A. Calcination/crystallization atmosphere effect

Samples were prepared using drying condition (A) summarized in Table I, which is the standard procedure for the production of ceramic powders by the organic steric entrapment method. ^{Reference Gülgün, Nguyen and Kriven16–Reference Gülgün, Kriven and Nguyen20} The effect of atmosphere was tested for samples with two different cation/organic ratios of 4:1 and 2:1.

Figure 2 and Fig. S1 are x-ray diffractograms for the samples made in air and forming gas at 700 and 800 °C using cation/organic ratios of 4:1 and 2:1, respectively. The black spectrum in both figures is the LiFePO₄ commercial powder from MTI Corp., used as a reference. These figures clearly show that by working in air it is not possible to obtain a single phase; rather a powder with more than five phases, including binary and ternary compounds [Li₃Fe₂(PO₄)₃, Fe₂O₃, Li₃PO₄], were obtained. However, by working in forming gas, powders with less than three phases mostly consisting of LiFePO₄ could be obtained. These results demonstrate that a more reduced atmosphere is required to avoid formation of oxidized secondary and ternary phases. If oxidized binary or ternary species are formed during any step of the synthesis process, it is difficult and requires further consumption of energy to remove them in subsequent stages of the synthesis, because usually they are thermodynamically stable. It would take high energy (time, temperature, and multiple grinding steps) to transform all of those phases to the desired LiFePO₄ single phase. Based on these results, forming gas (95% N₂ and 5% H₂) was selected as a calcination/crystallization atmosphere and as a standard condition to evaluate the other variables in the model system.

FIG. 2.

XRD patterns for samples made with a cation/organic ratio 4:1 in air and forming gas at 700 and 800 °C.

B. Solution pH effect

The advantage of producing materials using a wet chemical process is the ability to form a solution that has all the components very well dissolved and homogeneously distributed in the entire volume. In other words, it is to have an ionic mixing between all of the components. When all elements in the mixture dissolve, it is a solution and it is transparent; however, when the elements do not dissolve, the mixture is a dispersion and it is cloudy or milky. The best scenario is when a clear solution is formed because an ionic mixing of elements and homogeneity in the entire volume can be reached. When agglomerates, flocculates, or precipitates are formed, segregation of the components, heterogeneities, and multiphases in the final product can be expected. The ionic characteristics of each component and the pH of the solution can cause gelation, agglomeration, flocculation, or precipitation. As it was shown in the Table SI, the pH of the Fe source is strongly acidic (Fe nitrate, pH ≈ 0) and for the P source is approximately neutral (pH = 7.88); this wide pH difference can cause coagulation.

When solutions containing Fe and Li sources were mixed, a clear solution could be formed. However, when the solution containing the P source was added to the former solution, an agglomeration of components was observed [Fig. S2(a)]. It was noticed that by keeping the pH of the solution below 0.5, the agglomerates could be dissolved, and a clear solution formed [Fig. S2(b)], while by having the pH above 0.5, the agglomeration was irreversible, and a cloudy dispersion formed [Fig. S2(c)]. If agglomeration occurred, additional stirring homogenized the solution, but a clear solution was never reached. Instead, a cloudy and milky dispersion was formed. As was mentioned before, a transparent solution, where an ionic mixing is guaranteed, is crucial for single-phase formation. Therefore, nitric acid (HNO₃) was added dropwise, while the solution containing P source was added to the solution containing Li and Fe sources. HNO₃ was selected because it would not affect the stoichiometry of the system and would burn out during the calcination process. Based on these results, mixing the reagents at pH below 0.5 was established as a standard condition to produce a clear solution that guaranteed the ionic mixing of all components in the system.

C. Solution drying effect

Samples dried using conditions (I) and (IV), shown in Table I, were examined in a simultaneous DSC/TGA apparatus to evaluate endothermic and exothermic processes, such as pyrolysis, decomposition, and crystallization of precursors. These two extreme conditions were selected because in (I) the organics and nitrates released during drying were abundant, and this procedure has been the standard procedure used thus far in the production of ceramic powders by the organic steric entrapment method. ^{Reference Gülgün, Nguyen and Kriven16–Reference Gülgün, Kriven and Nguyen20} Moreover, in samples dried using condition (IV), the release of organics or nitrates was barely detected, which allowed the full availability of these two components (fuel and oxidizer) for the next steps (calcination/crystallization).

Figure 3 shows TGA and DSC curves of the dried samples Type (I) and Type (IV). Curves A and B correspond to Type (I) dried samples made with the cation/organic ratios 4:1 and 2:1, while curves C, D, and E correspond to Type (IV) dried samples made with the cation/organic ratios 2:1, 4:1, and 6:1. For dried samples of Type (I), thermal analyses indicated that there was an exothermic peak at a temperature around 350 °C. This peak could be associated with the initiation of the crystallization process. The total weight percent loss was about 11 and 26% for the 4:1 and 2:1 samples, respectively. These values agreed with differences in the initial amount of organics in both samples. It is also important to remember that these samples were dried on a hot plate at ≈ 350 °C and then calcined in air for 2 h at 250 °C (10 °C/min) before performing DSC/TGA analyses. Orange and white gases were observed during drying, and a reddish dried powder was obtained. For dried samples Type (IV), the total weight percent lost was about 78, 83, and 40% for 2:1, 4:1, and 6:1 samples, respectively. These total weight-loss percentages were larger [compared with those of samples Type (I)] because the samples were dried at a low temperature to remove only the water content. All the organics and nitrates remained in the sample and could be used in subsequent steps in the synthesis. In addition, the maximum weight loss could be observed at the same temperature where the large exothermic peak in the DSC curve was seen. These large weight losses can be attributed not only to the release of gases of organics and nitrates but also, to some sample losses from the sample container, due to the highly exothermic reaction taking place during the analysis. These intense exothermic reactions were observed not only during the thermal analysis but also during the powder production in the laboratory. In addition, the TGA curves indicated that full calcination occurred at low temperature, especially for sample 4:1 at 150 °C. The precise nature of the decomposition mechanism and released gases is beyond the scope of this paper.

FIG. 3.

(a) TGA and (b) DSC analysis of dried samples Type (I) made with 2:1 and 4:1 organic cation ratio (curves A and B, respectively) and dried samples Type (IV) made with the cation/organic ratios 2:1, 4:1, and 6:1 (curves C, D, and E, respectively).

Differential scanning calorimetric curves revealed two different exothermic peaks for samples made using ratios 2:1 and 6:1. These two peaks could be attributed to organics and nitrates decomposing at different temperatures. For the sample with ratio 2:1 (curve C), the decompositions were observed at 175 and 225 °C, whereas for the 6:1 sample (curve E), the decomposition reactions took place at 220 and 275 °C. In the tube furnace (for sample 2:1), a strong and long decomposition reaction was also observed, where yellow/white gases were released. Moreover, for sample 6:1, the decomposition reaction was short and not strong as the previous reactions, and the released gases were mostly white in color.

In the case of the sample with ratio 4:1 (curve D), a single large exothermic peak was sensed in the DSC analysis at 150 °C, which meant that not only did the organics and nitrates decompose at the same time but also the crystallization started at a very low temperature. All the energy available from these decomposition reactions was released at the same time and could be used to induce crystallization. This exothermic peak perfectly matched the large weight loss observed in the TGA curve. This decomposition reaction was also observed in the tube furnace; it was very violent and long, and the released gases were red/yellow in color. From the DSC analysis, it can be seen that there was an optimum ratio between organics and nitrates to produce a strong and single exothermic decomposition reaction. For this model system, the optimum ratio was 4:1. This was the best scenario to generate the maximum amount of energy in the system, which could be used to induce crystallization. Movements to a higher or lower ratio from this optimum shifted the decomposition reaction as well as the crystallization onset toward higher temperatures.

D. Crystallization

Figure 2 shows that at 700 °C, small amounts of secondary phases were still present in the sample made using drying and calcination/crystallization conditions (II) listed in Table I. However, upon increasing the temperature to 800 °C, those secondary phases recombined through a solid-state reaction to form a single phase. It is important to mention here that unless an ionic mixing was guaranteed and sufficient amount of polymer was added to the system to ensure steric entrapment and avoid segregation, formation of single phase at low temperature was unsuccessful. High temperature (800 °C) and solid-state diffusion would be required to form pure LiFePO₄ phase.

In contrast, by taking advantage of the strong, energetic, and highly exothermic decomposition reactions observed in the dried samples of Type (IV), crystallization at lower temperatures was possible. As it was seen in the thermal analysis (Fig. 3), the temperature, magnitude, and characteristics of the decomposition reaction were dependent on the ratio of organics/nitrates (i.e., fuel/oxidizer). Figures S3 and S4 and Fig. 4 show the phase evolution when temperature was increased (300–800 °C) for the dried samples Type (IV) with different cation/organic ratios of 2:1, 4:1, and 6:1. All samples were compared not only with the ICDD-PDF cards but also with commercial LiFePO₄ from MTI Corporation.

FIG. 4.

X-ray diffractograms of dried samples Type (IV) with 4:1 cation/organic ratio in the range 300–800 °C. Single-phase LiFePO₄ was obtained under all conditions.

Formation of a single-phase LiFePO₄ was possible at 500 °C using a 2:1 cation/organic ratio. Samples below 500 °C indicated formation of small amounts of crystalline phases, but an amorphous phase was still present. No secondary phases were detected in samples treated above 500 °C, which meant that ionic mixing and steric entrapment were successful. Here in this group of samples, the effect of the exothermic decomposition reaction on the crystallization process was evident. The energy released during the decomposition of organics and nitrates helped to induce crystallization of the single-phase LiFePO₄ at lower temperatures, as compared to other synthesis methods as well as to the methodology implemented for dried samples of Type (II).

For samples containing a 4:1 cation/organic ratio, the results were even more interesting. As can be seen in Fig. 4, the formation of a pure crystalline LiFePO₄ phase at 300 °C was successful. At 300 °C, peak broadening and less intense peaks than the reference sample suggested the formation of nanoparticles. At 800 °C, some extra peaks were detected, which corresponded to Fe₂P. These results agreed with the information obtained from the thermal analysis where a single strong exothermic peak was detected at 150 °C. Moreover, in the tube furnace, the decomposition reaction was detected at around 240 °C, and it seems that the energy released was enough to induce crystallization of the single phase. As no secondary phases were detected in the samples below 800 °C, it suggests that the ionic mixing and the amount of organics and nitrates were in the right proportion to produce the steric entrapment. Thereby, we avoided segregation and generated a single exothermic decomposition reaction that was strong enough to crystallize olivine-type LiFePO₄ at low temperatures.

Samples containing a 6:1 cation/organic ratio (Fig. S4) only produced amorphous powder below 500 °C; however, the onset of crystallization was noticed at around 500 °C. Secondary phases, especially binaries and oxidized species, such as Fe₂O₃, were detected. With increasing temperature and under reducing atmospheres, those secondary phases recombined through solid-state reaction to form a more pure powder. These results confirmed the information obtained in the thermal analysis where the exothermic peaks were detected at higher temperatures of 220 and 275 °C. This was observed in the tube furnace at around 290 °C. The fact that the decomposition of organics and nitrates took place at different temperatures led to the exothermic reaction being less strong (Fig. 3). This means that the energy released was insufficient to induce crystallization at a low temperature. The finding of secondary phases suggested that unless there was ionic mixing, the amount of organics was not sufficiently sterically entrapping the cations so as to avoid segregation. These results demonstrated that there is an optimum ratio (4:1) of organics/nitrates not only to affect the polymeric steric entrapment but also to generate a single strong exothermic decomposition reaction, which provides enough energy to induce crystallization of the desired phase at temperatures lower than those involved in conventional methods. Shifts from the optimum ratio provoked delays in the starting of the decomposition reaction because extra energy was required to decompose the extra amount of organics, and also because the amount of nitrates was not enough to react with the total amount of organics. This was the case for 2:1 ratios. For 6:1 ratios, there is also an imbalance of fuel/oxidizer, which made the exothermic reaction of organics and nitrates to occur at different temperatures and also produces a less energetic reaction. Furthermore and more importantly, the amount of organics was not enough to make an optimum steric entrapment that produced segregation of the components in the system and allowed formation of multiphase materials.

E. Microstructure characterization

The microstructure of LiFePO₄ produced from the dried sample type (II) at 800 °C is shown in the Figs. 5(a) and 5(b). Platelet-like crystals in the range between 0.5 to 10 μm were found in the samples. Evidence of crystal growth and necking (sintering) was found in this sample, which means that high temperature promoted mass diffusion toward the particle interfaces. This also proved that solid-state diffusion took place, and so that secondary phases recombined and disappeared at higher temperatures. Unless this sample was pure LiFePO₄ and the microstructure resembled the microstructure reported in the literature for triphylite, 800 °C was a high processing temperature, which was not only uneconomical for an industrial process but also produced large crystals and sintered powders, which affected the surface area.

FIG. 5.

SEM micrographs of pure LiFePO₄ formed using drying and calcination/crystallization (a) and (b) Type (II) and (c) and (d) Type (IV). (a) and (b): Platelet-like crystals in the range between 0.5 and 10 μm. (c) and (d) Porous secondary particles (soft agglomerates) formed from primary particles in the range of 20 nm–10 μm.

Moreover, the morphology of the powders produced by the polymeric steric entrapment method, and specially using dried sample Type (IV), was porous secondary particles (soft agglomerates) formed from primary particles in the range of 20 nm–10 μm [Figs. 5(c) and 5(d)]. These secondary particles or soft agglomerates resulted in this particular microstructure due to the intense exothermic decomposition reaction of organics and nitrates. These porous structures had a high SSA, which may be helpful for ion and electron diffusion in lithium ion batteries, specifically for LiFePO₄.

In general, for samples made at temperatures between 300 and 400 °C using optimized conditions and cation/organic ratios of 2:1 and 6:1 (Fig. S5), particles without any particular shape were found because of the lack of crystallization. Nanocrystals with platelet-like shapes and with a narrow distribution of sizes, in the range of 50–150 nm, could be observed in samples made at temperatures between 500 and 600 °C using a cation/organic ratio of 2:1. At higher temperatures (700–800 °C), crystals about 1 μm in size with platelet-like shapes and a narrow size distribution were produced for both conditions 2:1 and 6:1 cation/organic ratios. Additionally, crystal growth and necking between particles were noticed, which meant that solid-state diffusion started and presintering took place.

Figure 6 exhibits a pure crystalline LiFePO₄ powder obtained using optimized conditions and a cation/organic ratio of 4:1 at temperatures in the range 300–800 °C. At 300 °C [Fig. 6(a)], nanoparticles of roughly less than 100 nm were formed. These nanoparticle sizes were narrowly distributed and had a platelet-like shape. It can be seen that at 600 °C [Fig. 6(b)] and 800 °C [Fig. 6(c)], crystals still had sizes in the nanoscale range of about 100 nm. These results were very interesting because after increasing the furnace temperature up to 800 °C, there was not too much crystal growth or presintering. This effect was very important because it meant that small particle size and high-surface area can be kept even while working at high temperatures. It is believed that this behavior occurred because of the organic steric entrapment. The remnant carbon covered and coated the nanoparticles, thereby avoiding solid-state diffusion and hence crystal growth and sintering. It is important to mention that pure LiFePO₄ produced at low temperatures (300 °C) by the polymeric steric entrapment method was comparable with the LiFePO₄ produced by other wet chemical synthesis methods, such as the sol–gel and hydrothermal methods. ^{Reference Meligrana, Gerbaldi, Tuel, Bodoardo and Penazzi2,Reference Lin, Pan, Gao and Liu4} In contrast to these methods, however, expensive reagents (alkoxides), high-pressure or special-sealed containers, and large amounts of polymers were not used. That is why the most interesting sample in this study was the powder obtained using the optimized conditions and a cation/organic ratio of 4:1 at 300 °C: because it was pure crystalline LiFePO₄ produced at really low temperature, with particle sizes less than about 100 nm. The production conditions and the characteristics described suggested a material with high-surface area, carbon coating, and small diffusion paths for ions and electrons, which makes this a promising material for energy storage applications. TEM characterization [Fig. 6(d)] clearly shows agglomerates of about 1 μm formed by platelet-like nanoparticles of 20 nm. A diffraction pattern of the selected region in the micrograph (e) is shown in the micrograph (f), where clear periodic rings which correspond to a polycrystalline sample can be seen. The nanoparticle microstructures revealed in Fig. 6 were consistent with the peak broadening evident in the x-ray diffraction (XRD) analysis.

FIG. 6.

SEM and TEM micrographs of LiFePO₄ powder produced under optimized conditions and cation/organic ratio 4:1 at (a) 300 °C, (b) 600 °C, (c) 800 °C, (d) 20 nm length scale, (e) 0.5 μm length scale, (f) diffraction pattern of selected region, and (g) 50 nm length scale.

To conclude the characterization of LiFePO₄ and comparison with the reference, a particle size analysis, SSA, and chemical analysis were carried out under the same conditions. Table II summarizes the physical properties of the best samples of pure LiFePO₄ obtained in this work. Samples made at cation/organic ratio 2:1 at 400 °C are included because it was an amorphous precursor to produce pure LiFePO₄ using the polymeric steric entrapment method.

TABLE II.

Physical properties of the best samples of pure LiFePO₄ produced by the polymeric steric entrapment method.

It can be noticed that the values of the particle size analysis were very far away from the crystal sizes observed in the SEM and TEM micrographs. These results indicated that the particle size measured by a laser scattering particle size analyzer (Partica LA-950V2, Horiba, Kyoto, Japan), really measured the agglomerates, which were also observed in the low magnification SEM and TEM micrographs. What is important, however, is the fact that in comparing our results with the commercial references, both the crystal sizes and the agglomerated particles produced in this work were smaller than the reference powders. The SSA values (Table II) indicated that all powder produced in this study had twice the SSA of the MTI reference powder. These results were expected because of both the crystallite sizes (nanocrystals) and the sponge-like microstructure of the agglomerates observed in the SEM and TEM micrographs. Having powders with high surface area may be advantageous for electronic applications because it can improve the ionic and electronic conduction.

Chemical compositions of selected samples are shown in Table SII. These samples were chosen to see how the carbon content changed with temperature and cation/organic ratio. It was important to take into account that the nature of the exothermic reaction also affected the carbon and nitrogen content because both fuel and oxidizer were consumed during the reaction. The ICP chemical analysis had the disadvantage that the values were expressed in terms of weight percent (wt%), which takes into account the whole mass of the powder. Large variation could be expected in all elements measured because they were relative to the amount of other elements, such as C, N, H, and O, which were consumed or added during the synthesis process. For instance, multiphase samples containing oxidized phases had more oxygen in the whole mass, which affected the relative amount of other cations in the system.

Nevertheless, there was a tendency to retain about 10 and 1.5–2 wt% of carbon and nitrogen, respectively, in samples calcined/crystallized at low temperatures using cation/organic ratios of 2:1 and 4:1. Furthermore, the amounts of carbon and nitrogen decreased with increasing temperature. For Li ion battery applications, it has been demonstrated that carbon coating improves the electrical conductivity of the active particles, improving electrochemical performance. Therefore, the fact that, in this work, LiFePO₄/C composites have been produced in situ generates hopeful expectations for energy storage applications. Samples calcined/crystallized using cation/organic ratio 6:1 showed low contents of C and N. These results agreed with the low amount of organic (PVA) added during the synthesis.