Synthesis, characterization and electrochemical performance of mesoporous FePO4 as cathode material for rechargeable lithium batteries

Synthesis, characterization and electrochemical performance of mesoporous FePO4 as cathode material for rechargeable lithium batteries

Available online at Electrochimica Acta 53 (2008) 2665–2673 Synthesis, characterization and electrochemical performance of mes...

1MB Sizes 0 Downloads 73 Views

Available online at

Electrochimica Acta 53 (2008) 2665–2673

Synthesis, characterization and electrochemical performance of mesoporous FePO4 as cathode material for rechargeable lithium batteries Z.C. Shi 1 , A. Attia 2 , W.L. Ye, Q. Wang, Y.X. Li, Y. Yang ∗ Department of Chemistry and the State Key Lab for Physical Chemistry of Solid Surfaces, College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, PR China Received 1 April 2007; received in revised form 8 June 2007; accepted 25 June 2007 Available online 10 July 2007

Abstract Mesoporous FePO4 could deliver enhanced specific capacity of 160 mAh g−1 at first discharge process, 90% of theoretical capacity of pure FePO4 , and 135 mAh g−1 in the following cycles at 0.1 C rate. At 1 and 3 C rates, the capacities are 110 and 85 mAh g−1 , respectively, which is much higher than that of previously reported for modified FePO4 materials. Electrochemical impedance spectroscopy (EIS) tests proved that mesoporous structure in FePO4 materials enhanced the lithium ion intercalation/deintercalation kinetics as indicated by smaller charge transfer resistance (Rct ) of these materials. These results revealed that this mesoporous electrode material can be a potential candidate for high-power energy conversion devices. © 2007 Published by Elsevier Ltd. Keywords: Iron phosphate; Mesoporous materials; Cathode materials; Electrochemical performance; Energy conversion devices

1. Introduction In recent years, nanostructured electrode materials, such as mesoporous materials, have shown promising applications in energy conversion devices [1–5]. Fe-based cathode materials, such as Fe(III)PO4 and LiFe(II)PO4 , are attractive lithium intercalation electrode materials for their low cost, environmentally friendly and high theoretical specific capacity [6–12]. FePO4 shows a discharge process from 3.5 V down to 2.5 V and a theoretical specific capacity of 178 mAh g−1 upon 1 mol of lithium intercalation [6]. However, the practical specific capacity of FePO4 is quite low due to the poor kinetics of lithium intercalation/deintercalation process. To improve the electrochemical performance of FePO4 , Croce et al. added RuO2 to quartz FePO4 to enhance its specific capacity with a higher material interparticle electronic conductivity [8]. Both amorphous hydrated and ∗

Corresponding author. Fax: +86 592 218 5753. E-mail address: [email protected] (Y. Yang). 1 Current address: Department of Chemical and Materials Engineering, University of Alberta, Edmonton, Canada. 2 On leave from Department of Physical Chemistry, National Research Centre, El-Tahrir St., Dokki 12622, Cairo, Egypt. 0013-4686/$ – see front matter © 2007 Published by Elsevier Ltd. doi:10.1016/j.electacta.2007.06.079

anhydrous FePO4 were also found to have better electrochemical performance than that of pure crystalline FePO4 [9–11]. The specific capacity of FePO4 is still incomparable to its reported theoretical one. We first reported that importing mesoporous structure into FePO4 , by a surfactant (EO20 -PO70 -EO20 , Pluronic P123) self-assembly method resulted in enhancement of the electrochemical performance of FePO4 [2]. The preparation conditions and electrochemical performance of the mesoporous FePO4 were also primarily optimized. Afterwards, mesoporous FePO4 with smaller pore diameter acted as a good storage material especially as lithium ion intercalation cathode material [13]. By using cetyltrimethyl ammonium bromide (CTAB) as template, this mesoporous FePO4 had a mesostructure with a pore diameter of 3–4 nm as confirmed by small angle X-ray diffraction compared with the results of Guo et al. [14], where the pore diameters were not large enough for the fast transport of electrolyte and failed to improve the kinetics of lithium ion intercalation with a result of relative low capacity. In this work, systematic studies on synthesis, characterization and electrochemical performance of mesoporous FePO4 were conducted, aimed at improving FePO4 electrochemical performance.


Z.C. Shi et al. / Electrochimica Acta 53 (2008) 2665–2673

the samples were suspended in ethanol solution by the help of ultrasonic bath, then mounted on a carbon holder, and left to dry before examination.

2. Experimental 2.1. Samples preparation FePO4 samples were synthesized by a surfactant (EO20 PO70 -EO20 , Pluronic P123) self-assembly method similar to that reported in our previous work [2]. After solvent evaporation, precursor gels were firstly dried at 80 ◦ C for at least 10 h, and then calcinated in a muffle furnace in air for 10 h. We synthesized FePO4 samples with different P123/(Fe + P) molar ratio of 0.013, 0.026, 0.039, and 0.052 (named as 1P, 2P, 3P, and 4P) at different calcination temperatures of 400, 450, 500, 600, and 700 ◦ C. Therefore, we got three series of FePO4 samples named as follows—(1) 1P-FPO samples: 1P-400, 1P-450, 1P-500, 1P600, and 1P-700; (2) 3P-FPO samples: 3P-400, 3P-450, 3P-500, 3P-600, and 3P-700; (3) 450 ◦ C-FPO samples: 1P-450, 2P-450, 3P-450, and 4P-450. The residual carbon in FePO4 samples was analyzed by elemental analysis using an EA1110 (ThermQuest Italia S.P.A., Italy) instrument, the precision of the measurements was ±0.3%. X-ray diffraction (XRD) experiments were carried out by a Panalytical X-pert diffractometer (PANalytical, ˚ The Netherlands), using Cu K␣ radiation (λ = 1.54059 A). Fourier transform infrared (FTIR) spectra were recorded on an Avatar 360 spectrophotometer (Nicolet, USA) and the resolution of spectra collected at 2 cm−1 interval over the range measured of wavenumber. The mesoporous structure information (specific surface area and pore diameter distributions) of FePO4 samples were obtained from N2 sorption isotherm plot on TriStar3000 (Micromeritics, USA) based on Brunauer–Emmett–Teller (BET) and Barrett–Joyner–Halenda (BJH) equations, the error in determination of BET surface area and porosity was in the range of ±10%. The structure and morphology of FePO4 samples were further characterized by high resolution transmission electron microscopy (HRTEM) technique by using Tecnai F30 (Philips-FEI, Netherlands) where

2.2. Cell fabrication and testing The electrochemical performance of FePO4 samples were assessed using CR2025 coin cells. The preparation of the electrodes and the assembly of coin cells are the same as that reported in Ref. [2], in which the electrolyte was 1 M LiPF6 dissolved in EC + DMC (1:1, v/v) and all potentials mentioned in this work were recorded versus Li/Li+ electrode. Charge–discharge experiments were performed between 1.5 and 4.0 V at various current densities (0.1, 1, 3 and 0.1 C) consecutively using a LAND CT2001A Battery Cycler (Wuhan, China). Electrochemical impedance spectroscopy (EIS) experiments of these cycled coin cells were done using an impedance/gain-phase analyzer (Solartron SI 1260) combined with an electrochemical interface (Solartron SI 1287) at an equilibrium charge/discharge state. The impedance spectra were obtained by applying a 10 mV potential amplitude excitation over a frequency range from 106 to 10−3 Hz. Impedance data acquisition and analysis were performed by using the electrochemical impedance software ZPlot and Zview (version 2.90, Scribner Associates Inc., USA). 3. Results and discussion 3.1. Samples synthesis and phase analysis Table 1 shows the residual carbon in all synthesized samples and surface properties of the samples. Table 1 shows that the residual carbon is inversely proportional to the temperature used for calcination of samples, i.e., by increasing the temperature from 400 to 500 ◦ C, the amount of residual carbon decreased from 1.48 to 0.94 wt.% in 1P samples, and similiarly for 3P samples, the residual carbon decreased upon increasing the tem-

Table 1 Residual carbon and surface properties of FePO4 samples Properties

P123/(Fe + P) molar ratio

400 ◦ C

Residual carbon content (wt.%) in FePO4 samples

1P 2P 3P 4P

1.48 N/A 1.79 N/A

Specific surfaces area (m2 g−1 ) of FePO4 samples

1P 2P 3P 4P

101 N/A 108 N/A

Pore volume (cm3 g−1 ) of FePO4 samples

1P 2P 3P 4P

0.28 N/A 0.38 N/A

Pore diameter (nm) of FePO4 samples

1P 2P 3P 4P

4.8 N/A 7.5 N/A

N/A signifies that these samples were not tested.

450 ◦ C 1.30 1.67 1.74 1.72 87 108 113 127 0.26 0.27 0.40 0.50 5.0 6.2 9.0 10.0

500 ◦ C

600 ◦ C

700 ◦ C

0.94 N/A 1.38 N/A

0.09 N/A 0.09 N/A

0.06 N/A 0.06 N/A

61 N/A 99 N/A

2.7 N/A 1.9 N/A

1.4 N/A 1.7 N/A

0.20 N/A 0.37 N/A

0.02 N/A 0.01 N/A

0.01 N/A 0.01 N/A

7.0 N/A 11.0 N/A



Z.C. Shi et al. / Electrochimica Acta 53 (2008) 2665–2673

perature from 400 to 500 ◦ C from 1.79 to 1.38 wt.%. Low carbon content can be found in the samples synthesized at higher temperature of 600 and 700 ◦ C in our case. Accordingly, the specific capacities of FePO4 samples have been corrected considering the carbon content. Analogously, as mentioned in our previous work [2], strong peaks of trigonal FePO4 (ICDD 29-0715) can only be indexed in XRD pattern of the FePO4 samples synthesized at 600 and 700 ◦ C (shown in supporting information files, Figs. 1–3), which indicates that the crystallization of trigonal FePO4 occurred at the temperature between 500 and 600 ◦ C, however, below 600 ◦ C, no crystalline peaks detected and amorphous phases obtained instead [15]. This crystallization process proved to be an endothermal peak on the DTA curve (shown in supporting information files, Fig. 4). Some of phosphorous resources might be lost during the synthesis process because of the possibility of presence of some impurities such as Fe2 O3 and Fe3 PO7 in the crystalline sample, where Fe2 O3 is apt to form easily in most of the conditions including ours and Fe3 PO7 is detected in the XRD (shown in supporting information files, Figs. 1 and 2) at 600 and 700 ◦ C. It is well known that the vibrational motions of ionic crystal can be classified into internal and external optic modes. In FePO4 , the internal modes come from the intramolecular vibrations of the PO4 3− polyanion [16], and the external ones originate in the lattice vibrations. Fig. 1 shows the FTIR spectra of 1P-FPO samples in the 1500–400 cm−1 range. Here, we mainly discuss the internal modes of 1P-FPO samples. Similar to that of olivine FePO4 and other iron phosphate materials [15,17,18], FTIR spectrum (Fig. 1, curve e) of trigonal FePO4 (1P-FPO-700) exhibited a broad maximum between 900 and 1200 cm−1 which can be assigned to the P–O vibrations of PO4 3− polyanion. Band of 932 cm−1 came from the symmetric vibration of P–O bond and the other bands from the asymmetric vibrations of P–O bond. In the range of 500–700 cm−1 ,

Fig. 1. FTIR spectra of the 1P-FePO4 samples calcinated at (b), 500 ◦ C (c), 600 ◦ C (d), and 700 ◦ C (e) for 10 h.

400 ◦ C


450 ◦ C


two medium bands at 595 and 633 cm−1 , and several weak bands were all attributed to the asymmetric vibrations of O–P–O bond. Band at 432 cm−1 and its shoulder at the right side were attributed to the symmetric vibrations of O–P–O bond. In the case of amorphous FePO4 , the bonding between all kinds of atoms was only partially reserved, which resulted in poor symmetric structure. Therefore, IR inactive vibrations of trigonal FePO4 became active in amorphous FePO4 . Consequently, the two absorbed bands at about 1000 and 600 cm−1 in amorphous mesoporous FePO4 prepared at 400, 450 and 500 ◦ C (Fig. 1, curves a–c) were broadened. Moreover, the maximum frequency of the broad band shifted from 1027 to 1060 cm−1 , which indicates a new dominating vibration of P–O bond at 1060 cm−1 in the amorphous mesoporous FePO4 . The FTIR spectra of 3P-FPO samples showed the same result as that of 1P-FPO samples (shown in supporting information files, Fig. 5). 3.2. Mesoporous structure of FePO4 samples The nano-structured characteristics of mesoporous FePO4 are largely affected by preparation conditions. The FePO4 samples prepared at or below 500 ◦ C show typical type IV isotherm plots in the N2 sorption experiments (shown in supporting information files, Figs. 6–8), indicating the existence of mesoporous structure of the FePO4 samples. Figs. 2 and 3 show the pore diameter distribution plot of 3P-FPO samples and all FPO samples at 450 ◦ C, respectively. Table 1 shows also the specific surface area, pore volume and pore diameter of all FePO4 samples under different conditions. When the calcination temperature of 1P-FPO samples increased from 400 to 500 ◦ C, the specific surface area decreased from 101 to 61 m2 g−1 , and pore volume decreased from 0.28 to 0.20 cm3 g−1 . On the contrary, the pore diameter increased from 4.8 to 7.0 nm. Mesoporous information could not be obtained for 1P-FPO-600 and 1P-FPO-700 samples, and the specific surface area decreased significantly to 2.7 and 1.4 m2 g−1 , respectively. This indicated that the mesoporous structure collapsed dur-

Fig. 2. Pore diameter distribution plots of FePO4 samples calcinated at 400 ◦ C (), 450 ◦ C (), 500 ◦ C (), 600 ◦ C (), and 700 ◦ C (♦) for 10 h with P123/(Fe + P) = 0.039(3P).


Z.C. Shi et al. / Electrochimica Acta 53 (2008) 2665–2673

Fig. 3. Pore diameter distribution plots of FePO4 samples calcinated at 450 ◦ C for 10 h with P123/(Fe + P) = 0.013 (, 1P-450), 0.026 (, 2P-450), 0.039 (, 3P-450) and 0.052 (, 4P-450). The average pore diameters are 5.0, 6.2, 9.0, and 10.0 nm, respectively.

ing the calcination process in the temperature between 500 and 600 ◦ C, which accompanied with the formation of trigonal FePO4 . For 3P-FPO samples calcinated at temperature from 400 to 500 ◦ C, the specific surface areas decreased from 108 to 99 m2 g−1 , respectively; which is expected due to the anticipated crystal growth of the FePO4 particles upon temperature increase. The pore volumes were found to be 0.38 and 0.37 cm3 g−1 , while the pore diameters were 7.5 and 11.0 nm, respectively. Mesoporous information could not be obtained for samples 3PFPO-600 and 3P-FPO-700 (Fig. 2). It is obvious from the above data and data in Table 1 that, at the same P123/(Fe + P) molar ratio, the specific surface area and the pore volume decreased with calcination temperature, and the pore diameter distribution broadened at the same time. Moreover, samples prepared at 600 and 700 ◦ C did not show any mesoporous information because the mesoporous structure collapsed upon the crystallization of FePO4 and growth of the particles between 500 and 600 ◦ C. As for 450-FPO samples, with the P123/(Fe + P) molar ratio increased from 1P to 2P, 3P and 4P, the specific surface area increased from 87 to 108, 113 and 127 m2 g−1 , respectively,

Fig. 4. TEM images of mesoporous FePO4 samples calcinated at 450 ◦ C for 10 h with P123/(Fe + P) = 0.013 (a, 1P-450), 0.026 (b, 2P-450), 0.039 (c, 3P-450), and 0.052 (d, 4P-450).

Z.C. Shi et al. / Electrochimica Acta 53 (2008) 2665–2673

while the pore volume from 0.26 to 0.27, 0.40 and 0.50 cm3 g−1 , respectively and the pore diameter from 5.0 to 6.2, 9.0 and 10.0 nm, respectively as shown in Table 1. Therefore, under the same calcination condition of temperature and time, the specific surface area, pore volume and pore diameter were enlarged with the increase of the surfactant ratio, although the pore diameter distribution got broadened and the uniformity of pore structure decreased (Fig. 3). Fig. 4 shows the HRTEM images of mesoporous FePO4 samples prepared at 450 ◦ C with different four P123/(Fe + P) ratios. The particle sizes were found to be in the range of 100–200 nm and meso-structure can be observed in all of the four samples, however, the mesoporosity was not of organized structure as indicated by HRTEM (Fig. 4). 1P-450 and 2P-450 samples synthesized under smaller P123/(Fe + P) molar ratio, showed smaller pore diameters than the other two samples of 3P-450 and 4P-450 of higher P123/(Fe + P) molar ratio and they showed more homogeneous pore distribution. However, HRTEM image of sample 4P-450 revealed destructed pore structure. For the highest P123/(Fe + P) molar ratio, it seemed difficult for inorganic compounds in the pore wall to support the mesostructure with large pore diameter. 3.3. Electrochemical performance of mesoporous FePO4 cathodes In our electrochemical tests, the typical voltage profile of all FePO4 samples, with an average discharge voltage at 3.0 V (shown in supporting information files, Figs. 9–11), is similar to that results of reported FePO4 materials [8–11]. Figs. 5 and 6 show the specific discharge capacity upon cycling of 1P and 3P FePO4 samples calcinated at different temperatures and at different current density. All FePO4 samples prepared at 600 and 700 ◦ C delivered low capacity of 40–50 mAh g−1 (shown in supporting information files, Fig. 12). This inactive behavior might be attributed to the presence of a glassy surface phase [9,19] or due to presence of some impurities such as Fe2 O3 and Fe3 PO7 , where Fe2 O3 is possibly formed easily

Fig. 5. Specific discharge capacity upon cycling of 1P-FePO4 samples (1P-400 (), 1P-450 (), 1P-500 () and 1P-600 ()) at 0.1, 1, 3 and 0.1 C.


Fig. 6. Specific discharge capacity upon cycling of different FePO4 samples (3P-400 (), 3P-450 (), 3P-500 () and 3P-600 ()) at 0.1, 1, 3 and 0.1 C.

in our conditions in small concentration below the detection limit of XRD technique, and Fe3 PO7 was detected by XRD (Figs. 1 and 2 in supporting information files) and might be responsible for the poor lithium ion intercalation/deintercalation kinetics in the crystalline trigonal FePO4 . In agreement with our recent results [2], the specific capacities of 1P-FPO samples were not highly improved (Fig. 5). 1P450 and 1P-500 samples showed initial capacity of 140 mAh g−1 at 0.1 C at the first discharge progress, which suffered an irreversible capacity of about 40 mAh g−1 and decreased to approximately 100 mAh g−1 at the subsequent cycles. In comparison with 1P-FPO samples, 3P-FPO samples showed better initial capacity and less irreversible capacity upon subsequent cycling. This behavior can be associated to the higher specific surface area, larger pore volume and larger pore diameter, which obtained upon increasing the P123/(Fe + P) molar ratio (Fig. 6). Among 3P samples, mesoporous 3P-450 samples delivered the highest discharge capacity of 159 mAh g−1 at 0.1 C rate, which reached 93.5% of the theoretical capacity of FePO4 . Even though, 3P-450 sample showed initial discharge capacity less than that of 3P-400. In the subsequent cycles, the average discharge capacity of 3P-450 was 135 at 0.1 C rate, 110 at 1 C rate, and 85 mAh g−1 at 3 C rate which is higher than that of RuO2 -added quartz FePO4 (110 mAh g−1 at C/3 rate) [8] and of amorphous FePO4 (70 mAh g−1 at 1 C rate) [9,10,15,20]. 3P-400 samples showed a much lower capacity, because the calcination temperature was not high enough for completing the reaction. Although 3P-500 samples had a mesoporous structure, they delivered a lower capacity which might be a result for its smaller specific surface area and pore volume. From the above results, we can conclude that the best electrochemical properties obtained for the P123/(Fe + P) ratio was at temperature of 450 ◦ C among 3P materials and it was of 500 ◦ C for the 1P materials. This implied that the ideal calcination temperature for synthesizing mesoporous FePO4 with the highest specific capacity was around 450–500 ◦ C and is affected by the molar ratio of P123/(Fe + Fe). In order to distinguish the impact of surfactant molar ratio on the mesostructure of FePO4 samples, we synthesized a series of


Z.C. Shi et al. / Electrochimica Acta 53 (2008) 2665–2673

Fig. 7. Specific discharge capacity upon cycling of different FePO4 samples (1P-450 (), 2P-450 (), 3P-450 () and 4P-450 ()) at 0.1, 1, 3 and 0.1 C.

mesoporous FePO4 at 450 ◦ C with different P123/(Fe + P) molar ratios. Then the cathode performance of these samples with different specific surface area and pore diameter were investigated by charge/discharge measurements at various rates as shown in Fig. 7. Cell tests proved that 3P-450 sample delivered the highest capacity of 135 mAh g−1 at 0.1 C rate, while 2P-450 and 4P-450 samples only delivered a capacity of 115 mAh g−1 , and the capacity of 1P-450 decreased to 90 mAh g−1 . On further cycling at high rates of 1 C and 3 C, 3P-450 sample delivered the highest capacity of 110 and 85 mAh g−1 , respectively, while when cycled at 0.1 C rate again, all samples could be recovered the original high capacity, which confirmed the stability of mesoporous structure during cycling. 3P-450 sample exhibited the best cathode performance which might be explained due to its larger specific surface area and pore diameter which might facilitate the fast transport of electrolyte and lithium ion intercalation/deintercalation, than that of 1P-450 and 2P-450 samples. Even sample 4P-450 possessed the highest specific BET surface area, pore volume and pore diameter, it is not understandable why it has lower capacity than that of 3P-450 sample. However, 3P-450 showed better electrochemical performance when cycled compared to the rest of the other materials. At this point, even it is speculative; one can think about that there is ideal pore diameter which allows the fastest kinetics of lithium intercalation/deintercalation, after it, the lithium intercalation kinetics decreases again. 3.4. EIS characterization of lithium batteries based on mesoporous FePO4 cathodes One of the advantages of EIS is that it can study the surface of the electrodes in question compared to cyclic voltammetry which can only study the bulk of the electrode. EIS as non-transient technique not alter the electrode-interface which make this technique more precise one. In this work, to study the effect of mesoporous structure on the lithium intercalation/deintercalation process of FePO4 cathodes, EIS

Fig. 8. Typical EIS of (a) Li/meso-FePO4 cell from 106 to 10−3 Hz at 25 ◦ C, insets show its high-medium frequency segments and (b) equivalent circuit of impedance for this cell ( experimental data; — fit result of experimental data).

experiments carried out on the above coin cells at equilibrium state after 20 charge/discharge cycles. Fig. 8(a) presents a typical EIS of Li/mesoporous FePO4 coin cells at 2.87 ± 0.01 V after 20 charge/discharge cycles. The impedance behavior included three semicircles followed by linear part. The first semicircles were at high frequency range of 106 to 5000 Hz and the second semicircle at 5000–30 Hz (Fig. 8(a), lower inset) while the third semicircle at 30–0.2 Hz (Fig. 8(a) upper inset) followed by a linear part at very low frequency of 0.2–0.01 Hz. The fitting circuit used in this work is shown in Fig. 8(b), where Re represented the electrolyte resistance; CPE1 and R1 represented the elements of the interfacial layer between the FePO4 -carbon composite and the electrolyte and they signify the capacitive and resistive contribution from the surface film; while CPE2 and R2 represented the bulk capacitive and resistive contribution; and the CPE3 and R3 represented the double layer capacitance and charge transfer resistance, respectively. Since the diffusion coefficient was out of the interest in this discussion, we did not fit the linear part of the impedance. Fig. 8(a), showed that the non-linear curve fitting were in high accordance with the experimental data and that the χ2 and weighted Σ 2 were of 0.000419 and 0.0528, respectively (Table 1 in supporting information files). True capacitance and constant phase element (CPE) are related to each other according to the following equation: C = Q0 (max )n−1


The true capacitance is equal to CPE (its magnitude represented by Q0 if n = 1), and CPE capacitance properties decreases when n approaches zero, and ωmax is the angular frequency at which maximum imaginary impedance attains.

Z.C. Shi et al. / Electrochimica Acta 53 (2008) 2665–2673


Fig. 9. EIS of fresh Li/Li coin cells after two cycles () and seven cycles () of cyclic voltammetry between 0.5 and −0.5 V at frequency range of 1–5 mHz.

The reasons of having depressed semicircles may be attributed to non-flat electrode surface as early discussed by the pioneer work of de Levie [21], fractal geometry of the electrode was proposed recently as another reason for this depressed semicircle [22] which in consent of the early work of de Levie. The existence of depressed semicircle in this work is highly probably due to the mesoscopic nature of FePO4 material. Besides, since the electrolyte decomposition cannot be excluded, it may contribute to this behavior as well. Other contribution can originate from the formation of solid electrolyte interface (SEI) which can be formed, as reported previously, very fast on carbonaceous materials [23], and on oxide materials [24–26] especially at the early stages of the lithium intercalation or deintercalation [25]. SEI formation can be a direct interpretation of why Rct increase. The formation of SEI is likely to be multilayer of compact and porous in nature and can be formed during cycling [27,28]. To ascertain that the main contribution of the passivity and charge transfer resistance is mainly originated from the LiFePO4 and not from the lithium anode, a symmetrical coin cell containing Li as working and reference electrodes where used and studied by the EIS (Fig. 9) after different cycles of cyclic voltammetry. Fig. 9 shows that the contribution of the impedance from Li/Li coin cell is negligible compared to that one of the total impedance of Li/FePO4 coin cells (Fig. 8), which support that the main impedance is originated from the mesoporous Li/FePO4 but not from Li anode. Figs. 10 and 11 evaluate the EIS of Li/mesoporous FePO4 (450-FPO) cell to that of Li/crystalline FePO4 (3P-700) cell at 2.86 ± 0.01 V after 20 cycles in the frequency range from 106 to 10−3 Hz at 25 ◦ C. Fitting results of all the EIS data using the appropriate equivalent circuit mentioned (Table 1 in supporting information files). The χ2 and Σ 2 of fitting were low enough for us to confirm our appropriate choice of the equivalent circuit. Moreover, all Re (electrolyte resistance) were smaller than 10  with small difference, which could be due to the tiny difference of cell assembly and the interaction between the electrolyte and the dendritic formation of SEI after 20 cycles of charging–discharging and all CPE1 and CPE2 were on the same magnitude of 10−5 F because all of them originated from the same source, i.e., surface and bulk of electrode materials, respectively, while all CPE3 reached the magnitude of 10−3 F because it represented the charge transfer process ensued inside active cathode materials. Similarly to that of Li/mesoporous FePO4 (3P-450 sample) cell, Li/non-mesoporous crystalline FePO4 (3P-700 sample) cell

Fig. 10. (a) EIS of Li/FePO4 cell based on 3P-450 () and 3P-700 () samples at 2.86 V (open-circuit voltage) from 105 to 5 × 10−3 Hz at 25 ◦ C and (b) shows their high-medium frequency segments.

also showed an EIS which had the characteristic of two depressed semicircles and straight lines (Fig. 10(a)). Due to the large specific surface area (113 m2 g−1 ) and residual carbon (1.72 wt.%), many surface reactions with electrolyte arose on the particle surface of 3P-450 mesoporous FePO4 sample, which hence resulted

Fig. 11. EIS of Li/FePO4 cell based on 1P-450 (), 2P-450 (), 3P-450 () and 4P-450 (䊉) samples at 2.86 ± 0.01 V (open-circuit voltage) at 25 ◦ C: (a) shows the full spectra from 106 to 10−3 Hz and (b) shows their high-medium frequency segments.


Z.C. Shi et al. / Electrochimica Acta 53 (2008) 2665–2673

Table 2 Fitting results of EIS in Figs. 10 and 11 Sample



Re ()

R1 ()



R2 ()



R3 ()



1P-450 2P-450 3P-450 4P-450 3P-700

2.0E−3 1.5E−3 0.5E−3 0.4E−3 0.6E−3

0.31 0.23 0.07 0.08 0.08

3.2 3.6 9.4 5.6 8.9

6.1 5.2 11.0 8.4 14.8

1.2E−5 0.3E−5 1.6E−5 5.4E−5 5.5E−5

0.93 0.91 0.78 0.72 0.69

271 228 73 152 22

1.3E−5 1.2E−5 2.5E−5 1.8E−5 1.8E−5

0.75 0.82 0.86 0.85 0.87

1015 426 210 338 625

1.7E−3 1.7E−3 1.5E−3 1.9E−3 2.0E−3

0.70 0.79 0.80 0.75 0.72

in a larger surface impedance on the surface layer (Fig. 10(b)). Simultaneously, the mesoporous structure and carbon residual in those mesoporous FePO4 samples enhanced lithium ion intercalation kinetics, which resulted in a smaller Rct than that of 3P-700 sample. At the same time, the increase of the temperature resulted in growth of the particles which can enhance the resistivity according to lose of the interconnectivities to the carbon material. The mesoporous FePO4 samples enhanced electrochemical performance because of a small Rct during lithium intercalation/deintercalation process which probably due to the addition of carbon which can enhance the FePO4 electronic conductivity. To ascertain that the mesoporous is the key factor of enhancement of the electronic conductivity and not the residual carbon in the samples, we recall that P-450 samples, the carbon content were similarly with minor differences (Table 1), especially the last three samples (2P-450, 3P-450 and 4P-450) where carbon content were 1.67, 1.74 and 1.72 wt.%, respectively. Moreover, additional carbon of 20 wt.% in the form of carbon black was mixed with 75 wt.% FePO4 and 5 wt.% PVDF during the preparation of FePO4 composite electrode. Therefore, the difference of carbon content in these samples gets more negligible. The composite carbon is therefore, took a part in improving the conductivity between the particles of the FePO4 while the added carbon during the electrode fabrication was to enhance the conductivity between the current collector and the composite material. On the other hand, the mesoporosity was the main feature responsible for enhancement of conductivity of the FePO4 material. Fig. 11(a) shows the EIS of Li/mesoporous FePO4 cells based on 1P-450, 2P-450, 3P-450 and 4P-450 samples at 2.86 ± 0.01 V in the frequency range from 106 to 10−3 Hz at 25 ◦ C. From the high-medium frequency segments (Fig. 11(b)), we can notice that 3P-450 sample has the smallest impedance, as a result of its improved electrochemical performance. Fitting results of these EIS data indicated that FePO4 cathodes with different mesoporous structure had different impedance response (Table 2), with the surfactant ratio increased, specific surface area, pore volume and pore diameter of mesoporous FePO4 samples were increased (Table 1), which initiated the Rct to lessen because large specific surface area could provide large active sites for lithium ion intercalation/deintercalation, large pore channels could facilitate fast transport of electrolyte containing lithium ion, and large pore volume could accommodate more electrolyte and can lead to thin pore wall which reduce the resistance of lithium ion intercalation/deintercalation into/from inner active FePO4 materials. For 4P-450 sample, though possessed the largest specific surface area, pore volume and pore

diameter, it exhibited larger Rct than that of 3P-450 sample because of its less homogeneous pore distribution. Therefore, in general, large specific surface area, pore sizes, pore volume, and homogeneous pore distribution in the FePO4 cathode material will improve aptly its electrochemical performance (Table 1). Our results of impedance, even it is for two-electrode system, resemble that of the three-electrode system especially after charging–discharging. This observation was in agreement with the work of Zane et al. [29] where they showed that the only difference between three-electrode system and two-electrode system is only a small shift in the Rct values, being lower for the three-electrode system. 4. Conclusions In summary, poor lithium intercalation/deintercalation kinetics of FePO4 could be improved by importing a mesoporous structure, which could be tuned by changing template ratio and calcination temperatures. EIS tests proved that this mesoporous structure could enhance their electrochemical performance, especially at high current density, as indicated by low values of Rct of FePO4 cathodes. The results demonstrated the role of mesoporous structure in improving the fast transport and intercalation kinetics of lithium ions in the FePO4 materials. Moreover, this work presented a new method for preparing new materials with good rate of performance for batteries and other energy storage devices such as electrochemical supercapacitors. Acknowledgments The financial supports from the National Natural Science Foundation of China (Nos. 29925310 and 20021002) and the Ministry of Science and Technology of China (2001CB610506) are acknowledged. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.electacta.2007.06.079. References [1] J.W. Long, B. Dunn, D.R. Rolison, H.S. White, Chem. Rev. 104 (2004) 4463. [2] Z.C. Shi, Y.X. Li, W.L. Ye, Y. Yang, Electrochem. Solid-State Lett. 8 (2005) A396. [3] P. Liu, S.H. Lee, Y. Tracy, J.A. Turner, Adv. Mater. 14 (2002) 27.

Z.C. Shi et al. / Electrochimica Acta 53 (2008) 2665–2673 [4] E. Kim, D. Son, T.C. Kim, J. Cho, B. Park, K.S. Ryu, S.H. Chang, Angew. Chem. Int. Ed. 43 (2004) 5987. [5] L. Kavan, A. Attia, F. Lenzmann, S.H. Elder, M. Graetzel, J. Electrochem. Soc. 147 (2000) 2897. [6] A.K. Padhi, K.S. Nanjundaswamy, C. Masquelier, S. Okada, J.B. Goodenough, J. Electrochem. Soc. 144 (1997) 1609. [7] A.K. Padhi, K.S. Nanjundaswamy, C. Masquelier, J.B. Goodenough, J. Electrochem. Soc. 144 (1997) 1188. [8] F. Croce, A. D’Epiffanio, P. Reale, L. Swttimi, B. Scrosati, J. Electrochem. Soc. 150 (2003) A576. [9] Y. Song, S. Yang, P.Y. Zavalij, M.S. Whittingham, Mater. Res. Bull. 37 (2002) 1249. [10] Y.S. Hong, K.S. Ryu, Y.J. Park, M.G. Kin, J.M. Lee, S.H. Chang, J. Mater. Chem. 12 (2002) 1870. [11] P.P. Prosini, M. Lisi, S. Scaccia, M. Carewska, F. Cardellini, M. Pasquali, J. Electrochem. Soc. 149 (2002) A297. [12] Z.C. Shi, Y. Yang, Prog. Chem. 17 (2005) 604. [13] J. Santos-Pena, P. Soudan, C.O. Arean, G.T. Palomino, S. Franger, J. Solid State Electrochem. 10 (2006) 1. [14] X. Guo, W. Ding, X. Wang, Q. Yan, Chem. Commun. 8 (2001) 709. [15] S. Okada, T. Yamamoto, Y. Okazaki, J.-I. Yamaki, M. Tokunaga, T. Nishida, J. Power Sources 146 (2005) 570.


[16] P. Reale, B. Scrosati, Chem. Mater. 15 (2003) 5051. [17] A. Ait-Salah, J. Dodd, A. Mauger, R. Yazami, F. Gendron, C.M. Julien, Z. Anorg. Allg. Chem. 632 (2006) 1598. [18] C.M. Burba, R. Frech, J. Electrochem. Soc. 151 (2004) A1032. [19] Y. Song, P.Y. Zavalij, M. Suzuki, S. Whittingham, Inorg. Chem. 41 (2002) 5778. [20] C. Masquelier, P. Reale, C. Wurm, M. Morcrette, L. Dupont, D. Larcher, J. Electrochem. Soc. 149 (2002) A1037. [21] R. de Levie, Electrochim. Acta 9 (1964) 1231. [22] K.-N. Jung, S.-I. Pyun, Electrochim. Acta 51 (2006) 2646. [23] D. Aurbach, K. Gamolsky, B. Markovsky, G. Salitra, Y. Gofer, U. Heider, R. Oesten, M. Schmidt, J. Electrochem. Soc. 147 (2000) 1322. [24] S.S. Zhang, K. Xu, T.R. Jow, J. Electrochem. Soc. 149 (2002) A1521. [25] A. Attia, M. Zukalova, L. Pospisil, L. Kavan, J. Solid State Electrochem. 11 (2007) 1163. [26] F. Nobili, F. Croce, B. Scrosati, R. Marassi, Chem. Mater. 13 (2001) 1642. [27] C.R. Yang, J.Y. Song, Y.Y. Wang, C.C. Wan, J. Appl. Electrochem. 30 (2000) 29. [28] Y.G. Ryu, S.I. Pyun, J. Electroanal. Chem. 433 (1997) 97. [29] D. Zane, M. Carewska, S. Scaccia, F. Cardellini, P.P. Prosini, Electrochim. Acta 49 (2004) 4259.