INTRODUCTION

In the past few decades, intensive research effort has focused on energy storage devices, such as hydrogen cycle systems, supercapacitors, and lithium-ion batteries [1]. These latter are thought to have the greatest potential for use in portable electronics, wireless devices, and transportation applications owing to their high energy capacity and good cyclability. At the same time, the most widespread cathodes based on oxide materials (LiCoO2, LiMn2O4, and LiNixMnyCo1 –x–yOδ) have a variety of drawbacks, such as high cost, risk of explosion, and toxicity. The lithium iron phosphate LiFePO4 with the olivine structure is thought to be a possible alternative to the oxide materials, which is primarily due to its safety during cycling, low cost, and sufficiently high theoretical capacity (170 mAh/g) [2, 3]. Unfortunately, its relatively low ionic and electronic conductivities significantly limit its potential practical application areas [3, 4].

Among the widespread approaches for modifying electrode materials, which make it possible to overcome the above limitations, it is worthwhile to note heterovalent doping [5, 6]; the use of nanomaterials, including those with various morphologies [7, 8]; and the fabrication of composites, including those containing highly conductive additives [4, 9–11]. The last approach offers a comparatively simple and effective tool for raising the electrical conductivity of many solid electrolytes [12] and improving the cycling rate stability of lithium-ion battery electrodes [13–15]. Coatings are typically produced using carbon black and other carbon materials, including carbon nanotubes and graphene [1, 16–18]. At the same time, conductive polymers (polyaniline, polythiophene, poly(3,4-ethylenedioxythiophene) (PEDOT), and polypyrrole) are potentially attractive coating materials owing to their high electronic conductivity, electrochemical stability, and good mechanical properties such as flexibility and strength [19, 20]. As pointed out in a number of reports [21–24], they can be electrochemically active as well, raising the electrochemical capacity of composites. Previous reports described composites of lithium iron phosphate with polypyrrole [25], polyaniline [26], polythiophene [27], and PEDOT [28–30]. The use of the last material appears to be particularly attractive owing to its high electronic conductivity and the stability of composites containing it [29, 30].

To produce lithium iron phosphate/PEDOT composites, use is made of in situ chemical polymerization [30] and dynamic electropolymerization of 3,4-ethylenedioxythiophene (EDOT) on the surface of LiFePO4 particles and mechanical mixing of LiFePO4 with presynthesized PEDOT particles [31].

The electronic conductivity of some conductive polymers, including polyaniline, is known to rise appreciably on doping with acids, in particular with polymers having acid functional groups [32]. A similar effect was reported for a PEDOT/poly(styrene sulfonate) (PEDOT/PSS) composite material [30, 33]. The physicochemical properties of PEDOT, primarily its thermal stability and electrical conductivity, were improved by producing polymer films in the presence of surfactants [34, 35]. At the same time, the synthesis of LiFePO4/PEDOT composites via in situ oxidative polymerization in an acid medium may be accompanied by partial delithiation of the electrode material [36], leading to the formation of a considerable amount of impurities and making the composition of the material difficult to control. It is reasonable to expect that, owing to the high hydrophobicity of carbon, the use of carbon-precoated LiFePO4 can hinder delithiation of the material and oxidation of iron(II) ions on the surface of the particles during the in situ EDOT polymerization process. A similar effect can be produced by the use of surfactants, primarily nonionic ones, in the synthesis of composites. This approach would be expected to ensure not only “protection” of LiFePO4 from delithiation but also higher electrical conductivity and, accordingly, better electrochemical properties of the resultant lithium iron phosphate/PEDOT composites.

Given the above, the purpose of this research is to compare the electrochemical properties of LiFePO4/C/poly(3,4-ethylenedioxythiophene) composites prepared by both in situ oxidative EDOT polymerization and direct mechanical mixing of LiFePO4/C with presynthesized PEDOT particles, in particular in the presence of different surfactants.

EXPERIMENTAL

To synthesize LiFePO4 (LFP), an aqueous solution containing a stoichiometric ratio of iron(III) nitrate (Sigma-Aldrich, >98%), lithium nitrate (Sigma-Aldrich, >99%), and ammonium dihydrogen phosphate (Sigma-Aldrich, >98%) was held at 70°C with constant stirring until homogeneous suspension was obtained. Next, the resultant mixture was heat-treated at 300°C for 6 h. To obtain LFP/carbon (LFP/C) composite materials, the resultant powder was ground with 25 wt % saccharose and annealed at 600°C for 10 h in an argon atmosphere.

3,4-Ethylenedioxythiophene (Sigma-Aldrich, 97%) was polymerized in acetonitrile (Khimmed, reagent grade) at 25°C with constant stirring, using anhydrous FeCl3 (Fisher Chemical, extrapure grade) as an oxidant. To assess the effect of the amount of the oxidant on the properties of the resultant polymer, we varied the EDOT : FeCl3 molar ratio. The PEDOT samples obtained at EDOT : FeCl3 ratios from 0.20 to 0.25 had higher electrical conductivity. Because of this, in our subsequent preparations of poly(3,4-ethylenedioxythiophene), an EDOT : FeCl3 molar ratio of 0.22 was used. The PEDOT precipitate was washed with water and dried under vacuum at 30°C.

LiFePO4/C/poly(3,4-ethylenedioxythiophene) composites were prepared by both in situ EDOT polymerization and direct mechanical mixing of LiFePO4/C with presynthesized PEDOT particles. In the former case, EDOT was polymerized in the presence of LiFePO4/C as described above. The resultant composite materials will hereafter be referred to as LFP/C_5PEDOT_is. Here and in what follows, the numbers in the sample designations represent the weight percentage of the monomer (EDOT) relative to LFP/C. In addition, a composite of LiFePO4 and PEDOT was prepared by in situ polymerization using LFP powder precoated with the Triton X-100 (Sigma-Aldrich) surfactant at an LFP : Triton X-100 molar ratio of 1 : 10.8 (LFP_Tr_5PEDOT_is).

An LFP/C_7PEDOT composite was obtained by adding a suspension of presynthesized poly(3,4-ethylenedioxythiophene) in acetonitrile. The suspension was prepared using anhydrous FeCl3 and a 0.08 M EDOT solution in acetonitrile (EDOT : FeCl3 molar ratio of 1 : 5). The amount of PEDOT suspension added was such that the weight fraction of EDOT relative to LFP/C was 7%. In a similar way, we prepared LFP/C_Tr_7PEDOT and LFP/C_CTAB_7PEDOT composites, using LFP/C powders precoated with the Triton X-100 and cetyltrimethylammonium bromide (CTAB) surfactants, respectively (LFP : surfactant molar ratio of 1 : 10.8). In the case of the LFP/C_7PEDOT_Tr and LFP/C_7PEDOT_CTAB composites, poly(3,4-ethylenedioxythiophene) was synthesized in the presence of Triton X-100 and CTAB, respectively (EDOT : surfactant molar ratio of 7 : 1). In all cases, after the addition of the PEDOT suspension to the LiFePO4/C powder, the mixture was stirred for 10 min, washed with water and ethanol, and dried at 30°C in air.

In addition, we prepared an LFP/C_5PEDOTdry composite by adding 5 wt % presynthesized and dried PEDOT to LiFePO4/C, followed by grinding in an agate mortar.

The phase composition of the synthesized materials was determined by X-ray diffraction on a Rigaku D/MAX 2200 diffractometer (CuKα radiation) using the Rigaku Application Data Processing software package. The particle (crystallite) size was assessed from X-ray diffraction line broadening using the Debye–Scherrer formula and a LaB6 standard.

The synthesized composites were analyzed for carbon using a EuroVector EA3000 CHNS-O elemental analyzer. The carbon content of the LFP/C powder was determined to be 6.1%. The PEDOT content of the lithium iron phosphate/poly(3,4-ethylenedioxythiophene) composites was evaluated from the difference in carbon content between the synthesized samples and the starting LFP/C composite (Table 1). It is worth noting that the composites prepared using the surfactants could contain an amount of CTAB or Triton X-100.

Table 1.   Calculated content of poly(3,4-ethylenedioxythiophene) in the composites according to CHNS analysis data
Full size table

The microstructure of the synthesized materials was examined by scanning electron microscopy (SEM) on a Carl Zeiss NVision 40.

IR spectra were measured on a Nicolet iS5 Fourier transform IR spectrometer in attenuated total reflection mode (Specac Quest module, diamond crystal).

Electronic conductivity was measured at dc at 25°C with a Z500 PRO impedance meter (Elins, Russia) using samples in the form of pressed cylindrical pellets with silver contacts.

To prepare electrode paste, one of the synthesized composites (88%) was mixed with carbon black (10%), using poly(vinylidene fluoride) (2%) dissolved in N-methylpyrrolidone as a binder. The resultant paste was applied to a stainless steel grid (so that its surface density was 10–15 mg/cm2), pressed at 0.1 GPa, and then dried at 120°C under vacuum for 10 h.

Electrochemical tests were performed in hermetically sealed three-electrode cells, with a metallic lithium auxiliary electrode and reference. The cells were set up in a glove box under a dry argon atmosphere, using nonwoven polypropylene as a separator. The electrolyte used was a 1 M LiPF6 solution in a 1 : 1 : 1 mixture of ethylene carbonate, diethyl carbonate, and dimethyl carbonate. The cells were electrochemically cycled in galvanostatic mode at potentials in the range from 2.5 to 4.1 V and current densities from 20 to 1600 mA/g using a ZRU 50 mA–10 V charge–discharge system. Specific capacity was calculated per unit mass of the lithium iron phosphate.

RESULTS AND DISCUSSION

Table 2 lists the main absorption bands in the frequency range 600–1800 cm–1 (region of characteristic bands of heterocycles) in the IR spectrum of the sample prepared by the oxidative polymerization of 3,4-ethylenedioxythiophene. These data agree with previously reported results and point to PEDOT formation. Note that, according to Kvarnström et al. [38], the rather high intensity of the bands in the range from 1557 to 1517 cm–1 can be interpreted as evidence that the synthesized PEDOT is in a state characterized by high electrical conductivity. The conductivity of the synthesized bulk poly(3,4-ethylenedioxythiophene) is 0.12 S/cm, which is comparable to the conductivity of a polymer prepared by Kvarnström et al. [38] by a similar process but lower than that reported for thin PEDOT films by Brooke et al. [40].

Table 2.   Bands in the IR spectrum of the poly(3,4-ethylenedioxythiophene) prepared by oxidative EDOT polymerization
Full size table

The IR spectrum of the LFP/C_PEDOT_is composite provides little information because of the small percentage of poly(3,4-ethylenedioxythiophene) and, accordingly, the low intensity of its bands compared to the strong band due to the PO4 groups of the lithium iron phosphate.

All reflections in the X-ray diffraction pattern of the starting LFP/C composite correspond to the orthorhombic phase of LiFePO4 (PDF-2 card no. 40-1499) (Fig. 1). The crystallite (coherent scattering domain) size of the lithium iron phosphate is about 50 nm. The X-ray diffraction patterns of the PEDOT-containing composites prepared by mixing LFP/C with the poly(3,4-ethylenedioxythiophene) suspension contain, in addition, reflections from FePO4 (PDF-2 card no. 70-6685), which was probably formed as a result of the oxidation of the LFP by the oxidant (FeCl3) that was present in the PEDOT suspension added. In the case of the PEDOT-containing composites prepared using the surfactants, the percentage of FePO4 impurities was slightly higher. At the same time, the iron(III) phosphate content of the LFP/C_Tr_7PEDOT composite (prepared using the LFP/C powder precoated with Triton X-100) was somewhat lower (Fig. 1). It should be noted that the formation of a small amount of FePO4 does not reduce the electrochemical activity of the material, because during cycling it is lithiated in the discharge process, thereby maintaining the capacity of the sample. It is also worth noting that the presence of a hydrophobic carbon shell on the LiFePO4 particles and the use of the nonionic surfactant in poly(3,4-ethylenedioxythiophene) synthesis in the presence of lithium iron phosphate helps the latter retain its crystallinity. Indeed, in the case of in situ PEDOT synthesis on the surface of LiFePO4 (without a carbon shell), the latter amorphizes, whereas the LFP_Tr_5PEDOT_is remains crystalline (Fig. 1).

Fig. 1.
figure 1

X-ray diffraction patterns of the composites based on lithium iron phosphate and PEDOT: (1) LFP/C, (2) LFP_Tr_5PEDOT_is, (3) LFP/C_Tr_7PEDOT, (4) LFP/C_5PEDOT, and (5) LFP/C_5PEDOT_is.

Full size image

According to the electron microscopy results, the LFP/C material had a rather large surface area (Figs. 2a, 2b). Micrographs of the starting LFP/C composite show LFP particles up to 200 nm in size, which considerably exceeds the calculated crystallite size. This leads us to conclude that the LFP/C sample under consideration contains agglomerates of individual crystallites. The formation of a PEDOT-containing composite causes no significant changes in the morphology of the material (Fig. 2c).

Fig. 2.
figure 2

(a–c) Secondary electron and (d) backscattered electron SEM images of (a, b) the LFP/C composite and (c, d) the LFP/C_5PEDOTdry composite, prepared by chemical mixing with presynthesized PEDOT.

Full size image

At current densities in the range 20–1600 mA/g, the materials under study were subjected to galvanostatic discharge, which led to electrochemical lithium deintercalation. Next, we carried out cycling (charge–discharge, which corresponded to lithium extraction and insertion processes) at a constant current. The discharge capacity of the materials based on lithium iron phosphate with no carbon shell did not exceed 20–30 mAh/g at a current density of 20 mA/g, so they were not cycled at high charge–discharge rates.

Figure 3 shows the variation in discharge capacity during cycling at different current densities for the LFP/C- and PEDOT-based composites prepared in the presence of the surfactants. At current densities of ≤100 mA/g, the capacity of the LFP/C_7PEDOT_Tr composite, in which the PEDOT was prepared in the presence of Triton X-100, approaches that of the starting LFP/C composite. At the same time, further increasing the cycling rate leads to a sharp drop in the capacity of the sample (Fig. 3). The use of the cationic surfactant (cetyltrimethylammonium bromide) in PEDOT synthesis leads to a slight increase in the electrochemical capacity of the LFP/C_7PEDOT_CTAB composite in comparison with the LFP/C material, but only at low current densities (≤100 mA/g). In the rest of the range of cycling rates studied, the capacity of the synthesized material was comparable to or slightly lower than that of the starting material. Thus, the use of the cationic surfactant allows better results to be obtained in comparison with Triton X-100. It seems likely that the morphology of the poly(3,4-ethylenedioxythiophene) prepared in the presence of CTAB can favor a more uniform surfactant distribution over the LFP/C_7PEDOT_CTAB composite. In particular, data reported by Li et al. [34] suggest that, at a low cetyltrimethylammonium bromide concentration in aqueous solution (EDOT : CTAB molar ratio of 10 : 1), monomers form rodlike micelles, with only the two terminal groups of each micelle involved in polymerization. Thus, the polymer grows only at the ends of the micelles, forming nanorods. Increasing the amount of CTAB (EDOT : CTAB molar ratio of 3.3 : 1) leads to the formation of platelike micelles, which stack into blocks, forming plates. In our case, the EDOT : CTAB molar ratio is 7 : 1, which is most likely responsible for the formation of poly(3,4-ethylenedioxythiophene) in the form of nanorods. In the case of Triton X-100, a nonionic surfactant, weaker repulsion between EDOT and polymer particles may be observed, leading to stronger aggregation of the polymer particles and, accordingly, a less uniform distribution of the polymer over the LiFePO4-containing composite.

Fig. 3.
figure 3

Variation in discharge capacity during cycling of the (1) LFP/C, (2) LFP/C_7PEDOT-Tr, (3) LFP/C_7PEDOT_CTAB, (4) LFP/C_CTAB_7PEDOT, and (5) LFP/C_Tr_7PEDOT, prepared with the use of Triton X‑100 and CTAB by mechanical mixing of LiFePO4/C with PEDOT. The numbers at the data points specify current density (mA/g).

Full size image

A different situation occurs if a surfactant is preapplied to LiFePO4/carbon composite particles (Fig. 3). For example, in the case of CTAB the electrochemical capacity of the LFP/C_CTAB_7PEDOT composite approaches that of LFP/C except at a current density of 20 mA/g, where the capacity of the PEDOT-containing composite is considerably higher. Special attention should be paid to the fact that preprocessing of the lithium iron phosphate with Triton X-100, a nonionic surfactant, leads to a considerable increase in the capacity of the synthesized composite relative to the starting LFP/C material over the entire range of cycling rates studied, including high current densities. The observed effect is most likely due to the better uniformity of the nonionic surfactant coating on lithium iron phosphate/carbon composite particles owing to the hydrophobic carbon shell, which, accordingly, leads to a more uniform distribution of presynthesized PEDOT particles over the LFP/C surface. Besides, the nonionic surfactant can improve π–π stacking interactions in poly(3,4-ethylenedioxythiophene) [41], which would also lead to better PEDOT adhesion to LFP/C composite particles and the formation of a uniform, highly conductive coating.

Comparison of the cycling results for the composites prepared using Triton X-100 as a surfactant and for the starting materials (Fig. 4) clearly demonstrates the advantages of the composites. The potential difference between the charge and discharge curves of a given sample is substantially smaller, suggesting that, in this case, the resistance of the samples is lower [42].

Fig. 4.
figure 4

(1, 3) Charge and (2, 4) discharge curves of the (1, 2) LFP/C and (3, 4) LFP/C_Tr_7PEDOT composites in the fifth cycle at a current density of 20 mA/g.

Full size image

Based on the above data, only Triton X-100, a nonionic surfactant, was used in subsequent investigation to assess the effect of the process used to prepare the composites of LiFePO4/C and poly(3,4-ethylenedioxythiophene) (in situ oxidative polymerization of EDOT or mechanical mixing of LiFePO4/C with presynthesized PEDOT particles). At a low discharge rate, all of the samples except LFP/C_Tr_5PEDOT_is have high capacity compared to the starting LFP/C composite (Fig. 5). In particular, at a current density of 20 mA/g the discharge capacity of the synthesized samples increases in the order LFP/C_Tr_5PEDOT_is < LFP/C ≈ LFP/C_5PEDOT ≈ LFP/C_5PEDOT_is ≈ LFP/C_7PEDOT_Tr ≈ LFP/C_7PEDOT_CTAB < LFP/C_CTAB_7PEDOT < LFP/C_Tr_7PEDOT ≈ LFP/C_5PEDOTdry < LFP/C_Tr_7PEDOT_is. It is worth noting that, at high current densities, the capacity of the LFP/C_Tr_5PEDOT_is sample is comparable to or even exceeds that of the starting LFP/C material. At the same time, the capacity of the LFP/C_5PEDOTdry sample, prepared by grinding LFP/C with 5% presynthesized and dried PEDOT, decreases considerably faster with increasing charge–discharge rate than does the capacity of the other samples. It seems likely that the observed effect is due to the considerable poly(3,4-ethylenedioxythiophene) agglomerate size in the LFP/C_5PEDOTdry sample, which hinders lithium ion migration at high current densities (≥400 mA/g). This feature of the material is well illustrated by backscattered electron images. Dark areas in such images correspond to poly(3,4-ethylenedioxythiophene) (Fig. 2d), with the agglomerate size ranging from 1 to several microns.

Fig. 5.
figure 5

Variation in discharge capacity during cycling of the (1) LFP/C, (2) LFP/C_5PEDOTdry, (3) LFP/C_5PEDOT, (4) LFP/C_5PEDOT_is, (5) LFP/C_Tr_5PEDOT_is, (6) LFP/C_Tr_7PEDOT_is, and (7) LFP/C_Tr_7PEDOT, prepared with the use of Triton X‑100 as a surfactant. The numbers at the data points specify current density (mA/g).

Full size image

At high charge–discharge rates, an increase in capacity relative to LFP/C is only significant for the LFP/C_5PEDOT, LFP/C_Tr_5PEDOT_is, LFP/C_Tr_7PEDOT_is, and LFP/C_Tr_7PEDOT materials. Note that, in the case of the last composite, the capacity gain is somewhat higher, which is most likely due to the more uniform distribution of the PEDOT particles over the surfactant (Triton X-100) precoated LFP/C surface.

Comparison of the behavior of the materials at high current densities leads us to conclude that PEDOT synthesis without LFP/C is preferable because the capacity of LFP/C_5PEDOT exceeds that of LFP/C_5PEDOT_is. In the case of the LFP/C_5PEDOT_is and LFP/C_Tr_5PEDOT_is pair, the application of Triton X-100 to LFP/C also leads to an increase in discharge capacity, which can be accounted for in terms of hindered PEDOT particle agglomeration in the presence of the large amount of the nonionic surfactant and a better distribution of the polymer over the LFP/C surface. It is worth noting that, throughout the range of current densities studied, the discharge capacity of LFP/C_Tr_7PEDOT_is exceeds that of LFP/C_Tr_5PEDOT_is. Therefore, the increased percentage of the polymer has an advantageous effect on the electrochemical behavior of the composites. Besides, there are data in the literature that demonstrate that residual Triton X-100 improves the mechanical properties of PEDOT [41] and reduces the loss of the polymer during washing. At the same time, the preparation of PEDOT by in situ polymerization has no special advantages, and higher capacities at high current densities are offered by the material prepared by mechanical mixing of the LFP/C composite with PEDOT.

CONCLUSIONS

We have performed a comparative study of the electrochemical properties of LiFePO4/C- and poly(3,4-ethylenedioxythiophene)-based composites prepared by different methods: in situ oxidative EDOT polymerization in the presence of LiFePO4/C and mechanical mixing of LiFePO4/C with presynthesized PEDOT particles.

The results demonstrate that PEDOT synthesis in the presence of different surfactants (Triton X-100 and cetyltrimethylammonium bromide) causes no noticeable increase in the discharge capacity of the PEDOT-based composites. The composites prepared using prewashed and dried PEDOT particles have the best electrochemical performance at low charge–discharge rates, but their electrochemical capacity drops with increasing cycling rate because of the nonuniform PEDOT distribution over the material. In situ EDOT polymerization in the presence of LFP/C has no special advantages in terms of electrochemical cycling, and higher capacities at high current densities are offered by the material prepared by mixing the LFP/C composite with a suspension of presynthesized poly(3,4-ethylenedioxythiophene) in acetonitrile. Precoating of LiFePO4/C composite particles with Triton X-100 has been shown to ensure a more uniform PEDOT distribution over the material and an increase in discharge capacity relative to the starting LiFePO4/C composite over the entire range of charge–discharge rates studied.