are improved) [12]. When sample is heated to higher temperatures (500 ◦ C in Fig. 5), grain growth prevails over nucleation, the intergrain regions shrink and final conductivity decreases. Compared to literature data [1] (brown line in Fig. 5), our glass– ceramics composite exhibits almost two orders of magnitude better conductivity. This high value may be ascribed to as-mentioned precipitation of defect-rich nanograins and intergrain regions in glassy matrix after annealing. It should be also emphasised that conductivity values of glassy and nanocrystallised samples are not due to iron oxide precipitates. At room temperature, Fe3 O4 is an electronic conductor with conductivity value ca. 100 S cm−1 [13]. As one can see in Figs. 1 and 3, the peaks related to this phase are weak, indicating the inconsiderable content of it. Also, the conductivity of the glass and nanocrystallised sample was lower than 10−11 S cm−1 and 10−4 S cm−1 , respectively. As a consequence, it is expected that the whole contribution to conductivity comes from FeBO3 and LiFeBO3 phases. 4. Conclusions Novel nanocrystalline mixed conductor with similar composition to LiFeBO3 was prepared using original method of thermal nanocrystallisation of amorphous precursor. The material exhibits improved electronic conductivity, which is two orders of magnitude higher compared to polycrystalline one. Mainly, the electronic part of conductivity is improved after nanocrystallisation. It is expected that FeBO3 and LiFeBO3 phases contribute mostly to the conductivity. The electrochemical performance of synthesised best-conducting sample in the Li-ion battery will be studied in the near future.
References
Fig. 6. Impedance spectra collected for sample temperature of 273 ◦ C, taken during heating ramp (red) and cooling ramp, after nanocrystallisation (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
[1] B.L. Ellis, W.R.M. Makahnouk, Y. Makimura, K. Toghill, L.F. Nazar, Nat. Mater. 6 (2007) 749–753. [2] T.K. Pietrzak, Ł. Pawliszak, P.P. Michalski, M. Wasiucionek, J.E. Garbarczyk, Proc. Eng. 98 (2014) 28–35. [3] T.K. Pietrzak, P.P. Michalski, M. Wasiucionek, J.E. Garbarczyk, Solid State Ion. 288 (2016) 193–198. ´ [4] T.K. Pietrzak, M. Wasiucionek, I. Gorzkowska, J.L. Nowinski, J.E. Garbarczyk, Solid State Ion. 251 (2013) 40–46.
Please cite this article as: P. Michalski et al., Novel nanocrystalline mixed conductors based on LiFeBO3 glass, Solid State Ionics (2016), http://dx.doi.org/10.1016/j.ssi.2016.12.002
ARTICLE IN PRESS P. Michalski et al. / Solid State Ionics xxx (2016) xxx–xxx [5] J.E. Shelby, Introduction to Glass Science and Technology, 2005. [6] S. Li, L. Li, J. Liu, J. Jing, X. Li, X. Cui, Electrochim. Acta 155 (2015) 321–326. [7] X. Bian, Q. Fu, F. Du, Y. Gao, L. Zhang, B. Zou, G. Chen, Y. Wei, H, qiu, Chem. Mater. 27 (2015) 5745–5754. [8] S. Afyon, F. Krumeich, C. Mensing, A. Borgschulte, R. Nesper, Nature Sci. Rep. 4 (2014) 7113. [9] K. Hirose, T. Honma, Y. Benino, T. Komatsu, Solid State Ion. 178 (2007) 801–807.
5
[10] Manakova. E, S. Grazulis, D. Chateigner, R.T. Downs, A.T. Yokochi, M. Quiros, L. Lutterotti, J. Butkus, P. Moeck, A. Le Bail, J. Appl. Crystallogr. 42 (2009) 726–729. [11] I.G. Austin, N.F. Mott, Adv. Phys. 50 (2001) 757–812. ´ [12] T.K. Pietrzak, J.E. Garbarczyk, I. Gorzkowska, M. Wasiucionek, J.L. Nowinski, S. Gierlotka, P. Józwiak, ´ J. Power. Sources 194 (2009) 73–80. [13] R. Prakash, R.J. Choudhary, L.S. Sharath Chandra, N. Lakshmi, D.M. Phase, J. Phys. 19 (2007) 486212.
Please cite this article as: P. Michalski et al., Novel nanocrystalline mixed conductors based on LiFeBO3 glass, Solid State Ionics (2016), http://dx.doi.org/10.1016/j.ssi.2016.12.002
Contents lists available at ScienceDirect
Solid State Ionics journal homepage: www.elsevier.com/locate/ssi
Novel nanocrystalline mixed conductors based on LiFeBO3 glass ´ Przemysław P. Michalski* , Tomasz K. Pietrzak, Jan L. Nowinski, Marek Wasiucionek, Jerzy E. Garbarczyk Faculty of Physics, Warsaw University of Technology, Koszykowa 75, Warsaw 00-662, Poland
A R T I C L E
I N F O
Article history: Received 29 July 2016 Received in revised form 30 November 2016 Accepted 6 December 2016 Available online xxxx Keywords: Lithium–borate glass Thermal nanocrystallization Mixed conductor Li-ion batteries Cathodes
A B S T R A C T Novel nanocrystalline mixed conductors were obtained by thermal nanocrystallisation of lithium– iron–borate glasses (nominally LiFeBO3 ) prepared via melt-quenching method. After heat treatment, a glass–ceramics consisting of FeBO3 and LiFeBO3 phases was obtained. Grain sizes, estimated using Scherrer formula, were 45–50 nm and 70–80 nm, respectively. Value of electrical conductivity of the initial glass was estimated as 5.8·10−12 S cm−1 . After annealing at 475 ◦ C, a growth by a factor 2.4·106 was observed, leading to final conductivity value 1.4·10–5 S cm−1 at room temperature. The activation energy of electronic conductivity was lowered from 0.81 eV for glass to 0.18 eV for thermally treated material. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Due to their high gravimetric energy and power, lithium ion (Li-ion) batteries are preferred energy sources for mobile devices and electric cars. Cathode performance and its production costs are the biggest bottlenecks in Li-ion technology development. Therefore, there is a high need for novel and inexpensive cathode materials. Lithium iron borate (LiFeBO3 ) is one of the potential candidates. Due to low molar mass of the borate group, its gravimetric capacity reaches 220 mAhg−1 with one lithium ion exchanged. Also, thanks to high abundance of the substrate elements in Earth’s crust, the material can be produced at low cost; it is also environmentally benign. On the other hand, LiFeBO3 exhibits low electronic conductivity – when prepared in polycrystalline form, the value is 3.9 • 10–7 S cm −1 [1], which is too low to obtain high current rates in Li-ion battery. It is also known that electrochemical performance of LiFeBO3 is strongly reduced for grain sizes bigger than 100 nm [1]. Here, we would like to propose a novel method of materials’ preparation, based on thermal nanocrystallisation of initially amorphous precursors. This method allows to obtain materials with highly enhanced conductivity and nanometric-sized grains. Recently, we have successfully applied this approach in case of cathode-like V2 O5 –P2 O5 [2], LiF–V2 O3 –P2 O5 [3] and Li2 O–FeO– V2 O5 –P2 O5 [4] systems. In the latter case, the material with final
* Corresponding author. E-mail address: [email protected] (P. Michalski).
conductivity value of 10 −3 S cm−1 and microstructure consisting of 3–5 nm grains of LiFePO4 olivine was obtained. Inspired by previous results, we are determined to study other glass systems. A proposed borate system seems to be an ideal candidate for investigations because of a few reasons. First of all, boron (III) oxide is a good glass-former 5]. Also, boron-related materials are good Li+ conductors, which allows their use as electrolytes (e.g. LIBOB salt [6]), glassy coatings on cathode materials providing better electrochemical efficiency [7] or as glassy cathodes exhibiting high performance itself [8]. 2. Experimental Glasses of nominal composition 0.25Li2 O • 0.5FeO • 0.25 B2 O3 (LiFeBO3 ) were prepared. Starting reagents: Li2 CO3 (Aldrich, 99.9%), FeC2 O4 • 2H2 O (Aldrich, 99.9%) and H3 BO3 (Aldrich, 99.5%) were mixed and homogenised in Retsch planetary mill (400 rpm, 20 min). The obtained batches were melted for 10 min in furnace preheated to temperatures in 1000–1250 ◦ C range. Double crucible method [9] was used to prevent iron oxidation. The melted batches were rapidly poured on a metal plate and quickly pressed with another metal plate (melt-quenching method. X-ray diffractometry (Phillips X’Pert Pro with Bragg–Brentano configuration, using CuKa line with k = 1.5406 Å) was used to verify the amorphousness of obtained samples and for temperaturedependent investigations. In the latter case, Anton Paar HTK1200 oven was used and the measurements were conducted in nitrogen flow. The temperature was stabilised for 40 min before each measurement.
http://dx.doi.org/10.1016/j.ssi.2016.12.002 0167-2738/© 2016 Elsevier B.V. All rights reserved.
Please cite this article as: P. Michalski et al., Novel nanocrystalline mixed conductors based on LiFeBO3 glass, Solid State Ionics (2016), http://dx.doi.org/10.1016/j.ssi.2016.12.002
ARTICLE IN PRESS 2
P. Michalski et al. / Solid State Ionics xxx (2016) xxx–xxx
Temperatures of glass transition and crystallisation were obtained using differential thermal analysis method. TA Instruments SDT Q600 instrument was used. Measurements were performed with 1C min−1 and 10C min−1 heating rates, in argon flow. Electrical properties were studied in situ during temperature ramps, using Solartron 1260 Gain Phase/Impedance Analyzer integrated with a furnace and a temperature programming module (Eurotherm 2404). Before each measurement, the temperature was stabilised for 40 min. Maximum temperatures were chosen in 400–500 ◦ C range. Impedance spectra were collected in the 10 mHz– 10 MHz frequency range for each temperature in the ramp. The applied voltage amplitude was equal to 0.1 V.
3. Results and discussion 3.1. XRD of as-prepared samples For all melting temperatures in the investigated range, the materials obtained after quenching were amorphous with small traces of non-stoichiometric Fe3 O4 phase (Fig. 1). The best match was obtained for Fe2.897 O4 phase (PDF-2 reference code: 01-086-1337, later referred as Fe3 O4 ). The amount of this phase decreased with increasing melting temperature. This was the reason why the material obtained by melting at the highest temperature (1250 ◦ C) was chosen for further studies. The existence of non-stoichiometric Fe3 O4 phase is also an indication of presence of Fe2+ and Fe3+ ions in glass, which are relevant for electronic hopping conductivity.
3.2. DTA of as-prepared samples DTA curves obtained for 1 ◦ C min−1 (blue) and 10 ◦ C min−1 (red) are typical for glassy materials (Fig. 2) and consist of glass transition step which is interpreted as a change of heat capacity during devitrification, followed by a few exothermic peaks which may be attributed to crystallisation of different phases. The characteristic temperatures were collected in Table 1. For lower rates, the transitions were less pronounced and the temperatures could be only estimated. This fact can be explained by the nature of DTA experiment, where the heat difference between analyzed and reference sample is measured. This difference can be written as follows: dQ dQ dT dT = = cp dt dT dt dt
Fig. 2. DTA curves collected with heating rates 10 ◦ C min−1 (red) and 1 ◦ C min−1 (blue). Characteristic thermal events are marked by arrows. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
where cp is the heat capacity of the material. Assuming cp constant in current temperature, one can see that the heat difference dQ/dt is proportional to heating rate dT/dt. 3.3. Temperature XRD investigations The results of XRD temperature-dependent measurements are presented in Fig. 3. Due to existence of small Fe3 O4 inclusions in initial glass, this phase is the first to crystallise at temperatures in the 425–450 ◦ C range. These are in proximity of glass transition and first crystallisation peak temperatures for low heating rate. As soon as the glass devitrifies (Tg ), atoms movement is facilitated and they may form crystal structure based on initially present iron oxide units. The other phases – delithiated, hexagonal iron borate phase and monoclinic lithium iron borate phase – crystallise at ca. 525–550 ◦ C. This temperature may be ascribed to second crystallisation peak (Tc2 ), while the third one could be correlated with recrystallisation of lithium boron oxide (Li2 B4 O7 , reference code: 00-011-0408). The peak matched by this pattern firstly occurs at 525 ◦ C at 2H ≈ 22◦ . Up to 600 ◦ C, its intensity increases and then, for higher temperatures, decreases, to almost vanish at 650 ◦ C. Drawing conclusions based on DTA data collected for 1 ◦ C min−1 is reasonable, because similar heating rate occurred during stabilisation of temperature in XRD measurements. The FeBO3 /LiFeBO3 patterns were obtained from Crystallography Open Database [10] (COD IDs 1511131 and 1511129, respectively). After cooling down the sample to room temperature, the grain sizes were estimated using Scherrer formula, taking into calculations the most intense peaks of FeBO3 (the peaks of lithiated phase were hardly visible). The obtained result is 20–45 nm. The intensities of peaks of FeBO3 /LiFeBO3 phases in XRD data acquired during temperature-dependent measurements were relatively low, indicating slight content of these phases in obtained glass-ceramics. Therefore, in the next step, the as-prepared glass was isothermally annealed at 550 ◦ C for 12 h (purple line in Fig. 4). The temperature was chosen as the one, at which FeBO3 /LiFeBO3 peaks firstly occurred. During temperature-dependent measurements (green line), the sample was kept in one temperature for ca. Table 1 Temperatures of characteristic thermal events.
Fig. 1. X-ray diffractograms of glasses prepared with different temperatures of melting (shown on the right).
Heating rate
Tg [◦ C]
Tc1 [◦ C]
Tc2 [◦ C]
Tc3 [◦ C]
1 ◦ C min−1 10 ◦ C min−1
421 435
454 476
502 546
588 639
Please cite this article as: P. Michalski et al., Novel nanocrystalline mixed conductors based on LiFeBO3 glass, Solid State Ionics (2016), http://dx.doi.org/10.1016/j.ssi.2016.12.002
ARTICLE IN PRESS P. Michalski et al. / Solid State Ionics xxx (2016) xxx–xxx
3
One can roughly estimate the FeBO3 /LiFeBO3 phase ratio, by comparing the area of the most intense peaks of these phases (2H = 25.18◦ and 35.37◦ , respectively). The calculated ratio is 75/25. Similarly to the previous sample, grain sizes were estimated using Scherrer formula. The values of 45–50 nm and 70–80 nm were obtained for FeBO3 and LiFeBO3 , respectively. 3.4. Electrical investigations The room temperature conductivity value of the initial glass was too low to be measured. After extrapolation, it was calculated as 5.8 • 10−12 Scm−1 . The activation energy was equal to 0.81 eV (Fig. 5). Heat treatment at different temperatures between Tg and Tc2 led to irreversible conductivity growth. The best results were obtained for 475 ◦ C — after annealing, the conductivity value at room temperature increased by a factor 2.4 • 106 leading to the final value 1.4 • 10−5 S cm−1 , while the activation energy was lowered to 0.18 eV. One can also see that cooling ramps for annealing temperatures 425 ◦ C and 450 ◦ C consist of two parts with different slopes. The higher slope is present in higher temperatures. This phenomenon may be explained as follows: the activation energy value for ion is higher than for electrons. Due to that, at low temperatures, the
Fig. 3. XRD patterns collected during temperature-dependent measurements. FeBO3 peaks positions are marked by blue circles and LiFeBO3 – by red diamonds. Peak ascribed to Li2 B4 O7 is marked by asterisk. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
1 h (stabilisation and measurement time). It can be seen that longer annealing time led to further growth of FeBO3 and LiFeBO3 phases, as peaks related to these phases became more pronounced. What is more, the most intense peak matched by Fe3 O4 pattern present at 2H ≈ 35◦ in data collected during temperature ramp has split into two, matched by LiFeBO3 pattern. It means that present iron oxide phases possibly recrystallise to desirable lithium iron borate phases.
◦
Fig. 4. Comparison of XRD patterns collected at 550 C: during isothermal annealing for 12 h (purple); during temperature ramp (green). Dashed lines indicate peaks positions of FeBO3 (blue) and LiFeBO3 (red) peaks. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 5. Temperature dependence of electric conductivity presented in Arrhenius coordinates. Black points represent the data obtained during heating ramp, points in different colours – data collected during cooling ramps after heating to various temperatures. DTA curve for 1 ◦ C min−1 and the conductivity value for crystalline LiFeBO3 [1] (brown line) are shown for comparison. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Please cite this article as: P. Michalski et al., Novel nanocrystalline mixed conductors based on LiFeBO3 glass, Solid State Ionics (2016), http://dx.doi.org/10.1016/j.ssi.2016.12.002
ARTICLE IN PRESS 4
P. Michalski et al. / Solid State Ionics xxx (2016) xxx–xxx
electronic conductivity is predominant. The situation is different at higher temperatures, where ionic conductivity becomes more pronounced. This leads to rise of activation energy. In case of samples annealed at higher temperatures (475 ◦ C and 500 ◦ C), due to nanocrystallisation, the electronic conductivity dominates in the whole temperature region. On the other hand, the sample annealed at 400 ◦ C, which is lower than Tg , still represents the same linear behavior as glass (black curve). Impedance spectra collected at 273 ◦ C and 396 ◦ C are presented in Figs. 6 and 7, respectively. As it can be seen, regardless of temperature, both in heating and cooling ramps, spectra consist of two arcs. We postulate that they can be attributed to electronic and ionic parts of the conductivity, respectively. In heating ramps, both spectra were collected at temperatures lower than Tg . Therefore, in these temperatures, internal structure of a glass exhibit only short range order and is bereft of grains and intergrain regions, which can usually explain the presence of arcs in the impedance spectra. At higher temperatures, the ratio of diameters of high and low frequency arcs decreases, meaning that thermally activated motion of ions is becoming more effective. In cooling ramps, after nanocrystallisation, the situation is slightly different — the low frequency arc is becoming less visible as the temperature decreases, meaning that electronic transport is predominant. This can be explained based on Mott’s theory of electronic conductivity via electron hopping in disordered systems [11]:
s(T) = me c(1 − c)
e2 −Ea exp (−2aR) exp RkB T kB T
(1)
where R is the average distance between hopping centres (in this case: Fe2+ and Fe3+ ions), me = /me R2 , a is the inverse localisation length of the electron wave function, c is the fraction of occupied hopping sites for electrons and Ea is the activation energy of electronic conduction. During nanocrystallisation, a large number of nanometer-sized grains precipitate in glassy matrix. This creates an area of connected, highly defected grain–glass interfaces with the high concentration of Fe2+ and Fe3+ ions, between which the electron hopping takes place, creating easy-conduction paths (the c(1 − c) and R factors in
Fig. 7. Impedance spectra collected for sample temperature of 396 ◦ C, taken during heating ramp (red) and cooling ramp, after nanocrystallisation (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Eq. (1) are improved) [12]. When sample is heated to higher temperatures (500 ◦ C in Fig. 5), grain growth prevails over nucleation, the intergrain regions shrink and final conductivity decreases. Compared to literature data [1] (brown line in Fig. 5), our glass– ceramics composite exhibits almost two orders of magnitude better conductivity. This high value may be ascribed to as-mentioned precipitation of defect-rich nanograins and intergrain regions in glassy matrix after annealing. It should be also emphasised that conductivity values of glassy and nanocrystallised samples are not due to iron oxide precipitates. At room temperature, Fe3 O4 is an electronic conductor with conductivity value ca. 100 S cm−1 [13]. As one can see in Figs. 1 and 3, the peaks related to this phase are weak, indicating the inconsiderable content of it. Also, the conductivity of the glass and nanocrystallised sample was lower than 10−11 S cm−1 and 10−4 S cm−1 , respectively. As a consequence, it is expected that the whole contribution to conductivity comes from FeBO3 and LiFeBO3 phases. 4. Conclusions Novel nanocrystalline mixed conductor with similar composition to LiFeBO3 was prepared using original method of thermal nanocrystallisation of amorphous precursor. The material exhibits improved electronic conductivity, which is two orders of magnitude higher compared to polycrystalline one. Mainly, the electronic part of conductivity is improved after nanocrystallisation. It is expected that FeBO3 and LiFeBO3 phases contribute mostly to the conductivity. The electrochemical performance of synthesised best-conducting sample in the Li-ion battery will be studied in the near future.
References
Fig. 6. Impedance spectra collected for sample temperature of 273 ◦ C, taken during heating ramp (red) and cooling ramp, after nanocrystallisation (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
[1] B.L. Ellis, W.R.M. Makahnouk, Y. Makimura, K. Toghill, L.F. Nazar, Nat. Mater. 6 (2007) 749–753. [2] T.K. Pietrzak, Ł. Pawliszak, P.P. Michalski, M. Wasiucionek, J.E. Garbarczyk, Proc. Eng. 98 (2014) 28–35. [3] T.K. Pietrzak, P.P. Michalski, M. Wasiucionek, J.E. Garbarczyk, Solid State Ion. 288 (2016) 193–198. ´ [4] T.K. Pietrzak, M. Wasiucionek, I. Gorzkowska, J.L. Nowinski, J.E. Garbarczyk, Solid State Ion. 251 (2013) 40–46.
Please cite this article as: P. Michalski et al., Novel nanocrystalline mixed conductors based on LiFeBO3 glass, Solid State Ionics (2016), http://dx.doi.org/10.1016/j.ssi.2016.12.002
ARTICLE IN PRESS P. Michalski et al. / Solid State Ionics xxx (2016) xxx–xxx [5] J.E. Shelby, Introduction to Glass Science and Technology, 2005. [6] S. Li, L. Li, J. Liu, J. Jing, X. Li, X. Cui, Electrochim. Acta 155 (2015) 321–326. [7] X. Bian, Q. Fu, F. Du, Y. Gao, L. Zhang, B. Zou, G. Chen, Y. Wei, H, qiu, Chem. Mater. 27 (2015) 5745–5754. [8] S. Afyon, F. Krumeich, C. Mensing, A. Borgschulte, R. Nesper, Nature Sci. Rep. 4 (2014) 7113. [9] K. Hirose, T. Honma, Y. Benino, T. Komatsu, Solid State Ion. 178 (2007) 801–807.
5
[10] Manakova. E, S. Grazulis, D. Chateigner, R.T. Downs, A.T. Yokochi, M. Quiros, L. Lutterotti, J. Butkus, P. Moeck, A. Le Bail, J. Appl. Crystallogr. 42 (2009) 726–729. [11] I.G. Austin, N.F. Mott, Adv. Phys. 50 (2001) 757–812. ´ [12] T.K. Pietrzak, J.E. Garbarczyk, I. Gorzkowska, M. Wasiucionek, J.L. Nowinski, S. Gierlotka, P. Józwiak, ´ J. Power. Sources 194 (2009) 73–80. [13] R. Prakash, R.J. Choudhary, L.S. Sharath Chandra, N. Lakshmi, D.M. Phase, J. Phys. 19 (2007) 486212.
Please cite this article as: P. Michalski et al., Novel nanocrystalline mixed conductors based on LiFeBO3 glass, Solid State Ionics (2016), http://dx.doi.org/10.1016/j.ssi.2016.12.002