Electric field-assisted ultrafast synthesis of nanopowders: a novel and cost-efficient approach

Abstract

A current trend in materials processing is the synthesis of high-reactivity powders with a particle size averaged at the nanoscopic scale. This is normally achieved by considering chemical routes that allow processing of the materials at temperatures markedly lower than those required in the conventional method. Here we introduce a simple but novel approach that enables ultrafast synthesis of materials by using electric fields. The case of CaCu₃Ti₄O₁₂, where traditional chemical methods have usually revealed unable to synthesize a nanosized single-phase powder, is presented. In zero-field processing, the end-product powder prepared here via a modified polymeric precursor method exhibited, for instance, an average particle size of 300 nm. In the following, we show that thermal cycling of the precursor powder under an electric field input leads to a substantial drop in synthesis temperature, attributable to enhanced charge diffusion processes, ending with an average particle size sensibly reduced and, finally, rendering possible the production of nanopowders (<100 nm) by adjusting the maximum electric current allowed to flow across the material during processing.

---

1. Introduction

Functional materials are of common use in many electro-electronic devices like capacitors, gas sensors, thermistors and varistors,^1,2 frequently in the ceramic form. Processing of ceramics involves two main steps, namely, calcination at a low but sufficient temperature to synthesize the end-product powder of concern, and sintering at a high temperature for densification. Challenges within this topic include low-cost processing, maximization of densification and minimization of grain growth. Controlling the competition between the two latter is extremely difficult because the driving forces for both are proportional to the reciprocal of grain size and, hence, comparable in magnitude.^3,4 Solutions aimed at producing dense and fine-grained ceramics include lowering calcination temperatures so as to synthesize high-reactivity starting powders, i.e., powders with a particle size as small as possible. This can be accomplished by considering chemical routes (sol–gel, Pechini, etc.) that normally allow producing powders with nanosized (<100 nm) particles.^5–7 Compacts with such ultrafine particles present higher effective area and surface energy, implying increased grain boundary mobility and, therefore, sinterability towards lower temperatures when compared to processing of compacts formed by micrometric particles.

In the present work, we introduce a novel approach to this issue of synthesis of materials, which is based on application of an electric field during calcination of the precursor powder. Among the compounds we have been successfully testing, an example is here given for CaCu₃Ti₄O₁₂ (CCTO), a material which has recently attracted the attention of researchers because showing a giant dielectric constant with good stability over a wide range of temperatures.^8–11 This dielectric scenario is theoretically ideal for the manufacture of high-quality capacitors with, for instance, a significantly reduced size, as required in microelectronic devices. For CCTO, however, even well-established chemical routes like pyrolysis and combustion have unfortunately revealed ineffective to produce starting powders with a particle size averaged at the nanoscopic scale.^12,13 Here we demonstrate that, by using electric fields, synthesis of this compound can be achieved at times and furnace temperatures significantly lower than those required in zero-field approach, a novel and cost-efficient procedure which is shown to allow production of nanopowders.

2. Results and discussion

The starting point in this work was a precursor CCTO powder that we synthesized via a Pechini-derived chemical method (see details later in the Experimental section and in ref. 14), followed by annealing at 400 °C for 2 h to remove most organic compounds, resulting in an amorphous product as will be shown below. Optimization of the conditions to be finally considered for producing a single crystalline CCTO phase included performing simultaneous differential thermal (DTA) and thermogravimetry (TGA) analyses of such precursor powder, the results of which are depicted in Fig. 1. With increasing temperature, development of several physico-chemical processes, namely, three exothermic and two endothermic events is sensed. Concerning the two last processes, the one occurring at around 1015 °C originates from eutectic point-related liquid phase formation, as observed in CuO–TiO₂ binary systems,¹⁵ while that taking place at around 1150 °C results from CCTO melting. Here we will focus on those physico-chemical processes manifesting as exothermic peaks towards lower temperatures. A proper analysis of origin of these events included also collecting and examining the X-ray diffraction (XRD) data from the precursor powder after annealing at the following selected temperatures: 400, 600, 700 and 800 °C for 2 h, the patterns of which are displayed in Fig. 2a. Accordingly, the exothermic peak observed around 225 °C in Fig. 1 (DTA data), and involving a weight loss of about 1.7% (TGA data), is simply associated with combustion leading to partial elimination of residual organic compounds still present in the powder. No crystalline phase-related chemical synthesis occurs at this instance, noting that even after annealing at 400 °C for 2 h the powder still reveals amorphous (Fig. 2a).

Fig. 1 Differential thermal and thermogravimetric (DTA/TGA) analysis curves of the original CCTO precursor powder after pre-calcination at 400 °C for 2 h.

Fig. 2 (a) XRD patterns of the CCTO precursor powder after annealing at 400, 600, 700 and 800 °C for 2 h. The starting powder is basically amorphous, followed by formation of: TiO₂ (ICSD 200392), CuO (ICSD 67850) and CCTO (ICSD 32002; Miller index numbers given between parentheses). (b) SEM micrograph of the powder annealed at 800 °C for 2 h.

Still according to the XRD data in Fig. 2a, the exothermic peak developing at around 625 °C (Fig. 1) is to be associated with crystallization of CuO and TiO₂ phases, a stage during which there is a continuous removal of the organic compounds, equivalent to about 1.4% of weight loss (TGA data, Fig. 1). Towards this temperature region, we note that annealing the original powder for a considerable time period may reveal enough to start producing a more or less important amount of CCTO, as verified for the powder heat treated at 600 °C for 2 h (Fig. 2a). Finally, the exothermic process detected at around 710 °C (Fig. 1) is related to the chemical reaction (occurring at the highest rate) of the intermediate phases giving CCTO (Fig. 2a); in this stage, removal of the remaining organic compounds is completed, representing a weight loss of about 1.8% (TGA data, Fig. 1). The results in this Fig. 2a, in which incidence of residual amounts of CuO and TiO₂ for the powder heat treated at 700 °C for 2 h is still observed, allow concluding that calcination of this Pechini-derived precursor powder should be better performed at about 800 °C in order to produce a second phase-free CCTO powder. Fig. 2b is a scanning electron micrograph (SEM) corresponding to the powder after annealing at 800 °C for 2 h, showing the microstructure to consist of particles with an average size of 0.3 ± 0.1 μm. Similar results of optimal synthesis temperature falling towards the 700–800 °C range and final average particle size in the 200–400 nm range have been reported for end CCTO powders processed through wet-chemistry (pyrolysis and combustion) methods.^12,13

In the following, starting from the same precursor powder, we proceeded with evaluation of the synthesis kinetics of CCTO in the presence, now, of an electric field (E). The current (J) flowing across the sample under test was monitored during these experiments, the data of which are displayed as a function of furnace temperature in Fig. 3, with E going from low values to 240 V cm⁻¹. An abrupt rise of the current towards the last stage of heat treatment is observed, according to which a maximum current flowing across the specimen should be pre-set to avoid electrical-promoted physical damage of such a body; this safe current was, e.g., around J = 22.0 mA mm⁻² in Fig. 3. This scenario of a sudden increase in electric current during heat treatment reproduces well what is observed in the recently introduced processing approach called electric field-assisted flash sintering of materials,^16–18 with the difference that, in the present work, no densification of the material was observed at the end of the experiments.

Fig. 3 Current density (J) against furnace temperature during thermal cycling of the CCTO precursor powder (originally amorphous) under electric field (E) input. The current limit was pre-set to J = 22.0 mA mm⁻², followed by a holding time of 1 min. The indicated points (a) to (f) are discussed in the text, in connection with Fig. 4.

The inset in Fig. 3 refers to a magnification of the current density data towards low temperatures. Incidence of an ‘anomaly’ manifesting as J peak is noted. Moreover, the entire J curves shift towards lower temperatures as E increases, meaning progressively enhanced kinetics of the processes involved, whose origin is in the following analysed. The points (a) to (f) indicated in this Fig. 3 are related to representative instances where the experiments were stopped and XRD analyses conducted, the patterns of which are shown in Fig. 4. For point (a) we simply recuperated the 400 °C-XRD data shown in Fig. 2a, and according to which the starting Pechini-derived product is basically amorphous. The XRD data applying at point (b) are indicative that incidence of the J peaks (Fig. 3 inset) is indeed associated with crystallization of CuO and TiO₂; at this stage, we note that production of some trace of CCTO also occurred. At points (c) to (f), still according to Fig. 4, a single CCTO phase was synthesized. In other words, under field action, the powder transits in these single thermal cycling experiments from amorphous to crystallization through intermediate phases, followed by flash synthesis of CCTO. This is the ultrafast synthesis of end-product powders we introduce here, achievable in this work at a furnace temperature as low as 440 °C (for E = 240 V cm⁻¹), i.e., far below the temperature of about 800 °C required when considering conventional (zero-field) processing of CCTO (Fig. 1 and 2a).

Fig. 4 XRD patterns corresponding to the points (a) to (f) indicated in Fig. 3. (a) Starting powder at 400 °C; specimens at points (b) 515 °C (E = 200 V cm⁻¹), (c) 677 °C (E = 60 V cm⁻¹), (d) 562 °C (E = 200 V cm⁻¹) and (e) 440 °C (E = 240 V cm⁻¹); and at point (f) for a limiting current density reduced to J = 7.5 mA mm⁻² (E = 200 V cm⁻¹). The starting powder transits from amorphous to intermediate phases: TiO₂ (ICSD 200392) and CuO (ICSD 67850), and then to CCTO (ICSD 32002; Miller index numbers given between parentheses).

Fig. 5 shows the behaviour of the flash and crystallization temperatures (T_flash and T_cryst, respectively) against electric field, as we evaluated from Fig. 3. For simplicity, T_cryst was taken as the temperature at the current peak (Fig. 3 inset). Both T_flash and T_cryst decrease with raising E, and assume an identical value for E ≥ 220 V cm⁻¹, promoted by a steeper drop of T_flash. For E = 0 V cm⁻¹ we have in Fig. 5 included T^(DTA)_cryst and T^(DTA)_synth which refer, respectively, to the temperatures of crystallization and synthesis as extracted from the DTA curve depicted in Fig. 1 (we have also for simplicity taken the peak temperature values). Notice, quite important, that these data are coherent in terms of trend with the data appraised from processing under field input, validating once more our original interpretation that these physico-chemical processes are responsible for the distinctive features (anomalies) observed in Fig. 3. Overall, Fig. 5 summarizes well the scenario according to which there is, under field input, accelerated kinetics of crystallization (of the intermediate phases) and chemical reaction giving the single CCTO phase during thermal cycling of the original precursor powder. For crystallization, just to start with, this picture is in principle to be expected because a decrease in Gibbs free energy for formation of nuclei with the critical size (ΔG_C), demonstrable to satisfy ΔG_C = (16/3)πγ_S³/[|ΔG_V| + (1/2)ε_o(K_f − K_i)E²]², applies in the presence of an electric field,^19,20 as long as the dielectric constant of the new medium (K_f) is higher than in the original medium (K_i). ΔG_V refers to the strain energy per unit volume, ε_o is the permittivity of vacuum, and γ_S the surface energy per unit area. The K values measured in this work (at room temperature and 100 kHz, on pressed powders with comparable void volume fraction) were K_(a) = 85 and K_(b) = 110 for the powders from points (a) and (b), in connection with Fig. 3. This reasoning also applies for the synthesis event arising from reaction of the intermediate phases, where formation of the end CCTO product also involves crystallite nucleation and growth processes, and obeys equal thermodynamic approach for ΔG_C; in relation to Fig. 3, we found K_(f) = 143 for the CCTO powder from point (f).

Fig. 5 Evolution of crystallization (T_cryst) and flash synthesis (T_flash) temperatures with varying the electric field (E) applied during the processing experiments. The dashed lines are guide for the eyes, delimitating the regions of amorphous precursor powder, intermediate-phase powder and end-product (single-phase) powder, as identified in this work. The crystallization, T^(DTA)_cryst, and (CCTO) synthesis, T^(DTA)_synth, temperatures extracted from the DTA results (Fig. 1) have been included for E = 0 V cm⁻¹.

In summary, increasing E under K_f > K_i has the expected effect of decreasing progressively ΔG_C, implying improved kinetics of the physico-chemical (crystallization and synthesis) processes, as observed in this work. Besides the effect from a decrease in ΔG_C, incidence of synthesis as a flash event (Fig. 3 and 4) might involve, in parallel, some contribution from Joule heating (power dissipation, P = J × E) modulating the real sample temperature and/or, most likely, a generation of defects in the form of avalanche promoted by the applied field, as also proposed elsewhere for flash sintering.^17,21 This is a complex issue to be further explored in future works. We just note, first, that CCTO formation was here still stimulated even after drastically reducing J at flash from 22.0 to, for instance, 7.5 mA mm⁻² (Fig. 3 and 4); the latter current value represents a low power dissipation density of 0.180 W mm⁻³ when E = 240 V cm⁻¹ to only 0.045 W mm⁻³ when E = 60 V cm⁻¹. Second, also quite important, note that power dissipation at the onset of flash is, in each (field) case, comparatively even lower.

Fig. 6 shows two representative micrographs of the end CCTO powder after synthesis at E = 200 V cm⁻¹ under maximum currents pre-set to J = 22.0 mA mm⁻² (Fig. 6a) and J = 7.5 mA mm⁻² (Fig. 6b). The values of average particle size (APS) processed for all the powders whose microstructures were imaged are illustrated in Fig. 7. Compared to conventional synthesis (closed circle), it is clear that application of an electric field during thermal cycling of the precursor powder inhibits sensibly particle coarsening, traducing into a reduced APS in the end-product powder (open symbols). To distinguish among flash synthesis conditions and microstructure characteristics, the open circles in this graph correspond to synthesis conducted allowing the current density at flash to reach J = 22.0 mA mm⁻². What is stimulating is to observe that APS tends to the nanoscopic scale as E is increased. Of finally great relevance to this work is the realization that lowering the maximum current (across the material) at flash to J ≤ 12.0 mA mm⁻² (for E = 200 and 240 V cm⁻¹, for instance) allowed us to definitively synthesize CCTO nanopowders, as can be seen, for instance, in the image shown in Fig. 6b. These are the data illustrated in Fig. 7 towards the scratched region, and plotted in the Fig. 7 inset in terms of APS vs. J to better visualize the effect from J. All these results indicate that using electric fields during materials processing opens a new strategy for synthesizing nanopowders at times and furnace temperatures significantly lower than in conventional approach. Although the physical mechanism (field-modulated temperature and/or defect avalanche generation) behind origin and development of the flash synthesis phenomenon we have introduced here needs to be further and carefully investigated, the present data at least show that lowering J at flash has the effect of sensibly decreasing APS, a fact attributable to a diminished Joule heating contribution to the real and final processing temperature.

Fig. 6 SEM micrographs of CCTO powders flash-synthesized at E = 200 V cm⁻¹ under pre-set current density limits of (a) 22.0 mA mm⁻² and (b) 7.5 mA mm⁻².

Fig. 7 Average particle size (APS) upon variation of the electric field applied to promote flash synthesis of CCTO. The open circles refer to experiments conducted with the current density limited to J = 22.0 mA mm⁻². The remaining data correspond to flash under: E = 200 V cm⁻¹ with the current limited to J = 10.0 mA mm⁻² (open triangle up), J = 7.5 mA mm⁻² (open triangle down) and J = 5.0 mA mm⁻² (open square), and finally E = 240 V cm⁻¹ with the current limited to J = 12 mA mm⁻² (open diamond). The figure inset refers to the effect of the pre-set maximum current during flash synthesis on APS for flash under E = 200 V (open stars) and E = 240 V cm⁻¹ (closed stars). The dashed lines are just a guide for eyes.

3. Conclusions

Synthesizing a CaCu₃Ti₄O₁₂ (CCTO) powder with an average particle size as small as possible was explored in this work. We have demonstrated that materials processing can be remarkably accelerated by using electric fields, a scenario that apparently arises from a decrease in Gibbs free energy for formation of nuclei with critical size. The consequences are development of crystallization and chemical reaction processes at times and temperatures significantly lower than in zero-field processing approach and, hence, synthesis of a finer end-product powder. In particular, it is shown that adjusting both electric field strength and maximum current allowed to flow across the precursor powder during thermal cycling makes possible production of a single-phase CCTO nanopowder where zero-field processing fails or may fail, as found here and elsewhere for this compound.

4. Experimental procedure

CaCu₃Ti₄O₁₂ (CCTO) powder was obtained by applying a modified (Pechini-derived) polymeric precursor method.^5,14 Calcium, copper and titanium citrates were separately prepared using calcium carbonate (CaCO₃ – Synth, 99%), copper nitrate (Cu(NO₃)₂·3H₂O – Labsynth, 98%) and titanium isopropoxide (Ti[OCH(CH₃)₂]₄ – Alfa Aesar, 97%) mixed in stoichiometric amounts with citric acid (CA), previously dissolved in distilled water (0.1 g ml⁻¹), at the molar ratios of 1 : 3 (Ca : CA) for calcium and 1 : 6 (Cu, Ti : CA) for copper as well as titanium. Ethylene glycol (EG) was then added to these solutions, at the mass ratio of CA : EG = 60 : 40, to promote citrate polymerization by polyesterification. The (Ca, Cu, Ti)-containing resin was prepared by mixing these polyesters, followed by pH adjustment to 9 by adding ammonium hydroxide. A stable resin with a blue colour and transparent appearance was then obtained, and heated to 120 °C to eliminate excess water, followed by heat treatment at a temperature as low as 400 °C for 2 h to remove most organic compounds. Thermal analyses of this CCTO precursor powder were performed using simultaneous differential thermal (DTA) and thermogravimetry (TGA) analyses, with data collected using a SDT 2960 apparatus – TA Instruments, at a heating rate of 10 °C min⁻¹, in a flow of synthetic air (O₂/N₂ – 1/4), from room temperature to 1200 °C. Phase development in such a powder was monitored by X-ray diffraction (XRD) measurements conducted after annealing this powder at selected temperatures, as discussed in the text. These measurements were carried out at room temperature using a Rigaku RINT 2000/PC equipment, operating with CuKα radiation, in continuous mode, with 2θ varying from 20° to 80° in steps of 0.02°. On the basis of these analyses, the optimal conditions to produce a single-phase CCTO powder were established. The non-conventional approach explored in this work consisted in subjecting the precursor powder to heat treatment in the presence of an electric field (in the direct current mode), with values going from low to 240 V cm⁻¹. Processing was conducted at a heating rate of 10 °C min⁻¹, while the maximum current flowing across the samples was pre-set to values ranging from 22.0 to only 5.0 mA mm⁻². A Sorensen 300-2 DC power supply was used, while the current was measured with a digital Keithely 2000 multimeter. Phase development in the field-processed materials was also monitored by X-ray diffraction analyses (XRD, Rigaku RINT 2000/PC equipment). The microstructures of the end-product powders were analyzed through a field-emission scanning electron microscope (SEM-FEG, JEOL JSM-7401F), after which the average particle size (APS) was estimated by applying the linear intercept method.²² The dielectric constant of representative samples was measured at room temperature and 100 kHz using a Solartron SI 1260 impedance analyzer, on powders with comparable porosity volume fraction after compaction.

Acknowledgements

The authors are grateful to CAPES, a Brazilian funding agency, for the support through grants no. BEX 3276/14-7 and BEX 9291/13-0. S. K. Jha and J.-M. Lebrun are also acknowledged for technical assistance with some of the experiments related to field-assisted processing, as well as B. E. Francisco for help with some of the XRD measurements.

References

1. Y.-M. Chiang, D. P. Birnie III and W. D. Kingery, Principles for Ceramic Science and Engineering, Physical Ceramics, John Wiley & Sons, Inc., NY, USA, 1997.
2. K. Wang, F.-Z. Yao, W. Jo, D. Gobeljic, V. V. Shvartsman, D. C. Lupascu, J.-F. Li and J. Rödel, Adv. Funct. Mater., 2013, 23, 4079–4086.
3. I. Chen and X.-H. Wang, Nature, 2000, 404, 168–171.
4. C. P. Cameron and R. Raj, J. Am. Ceram. Soc., 1988, 71, 1031–1035.
5. M. P. Pechini, U.S. Pat. no 3,330.697, 1967.
6. R. S. Silva, M. I. B. Bernardi and A. C. Hernandes, J. Sol-Gel Sci. Technol., 2007, 42, 173–179.
7. R. S. Silva, A. C. Hernandes and J.-C. M'Peko, Mater. Res., 2012, 15, 522–529.
8. D. C. Sinclair, T. B. Adams, F. D. Morrison and A. R. West, Appl. Phys. Lett., 2002, 80, 2153–2155.
9. T. B. Adams, D. C. Sinclair and A. R. West, Adv. Mater., 2002, 14, 1321–1323.
10. J. Li, M. A. Subramanian, H. D. Rosenfeld, C. Y. Jones, B. H. Toby and A. W. Sleight, Chem. Mater., 2004, 16, 5223–5225.
11. P. Fiorenza, V. Raineri, M. C. Ferrarelli, D. C. Sinclair and R. L. Nigro, Nanoscale, 2011, 3, 1171–1175.
12. J. Liu, Y. Sui, C.-G. Duan, W.-N. Mei, R. W. Smith and J. R. Hardy, Chem. Mater., 2006, 18, 3878–3882.
13. J. Liu, R. W. Smith and W.-N. Mei, Chem. Mater., 2007, 19, 6020–6024.
14. L. M. Jesus, J. C. A. dos Santos, D. V. Sampaio, L. B. Barbosa, R. S. Silva and J.-C. M'Peko, J. Alloys Compd., 2016, 654, 482–490.
15. M. A. de la Rubia, J. J. Reinosa, P. Leret, J. J. Romero, J. de Frutos and J. F. Fernández, J. Eur. Ceram. Soc., 2012, 32, 71–76.
16. M. Cologna, B. Rashkova and R. Raj, J. Am. Ceram. Soc., 2010, 93, 3556–3559.
17. J.-C. M'Peko, J. S. C. Francis and R. Raj, J. Am. Ceram. Soc., 2013, 96, 3760–3767.
18. J.-C. M'Peko, J. S. C. Francis and R. Raj, J. Eur. Ceram. Soc., 2014, 34, 3655–3660.
19. W. Liu, K. M. Liang, Y. K. Zheng, S. R. Gu and H. Chen, J. Phys. D: Appl. Phys., 1997, 30, 3366–3370.
20. K. S. Naik, V. M. Sglavo and R. Raj, J. Eur. Ceram. Soc., 2014, 34, 4063–4067.
21. R. Raj, J. Eur. Ceram. Soc., 2012, 32, 2293–2301.
22. American Society for Testing and Materials – ASTM. E1382: Standard test methods for determining average grain size using semiautomatic and automatic image analysis. ASTM; 1991. Annual Book of ASTM Standards, v. 03.01.

---