Among all the possible secondary battery systems, Li metal-based batteries have attracted increasing interest in recent decades due to the unapproachable theoretical specific capacity (3860 mAh g-1) and low redox potential (-3.04 V vs. NHE) of Li. However, the growth of lithium dendrites will cause security problems which restrict the application of lithium metal. According to previous work, the current collector could play an important role in the protection of the lithium metal. In this work, Here we demonstrate that a hollow carbon nanofiber with proper interior to exterior radius ratio can enable Li-ions to deposit on the inner surface of the channels selectively due to the drifting effect from the structural stresses. Based on this principle, a lotus-root like structure is further designed to realize a dendrite-free hybrid Li anode with a high Li loading capability. The lotus-root like carbon nanofiber (LCNF) anode, with being coated by a lithiated Nafion (LNafion) layer as artificial solid electrolyte interface (SEI), achieves a capacity of >3600 mAh gcarbon-1 for Li deposition/stripping along with a greatly improved CE.
Result and discussion
Fig. 1 shows the schematic diagrams of LCNF and the solid carbon nanofiber (SCNF, the sample prepared via the same route as LCNF but without PS addition in the precursor solution) before and after Li deposition. The LCNF was prepared by a simple electrostatic spinning route followed by high-temperature carbonization process. For the carbonizing process of PAN, it is widely accepted that PAN undergoes a cyclization reaction and forms a highly conjugated structure, which rending the resulting material both insoluble and infusible. At the PAN/PS ratio of 1:0.5, the LCNT with more than 10 parallel channels (~50 nm) was obtained. The transmission electron microscopy (TEM image (Fig. 1F) clearly demonstrate that the longitudinal straight multi-channels are homogeneously distributed in carbon nanofibers parallelly. Carbon nanofibers with uniform size (~500 nm) are intertwined into the freestanding LCNF paper, and the external surface of the fibers is smooth and clean (Fig. 1J). Compared with LCNF, SCNF has no channels inside (Fig. 1E) and appears predisposed to fracture (Fig. 1I). Fig. 2G and 2H show the TEM images of Li-LCNF and Li-SCNF (Li deposition state, after 30 cycles at a current density of 1 mA cm-2 for 8 mAh cm-2 (it amounts to 3640 mAh gcarbon-1 or 0.94 g Li infilled 1.0 g carbon calculated based on the areal density of LCNF of 2.2 mg cm-2) . It can be clearly seen that a large amount of Li uniformly deposits within the inner channels while the outer surface of the nanofibers is relatively clean (Fig. 1L). The diameter of the Li-filled LCNF keeps almost the same as that of the original LCNF before Li deposition. As expected, Li ions are preferentially deposited on the inner surface of the channels, and thus the outer surface of the Li-LCNF anode is free of Li dendrites. In sharp contrast, for the Li-SCNF sample with the same amount of Li deposition, Li ions only deposits on the outer surface (Fig. 1G), and lots of mossy deposits can be observed (Fig. 1K).
In summary, a new strategy to control Li deposition via matrix geometry design is proposed. We demonstrate that a hollow carbon nanofiber with proper interior to exterior radius ratio can control Li deposition via the drifting effect from the structural stresses. The lotus-root like carbon nanofiber matrix we further developed enables Li preferentially to deposit on the inner surface successfully and therefore the stable carbon shell provides a strong barrier to suppress Li dendrite. Furthermore, the multichannel structure also endows the LCNF anode with a high Li loading capability due to its abundant inner hollow space. With the aid of the Nafion SEI, the Li-LCNF anode realizes a high specific capacity of >3600 mAh gcarbon-1 for Li deposition/stripping along with a greatly improved CE. The CE reaches over 99% and maintains in the following long-term cycles.