Figure 6 SEM images of the LFP electrode fabricated with the water-based slurries of the pristine CNTs (a), SCNT-M (b), SCNT-S (c) and SNCNTs (d).
The electrode performances of the LFP electrodes prepared using water-based slurries were tested via assembly coin-type half cells. As shown in Figure 7a, the LFP electrode with SNCNTs conductive additive shows the highest discharge capacities among all the samples. It affords the average reversible capacity of 163.4 mAh g-1 at 0.2 C, obviously higher than the capacities of the LFP electrode with CNTs (155.4 mAh g-1). The reversible capacity of LFP-SNCNT remains 134.4 mAh g-1 at 5 C, 5.5% higher than that of LFP-CNT (127.4mAh g-1), showing excellent rate capability. The cycling performances of the as-prepared electrodes were shown in Fig 7b. The capacity retention of the LFP-SNCNT electrode is 99.5% after 100 cycles at 1C, obviously higher than that of LFP-CNT (96.3%). In the charge-discharge curves at 0.5 C (Figure 7d), extended charge and discharge platforms at ~3.4 V are observed for LFP-SCNT-S and LFP-SNCNT. It indicates that more active material particles have been involved in the charge-discharge processes, due to the connection by well-established conductive networks. Figure 7f shows the impedance curves of the as-fabricated LFP electrodes. The charge transfer resistance (R ct) of LFP-SNCNT is the smallest among all the simples. Although the conductivity of SNCNTs (806 S/m) is smaller than that of SCNT-S (1230 S/m), our results indicate that the better dispersion of SNCNTs in the water-based slurry and the as-resulted electrode due to their better hydrophilicity has played a dominant role in reducing the charge transfer resistance of the electrode.
Although the final discharge capacity at 2.0 V for LFP-SCNT-M is close to that for LFP-SCNT-S (Figure 7c), the discharge platform of LFP-SCNT-M drops earlier. The drop in the discharge platform for LFP-SCNT-M is likely due to not only the higher bulk charge transfer resistance aroused by the worse dispersion of CNTs, as compared to LFP-SCNT-S, but also the higher contact resistance between CNTs and LFP particles aroused by the uneven distribution of S heteroatoms in SCNT-M. The conductivity and hydrophilicity of SCNT-S are superior to SCNT-M, which has led to better electrode performance. On the other hand, in the synthesis of SCNT-M that uses MgSO4 as dopant, the S-doping reaction between carbon species and MgSO4 was significantly affected by the solid phase contact between the two reactants. In order to realize a good contact between CNTs and MgSO4, the amount of MgSO4 was kept at a high level. It means that the operation cost for adding and removal of Mg salts is high, and that the usage efficiency of MgSO4is limited. Therefore, we conclude that the gaseous S doping possesses the advantage of both lower cost and better product quality, as compared to the solid phase S doping.