Md. Selim Arif Sher Shah‡
a,
Shoaib Muhammad‡b,
Jong Hyeok Parkac,
Won-Sub Yoon*b and
Pil J. Yoo*ac
aSchool of Chemical Engineering, Sungkyunkwan University, Suwon 440-746, Republic of Korea. E-mail: pjyoo@skku.edu
bDepartment of Energy Science, Sungkyunkwan University, Suwon 440-746, Republic of Korea. E-mail: wsyoon@skku.edu
cSungkyunkwan Advanced Institute of Nanotechnology, Sungkyunkwan University, Suwon 440-746, Republic of Korea
First published on 22nd January 2015
SnO2, a candidate material for anodes in Li-ion batteries (LIBs), usually suffers from severe volume change (>300%) during charge–discharge cycles. This problem leads to undesirable continuous capacity fading, hindering its practical utilization. To address this issue, nanostructured SnO2 and its composites with carbon nanomaterials, especially graphene, have extensively been studied. Although the stability issue has improved substantially, these materials still suffer from low capacity characteristics, which are far from the theoretical capacity of SnO2. Motivated by this background, in this work, we synthesized a novel ternary nanocomposite of SnO2, reduced graphene oxide (rGO), and a conducting polymer, poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS), as a high performance anode material in LIBs. PEDOT:PSS together with rGO is expected to efficiently accommodate the volume change in SnO2 during cycling. Transmission electron microscopic observation reveals 2–3 nm-sized SnO2 nanoparticles are uniformly dispersed over rGO nanosheets while having a PEDOT:PSS coating. The capacities of the synthesized composites were dependent on the PEDOT:PSS concentration. The reversible capacity of the composite with 5 wt% PEDOT:PSS was maintained at 980 mA h g−1 with a coulombic efficiency over 99% even after 160 cycles. This capacity value is equivalent to 1185 mA h g−1 on the basis of only SnO2 in the composite. The high capacity of the ternary nanocomposites is attributed to the ultra-small size of SnO2 nanoparticles, enhanced electronic and ionic mobility, and facilitated volumetric relaxation synergistically offered by rGO nanosheets and the PEDOT:PSS coating.
Composites of SnO2 with different carbon nanomaterials may be superior to SnO2 alone because the carbon materials act as a cushioning layer to buffer against volume change and to improve electrical conductivity of the composites. In this regard, graphene, a single-atom-thick planar carbon nanomaterial, is considered as a promising 2D material due to its excellent electrical conductivity, large surface area (>2600 m2 g−1), and high structural flexibility.34–36 Therefore, SnO2–graphene nanocomposites have extensively been studied as a promising anode material for advanced LIBs. However, the problem of pulverization and capacity fading has not been completely eliminated.37–39 Such a limitation indicates that there is still much room to increase the capacity of such composite systems.
A good strategy to overcome the limitation of SnO2–graphene composites is to make ternary nanocomposites, in which the third component further accommodates the volume change more efficiently in comparison to the binary composites. There are several reports where nanostructured SnO2 materials were sandwiched between graphene layers. For example, Prabakar et al. reported alternating stacks of SnO2 with graphene oxide (GO) and amine functionalized graphene, which showed an excellent capacity of 872 mA h g−1 after 200 cycles at a current density of 100 mA g−1.40 N-doped graphene–SnO2 sandwiched papers were demonstrated by Wang et al., wherein a capacity over 800 mA h g−1 at 50 mA g−1 after 50 cycles and enhanced cycle stability of the composite were reported.41 The excellent electrochemical behaviour was attributed to the N-doped graphene and optimized structural features. In another report, Sun et al. described sandwiched graphene–SnO2 nanorod-carbon nanostructures for ultra-high lithium storage properties.4 They achieved a capacity of 1419 mA h g−1 after 150 cycles at 0.1 C, as compared to 389 mA h g−1 for graphene–SnO2 nanorod complexes. Very recently, Bhaskar et al. demonstrated ternary composites of SnO2 hollow spheres with poly(3,4-ethylenedioxythiophene) (PEDOT) and graphene oxide.42 They reported a capacity of 608 mA h g−1 at a current density of 100 mA g−1 after 150 cycles. Since they employed non-electroconductive graphene oxide instead of graphene or reduced graphene oxide, obtained capacity value was relatively low. To date, only a few reports have described a capacity that is close to the theoretical value of SnO2. In this regard, there is an urgent need to develop new electrode materials possessing remarkably enhanced capacities and high rate performance for development of the next-generation LIBs.
In the present work, we report ternary composites of SnO2 nanoparticles with PEDOT:PSS and reduced graphene oxide (rGO). The composites were synthesized through simple yet robust one-pot wet chemical method, namely, complexation and precipitation of SnO2 nanoparticles on rGO under reflux followed by the addition of PEDOT:PSS (Scheme 1). The synthetic conditions are mild, and the maximum temperature of 100 °C lasts for 1.5 h, followed by heating at 65 °C for 15 h; water is used as the only solvent. Under these synthetic conditions, we obtained small SnO2 nanoparticles (2–3 nm in size) while excluding the use of surfactants or harsh chemicals. The ultra-small SnO2 nanoparticles are uniformly distributed over rGO nanosheets and the composites are stably enwrapped with PEDOT:PSS. As a result, the undesirable tendency of SnO2 nanoparticles to self-agglomerate was largely eliminated. Moreover, the ultra-small SnO2 nanoparticles led to a decrease in the lithium diffusion length, which, together with the efficient accommodation of volume change of SnO2 during charge–discharge cycles, ultimately gives rise to high capacity. Accordingly, the ternary nanocomposite shows high lithium storage capacity and excellent capacity retention characteristics, which are much greater than those of SnO2-based binary composites. The specific capacities of the ternary composites vary depending on the amount of PEDOT:PSS.
Fig. 1 (a) X-ray diffractograms of SnO2, GS, and GSP. (b) FTIR spectra of GS, SP, and GSP. (c) Raman spectra of PEDOT:PSS, GS, and GSP. |
Raman spectral analysis was carried out to provide a detailed understanding of the chemical composition of the synthesized materials. As depicted in Fig. 1c, the Raman spectrum of PEDOT:PSS clearly shows bands at 1568 and 1503 cm−1 for the CαCβ asymmetric stretching mode, 1430 cm−1 for the symmetric CαCβ stretching frequency, 1361 cm−1 for the Cβ–Cβ stretching vibration, 1258 cm−1 for the Cα–Cα inter-ring stretching mode, 1097 cm−1 for C–O–C stretching, 994 and 579 cm−1 to oxyethylene ring deformation, 700 cm−1 for symmetric C–S–C deformation, and 441 cm−1 for SO2 bending vibration.47,42 Two main peaks are observed in the Raman spectra of GS at 1331 and 1598 cm−1, which are assigned, respectively, to the D- and G-band of graphene.48 The ternary composite, GSP, displays all the peaks present in PEDOT:PSS and GS.
X-ray photoelectron spectroscopy was used to further characterize the composites. The survey XPS spectrum of GSP5 (Fig. S1 in the ESI†) shows that the composite consists of Sn, C, O and S. Fig. 2a shows a high resolution C 1s XPS spectrum of the composite GSP. The peak at 284.6 eV is for the CC of graphene.43 The binding energy of C–OH of graphene and C–S of PEDOT:PSS in the α position is observed at 285.8 eV. CO of rGO and CC–O in the β position of PEDOT result in a peak at 287.06 eV.42 At the same time, the COOH peak of rGO and C–O–C bonding in the ethylene bridge of PEDOT occurs at 288.5 eV. The presence of PEDOT:PSS was further proven by characteristic XPS spectra of S 2p, as depicted in Fig. 2b. Spin–orbit splitting contributions of S 2p in PEDOT were found at 163.9 eV for S 2p3/2 and at 165.3 eV for S 2p1/2. Another doublet arises at 167.9 and 169.3 eV, respectively for S 2p3/2 and S 2p1/2 of PSS.49
The higher energy doublet originated from positively charged sulphur of PSS because a highly electronegative oxygen atom is attached to the S atom. All sulphur contributions have a separation between 2p1/2 and 2p3/2 of ∼1.4 eV, which is close to the literature values. From the XPS spectrum of S, we can unequivocally conclude that the ternary composite contains PEDOT:PSS. Fig. 2c displays high resolution XPS spectra of Sn 3d. As shown, a doublet is observed at 487.6 and 496.0 eV for Sn 3d5/2 and 3d3/2, respectively.41 The separation between the two peaks of the doublet was determined to be 8.4 eV. These observations indicate that Sn is present as Sn4+ in the composite, i.e., as SnO2.
Fig. 3a shows a transmission electron microscope (TEM) image of the composite GSP. From the micrograph, it is evident that 2–3 nm SnO2 nanoparticles coexist with rGO and PEDOT:PSS. This result, apart from XRD analysis, further confirms the presence of ultra-small SnO2 nanoparticles in the composites. A high resolution TEM (HRTEM) micrograph of GSP shown in Fig. 3b clearly depicts highly crystalline and ultra-small SnO2 nanoparticles were obtained following a relatively mild synthetic procedure. TEM micrographs clearly show that the SnO2 nanoparticles do not form any obvious aggregation. This is an important property, particularly for battery applications, as it mitigates the risk of possible pulverization. Fringe lines with a spacing of 0.33 nm are clearly visible in Fig. 3b, implying (110) planes of rutile SnO2.50 Fig. 3c is a typical scanning electron microscope (SEM) image of the composite. In the inset in Fig. 3c, an energy dispersive X-ray spectrum (EDS) of the composite is provided, demonstrating that the composite consists of C, O, Sn and S. Elemental mapping is shown on the right side of c. The amount of SnO2 nanoparticles in the composite was determined by thermogravimetric (TG) analysis. Fig. S2† is a TG trace of GSP5 showing that the amount of SnO2 nanoparticles in the composite is ∼74%.
Fig. 4a provides voltage curves of the as-synthesized pure SnO2 in the potential range of 0.005 to 3.0 V vs. Li/Li+ performed at constant current density of 80 mA g−1 (0.1 C). The first charge and discharge capacities were 624 and 1916 mA h g−1 with a coulombic efficiency of merely 32.6%. Other selected voltage profiles for SnO2 are also shown. Fig. 4b represents corresponding differential curves (dQ/dV). The differential curve (1st cycle) of SnO2 display two cathodic peaks at 1.05 and 0.22 V; the first peak corresponds to the reduction of SnO2 to Sn (reaction (4)), while the later peak is due to the alloying of Sn with Li according to reaction (5) and insertion of Li into rGO (reaction (6)).51,52 It is notable that there is a shoulder to the broad peak at 1.05 V. This may be due to the formation of a solid electrolyte interface (SEI). It may be noted that, in the second cycle, the reduction peak at 1.05 V disappears and a weak peak arises at 0.72 V, which may be due to the reduction of SnO2 to Sn. The first anodic cycle shows two peaks at 0.49 and 1.18 V, which are due to de-alloying of LixSn and oxidation of Sn to SnO2. In the second cycle, the oxidation peak at 1.18 V disappears, implying that only the conversion of Sn to LixSn occurs in the successive cycles.52 However, the alloying peak in the cathodic scan and de-alloying peak in the anodic scan exist in the second cycle. After repeated cycles (e.g. 50 times), intensities of all the peaks from both cathodic and anodic processes decrease largely and then nearly disappear after around 100 cycles.
Fig. 4 Charge–discharge profiles and the corresponding differential curves of (a and b) SnO2 nanoparticles and (c and d) GSP5. |
Galvanostatic charge–discharge profiles of the composite anode GSP5 were evaluated in the same condition as with pure SnO2. The first discharge and charge capacity achieved with the cell were 1759 and 1029 mA h g−1, as observed in Fig. 4c. Other representative charge–discharge profiles are also shown in Fig. 4c for the sake of comparison. To investigate the electrochemical processes occurring during charge–discharge cycles, plots of dQ/dV (known as differential capacity) vs. potential, known as differential curves, are displayed in Fig. 4d. A plateau in the charge–discharge curve appears as a peak in the differential curve. The plot of first discharge (cathodic) cycle exhibits four peaks at 1.32, 0.84, 0.42, and 0.18 V. The peak at 1.32 V indicates the reduction of the electrolyte and formation of a solid electrolyte interface (SEI) at the interface of the electrode and electrolyte. The broad peak at 0.84 V corresponds to the reduction of SnO2 to SnO and formation of Li2O. SnO was further reduced to metallic Sn, and Li2O was formed at 0.42 V.53 The fourth peak at 0.18 V is due to the alloying of Li and Sn to form LixSn (0 ≤ x ≤ 4.4) and the intercalation of Li+ ions into rGO. All the reactions occurring at the electrode are given below.
Electrolyte + Li+ → solid electrolyte interface | (1) |
SnO2 + 2Li+ + 2e− → SnO + Li2O | (2) |
SnO + 2Li+ + 2e− → Sn + Li2O | (3) |
SnO2 + 4Li+ + 4e− → Sn + 2Li2O | (4) |
Sn + xLi+ + xe− → LixSn (0 ≤ x ≤ 4.4) | (5) |
rGO + xLi+ + xe− → LixC | (6) |
In the second cycle of the cathodic scan, the peak at 1.32 V disappears. This may be due to the fact that no additional SEI formation occurs on the electrode. All other peaks are present in the successive scans. During the first anodic (charge) scan, two peaks arise at 0.49 and 1.23 V. The oxidation peak at 0.49 V is assigned to the de-alloying of LixSn to give rise to Sn and Li, whereas the peak at 1.23 V is due to the oxidation of Sn to SnO and SnO2. The reversible capacity of SnO2 (determined to be 782 mA h g−1) is on the basis of the reversibility of reaction (5). However, we observe that the anodic peak at 1.23 V is present in all the cycles, and the intensity of the peak increases after 100 cycles. This experimental fact implies that reaction (4) is also reversible, at least to some extent, and it contributes to the extra capacity of SnO2 in the composites.25,31,54,55 Apart from this, a new peak at 2.29 V emerges in the anodic scans after the second cycle. In contrast, it is obvious from Fig. 4b that reaction (4) is not reversible for the SnO2-only anode.
The cycling stability of the composites was studied at a constant current density of 80 mA g−1 (0.1 C) in the potential range of 0.005 to 3.0 V vs. Li/Li+, as shown in Fig. 5a. For a comparison, the performance of pure SnO2 was also observed under the same experimental conditions. As mentioned earlier, pure SnO2 shows first charge and discharge capacities of 624 and 1916 mA h g−1 with a coulombic efficiency of merely 32.6%. The capacity of SnO2 fades continuously and preserves a discharge capacity of 42 mA h g−1 after 100 cycles. A binary composite of SnO2 and GS shows better performance compared to pure SnO2. The initial discharge capacity of GS is 2071 mA h g−1, which, after continuous decay, ends up being 476 mA h g−1 after 100 cycles, as depicted in Fig. S3.† The low capacity value of GS is consistent with results from other reports.42,56,57
Fig. S3† further shows the discharge capacities of the composite SP5 or simply SP. The initial discharge capacities of the composites GSP2, GSP5, and GSP7 are respectively, 2015, 1759, and 1699 mA h g−1, whereas the charge capacities are 1145, 1029, and 839 mA h g−1, respectively. The coulombic efficiency values of the GSP2, GSP5, and GSP7 composites are 49.9, 58.5 and 49.4%, respectively. Low values of the coulombic efficiencies of first cycles may be due to the irreversible conversion of SnO2 to Sn and Li2O and the formation of SEI on the surface of the active material.40 The charge–discharge data of GS and SP5 are provided in the ESI.† Comparison of charge–discharge characteristics of all samples clearly demonstrate that the ternary composites, especially those with a high concentration of PEDOT:PSS (>2 wt%), show stable capacity from ∼10 cycles. This was further proved by the charge–discharge characteristics of GSP10 displayed in Fig. S4,† which shows stable capacity from the 10th cycle. The GSP10 composite exhibits a specific capacity of 709 mA h g−1 after 229 cycles with a coulombic efficiency >99%. It is notable that the specific capacity of the ternary composites increases first, then decreases with increasing amounts of PEDOT:PSS. The initial capacity loss in these electrodes is mainly due to the formation of SEI. Note that the capacity values were calculated on the basis of the total mass of each electrode. It is obvious that all the composites exhibit higher capacity than pure SnO2, and the ternary composites display even better capacity than the binary composites of GS or SP.
Notably, the composite GS is better in retaining the capacity compared to the composite SP. The severe capacity fading in SnO2, GS, and SP may be attributed to the pulverization of the electrodes. However, addition of PEDOT:PSS in GS increases the capacity of the ternary composites first and then decreases the capacity depending on the concentration of PEDOT:PSS. Capacity fading took place even with composite GSP2, which was not as severe as that with SnO2, GS and SP. The discharge capacity of GSP2 after 100 cycles was 515 mA h g−1. On the other hand, composites GSP5 and GSP7 show stable capacity from about 10 cycles. Therefore, we can conclude that the rGO together with PEDOT:PSS efficiently accommodate the volume change of SnO2 during the charge–discharge process, leading to stable capacity. The composite GSP5 shows a reversible capacity of about 974 mA h g−1 after 160 charge–discharge cycles with a coulombic efficiency over 99%, indicating that a highly reversible capacity of 1185 mA h g−1 can be achieved for SnO2 in the composite. This capacity is almost greater than by a factor of three than the commercial graphite anode, which shows a reversible capacity of 372 mA h g−1. Moreover, it is somehow close to the reversible capacity of 1494 mA h g−1 for nano-sized SnO2. On the other hand, GSP7 shows a discharge capacity of 851 mA h g−1 after 100 cycles, with a coulombic efficiency ∼99%. Good rate performance is essential for Li ion batteries. The rate performance of the composite is depicted in Fig. 5b, which shows that the composite exhibits excellent rate performance at different current densities. Even at high current densities of 1 C and 2 C, the composite delivers specific capacity values of 566 and 424 mA h g−1 at the seventh cycle, respectively. Moreover, the specific capacity can recover to 962 mA h g−1 when the current density returned to C/10.
Electrochemical impedance spectra (EIS) data for SnO2, SP, GSP2, GSP5, and GSP7 were collected and are shown in Fig. 5c. The EIS plots show semi-circles in the high frequency region and sloping straight lines at low frequencies. From these plots, it is clear that the ternary composites GSP5 and GSP7 show minimum charge-transfer resistance, which is due to the presence of highly conducting rGO and intimate contact of SnO2 nanoparticles with rGO and the conducting polymer PEDOT:PSS. Although their impedance values differ marginally, the composite of GSP5 displays more capacity than GSP7. It is notable that the impedance values of all the ternary composites are lower than that of GS due to the inclusion of PEDOT:PSS. Due to this reason together with insufficient cushioning to volume change, GS shows continuous capacity fading. On the other hand, SnO2 shows the highest charge-transfer resistance. From the above mentioned discussion, we can conclude that the high specific capacity of the composite GSP5 is obtained due to the following three aspects: (i) the ultra-small size of SnO2 nanoparticles, which decreases the lithium diffusion length in the composite matrix and partially accommodates volume change during charge–discharge cycling, eventually leading to the increased specific capacity, (ii) the highly conducting medium offered by rGO and PEDOT:PSS provides low charge-transfer resistance, and (iii) efficient buffering capability against volume change offered by the inclusion of rGO and PEDOT:PSS.
Footnotes |
† Electronic supplementary information (ESI) available: Survey XPS spectrum and thermo gravimetric analysis are shown in Fig. S1 and S2, respectively. Fig. S3 depicts cycle performance of GS and SP5. On the other hand, charge–discharge performance and coulombic efficiency of GSP10 is displayed in Fig. S4. See DOI: 10.1039/c4ra15913f |
‡ The authors contributed equally to this work. |
This journal is © The Royal Society of Chemistry 2015 |