Supplementary Figure 1 A schematic representation of the different reaction mechanisms observed in electrode materials for lithium batteries. Black circles: voids in the crystal structure, blue circles: metal, yellow circles: lithium. Copyright is the Royal Society of Chemistry, 2009 1.
Supplementary Figure 2 The Raman spectrum of graphene oxide (GO). Two peaks at 1351 and 1590 cm -1 correspond to the D and G bands of GO. Supplementary Figure 3 AFM characterization of graphene oxide nanosheet. The thickness of GO nanosheet is about 1 nm.
Supplementary Figure 4 XRD, SEM and EDS characterization of graphene-enveloped sulphur. (a) XRD pattern of S@GO particles. (b, c) SEM images of graphene-encapsulated sulphur. (d) EDS spectrum captured for the region shown in (a). (e, f) EDS carbon mapping, sulphur mapping of the region shown in (c). Supplementary Figure 5 XRD patterns of (a) Ni3S2@cG/Ni and (b) Ni3S2/Ni. The peaks can be indexed to Ni3S2, NiS (marked with ) and nickel (marked with ).
Supplementary Figure 6 Morphology of Ni3S2@cG/Ni, Ni3S2@0.5cG/Ni and Ni3S2/Ni. (a, b) TEM images of Ni3S2@cG/Ni. (c, d) TEM images of Ni3S2@0.5cG/Ni. (e, f) TEM images of Ni3S2. (g) Typical SEM image of Ni3S2@0.5cG/Ni. Crumpled graphene encapsulated particles, non-crumpled graphene encapsulated particles and bare particles are co-exist in this electrode. Scale bars, (a, c) 50 nm, (b, d) 20 nm, (e) 20 nm, (f) 10 nm, (g) 1 μm.
Supplementary Figure 7 TEM characterization and MD simulation of morphological evolution during synthetic process. TEM images of Ni3S2@cG at the reaction time of 0.5 h (a, b), 1 h (c, d), 2 h (e, f), 4 h (g, h). (i-l) MD simulation snapshots of the graphene s crumpling process at different reaction stages (Detailed process are shown in Supplementary Movie 1), which can represent the morphologies at the 0.5, 1, 2, 4 h, respectively. Scale bars, (a, c, e, g) 50 nm, (b, d, f, h) 10 nm.
Supplementary Figure 8 Schematic illustration of the crumpling mechanism of GO. Supplementary Figure 9 Charge/discharge curves. (a) Charge/discharge curves of Ni3S2@cG/Ni and Ni3S2/Ni. The voltage hysteresis is calculated from the difference value between charge and discharge voltage at the half reversible capacity, noted as ΔV(Q/2). (b) Charge/discharge curves of Ni3S2@cG/Ni at 1, 2, 4, 6, 8 A g -1.
Supplementary Figure 10 Rate performance of Ni3S2/Ni electrode with different voltage windows. Supplementary Figure 11 Electrochemical performance of Ni3S2@0.5cG/Ni. (a) Rate performance of Ni3S2@0.5cG/Ni electrode in different voltage windows. (b) Charge/discharge curves at different rate in the voltage windows of 0.01-3 V.
Supplementary Figure 12 Morphological evolution of crumpled graphene during electrochemical process. (a) Typical discharge/charge curve of Ni3S2@cG/Ni, red dots represent different voltage states of (i) pristine state, (ii) discharging to 1.5 V, (iii) discharging to 1 V, (iv) discharging to 0.01 V, (v) charging to 1.5 V, (vi) charging to 2.5 V. (b-g) SEM images of Ni3S2@cG/Ni electrode at the states of before testing (i), ii-vi, repectively. Crumples unfold with the expanding of nickel sulphides during the discharging process (b-e) and reversible fold during the inverse charging process (e-g). All scale bars in the figure are 1 μm.
Supplementary Figure 13 Schematic illustration of open cell nanobattery setup inside the TEM chamber. Supplementary Figure 14 The calculated potential energy of the graphene sheet in a period of 5 cycles of expansion & constriction for various circumference ratios β. (a) β = 0.965. (b) β = 0.975. (c) β = 0.986, which is also the same as Fig. 4h in the manuscript. (d) β = 0.996. (e) β = 1.007.
Supplementary Table 1 Elemental analysis results of Ni3S2@cG/Ni and Ni3S2@0.5cG/Ni. The mass of nickel sulphides and graphene were calculated based on mass of sulphur and carbon, which data was given by Vario EL Cube elemental analyzer. For the Ni3S2@cG/Ni, the mass of active materials (nickel sulphides and graphene) is ~2 mg and ~5 wt.% of that is graphene. For the Ni3S2@0.5cG/Ni, the mass of active materials is ~2 mg and ~2.5 wt.% of that is graphene. Sample Mass of active materials (mg) Mass graphene (mg) 1 2.01 0.0966 Ni3S2@cG/Ni 2 2.09 0.111 3 1.96 0.0976 Average 2.02 0.102 1 1.85 0.0469 Ni3S2@0.5cG/Ni 2 2.07 0.0511 3 1.91 0.0487 Average 1.94 0.0489
Supplementary Note A modified Hummers method for synthesis of graphene oxide The graphene oxide in our experiment is exfoliated from the graphite using a modified Hummers method. Firstly, 1.0 g of natural graphite powder mixed with 23 ml of 98% H 2 SO 4. The mixture was stirred at room temperature for 24 h followed by gradually adding 0.1 g of NaNO 3 at 40 o C with water bath environment. After 5 min, 1.5 g of KMnO 4 was added and stirred for 30 min. Next, 5 ml of deionized water was added and stirred for 5 min followed by another 5 ml deionized water. After 5 min, the water bath treatment was stopped, 140 ml of deionized water and 10 ml of 30% H 2 O 2 aqueous solution was added with stirring for 5 min. This mixture was washed by 1 M HCl twice followed by deionized water wash to ph = 7 and got precipitation. The precipitation was dissolved in 100 ml deionized water and sonicated for 1 h. A stable graphene oxide (GO) dispersion was obtained after removing the remaining precipitation. Demonstration of graphene s crumpling mechanism In order to demonstrate the crumpling mechanism, the TEM characterization on the samples at different reaction stages and MD simulation of synthetic process was performed. We captured the TEM images of the samples with different reaction time (0.5, 1, 2, 4 h). The TEM results reveal the graphene s morphological evolution: from partial-encapsulated non-crumpled state (Supplementary Fig. 7a, b), slightly wrinkled sheets (Supplementary Fig. 7c, d), initially crumpled shells (Supplementary Fig. 7e, f), to deeply-crumpled pinecone-like structure (Supplementary Fig. 7g, h). From the second reaction stage (reaction time of 1 h, Supplementary Fig. 7c, d), the graphene completely encapsulate the nickel sulphides and the crumpled degree is increasing with the crystal
growth of nickel sulphides, which indicate that the interactions between nickel sulphides and graphene play important roles in the crumpling of graphene. To further investigate the crumpling mechanism, we performed MD simulation of synthetic crumpling process of graphene. The methodology is very similar to that described in the Methods section of the manuscript. Starting with a graphene wrapped nanoparticle, MD simulations here focus on the crumpling behavior of the graphene in such a process: First, the nanoparticle slightly expands due to the melting of sulphur (Supplementary Fig. 7i). Then, the reaction between nickel and sulphur happens accompanied with volume contraction due to the formation of nickel sulphides. Meanwhile, the slight wrinkles form due to the adhesion and friction forces (Supplementary Fig. 7j), which is simulated by varying LJ parameters. Next, the graphene shell was anchored onto the nanoparticle due to the chemical bonding of cg-ni 3 S 2, which has been demonstrated by XPS results. Stronger chemical bonding interaction between graphene and nickel sulphides can effectively prevent the sliding and collapse of crumpled structure, which leads to the assembly of deeply crumpled graphene (Supplementary Fig. 7k, l). Volume variation and chemical bonding formation were simulated by varying the values of LJ parameters σ and ε. Subject to these considerations, MD results show the crumpling formation process on an atomic scale, which is also consist with the experimental observations (Supplementary Fig. 7a-h). Linear fitting of EIS results to evaluate diffusion performance The lithium ion diffusion coefficients (D Li ) of Ni 3 S 2 @GO/Ni and Ni 3 S 2 /Ni before and after cycling were calculated according to the following equations 2 : D Li = R 2 T 2 /2A 2 n 4 F 4 C 2 σ 2 (1)
Z = R D + R L +σω -1/2 (2) where D Li is the lithium ion diffusion coefficient, R is the gas constant, T is the absolute temperature, A is the surface area of the cathode, n is the number of electrons per molecule during oxidization, F is the Faraday constant, C is the concentration of lithium ion, and σ is the Warburg factor which could be obtained by Eq. 2. By linear fitting of Z and ω -1/2, the slope can be described as σ value. Combining with Eq. 1, the square of σ value have an inverse relationship with D Li. Discussion on the effect of mixed phases According to previous report 6, Ni 3 S 2 and NiS behave differently during electrochemical reaction. The electrochemical reaction of Ni 3 S 2 was thought as a one-step process: Ni 3 S 2 + 4Li 3Ni + 2Li 2 S, while NiS follows two-step reactions: 3NiS + 2Li Ni 3 S 2 + Li 2 S, Ni 3 S 2 + 4Li 3Ni + 2Li 2 S. The mixed phases of Ni 3 S 2 and NiS both can be detected in Ni 3 S 2 @cg/ni and Ni 3 S 2 /Ni (Fig. 5a, b). Therefore, the mixed phases contribute to the electrochemical performance of both Ni 3 S 2 @cg/ni and Ni 3 S 2 /Ni. Based on the previous reported and our results, it can be concluded that the mixed phases can improve the electrochemical performance in the following two points: First, the mixed phase can provide high grain boundary density thus enhance the interfacial lithium storage, which can result in high initial capacity 4, 5. In our work, high initial capacity can be achieved in both Ni 3 S 2 @cg/ni and Ni 3 S 2 /Ni at low rate (0.5 A g -1 ) due to the interfacial lithium storage of NiS and Ni 3 S 2. With charge/discharge cycles improving, the capacity of Ni 3 S 2 /Ni fades quickly due to the large phase change and aggregation during electrochemical process when lack of crumpled graphene encapsulation. Second, kinetic properties can be improved by constructing mixed phases to vary the spacing of
interfaces down to the nanoscale regime, in which the grain boundaries act as channels to Li + transport 5, 6. In our work, relatively high Li + diffusion coefficient (D Li ) can be obtained before the charge/discharge cycling without crumpled graphene encapsulation (Fig. 3i), which can be attributed to the Li ion transport channels provided by abundant grain boundaries. After 100 cycles, large phase change seals the ion transport channels and yields the fast fading of D Li (Fig. 3i). Supplementary References 1. Palacin, M.R. Recent advances in rechargeable battery materials: a chemist s perspective. Chem. Soc. Rev. 38, 2565-2575 (2009). 2. Liu, Z., et al. Low-temperature behavior of Li 3 V 2 (PO 4 ) 3 /C as cathode material for lithium ion batteries. J. Sol. St. Electrochem. 16, 1917-1923 (2012). 3. Han, S. C., et al. Charge-discharge mechanism of mechanically alloyed NiS used as a cathode in rechargeable lithium batteries. J. Alloys. Compd. 361, 247 251 (2003). 4. Hu, Y. Y., et al. Origin of additional capacities in metal oxide lithium-ion battery electrodes. Nat. Mater. 12, 1130 1136 (2013). 5. Rahman, M., Wang, J.Z., Hassan, M.F., Wexler, D. & Liu, H.K. Amorphous carbon coated high grain boundary density dual phase Li 4 Ti 5 O 12 -TiO 2 : a nanocomposite anode material for Li-ion batteries. Adv. Energy Mater. 1, 212-220 (2011). 6. Zhang, J.-J., Wei, Z., Huang, T., Liu, Z.-L. & Yu, A.-S. Carbon coated TiO 2 -SiO 2 nanocomposites with high grain boundary density as anode materials for lithium-ion batteries. J. Mater. Chem. A 1, 7360-7369 (2013).