A typical TEM image of the S-doped Fe@C core-shell nanoparticles is shown in Fig. 2(a). Their diameters are ranged from 10-40 nm and the thickness of the shell is 1-3 nm. According to XRD results (Fig. 2(b)), the core of the particles is Fe₃C and FeS. The peaks at around 29.8°, 33.7°, 3.2° and 53.2° can be assigned to the reflections of FeS, which forms from pyrolysis of iron pentacarbonyl and carbon disulfide, and the rest of the peaks are related to the Fe₃C phase. After acid-treatment, the ferrous cores (Fe₃C and FeS) are eliminated, leaving hollow cage structure as shown in Fig. 2(c) and their graphitic shells are enclosed without distinguished nanopores. In the XRD pattern (Fig. 2(d)), the peaks appearing at 26.2°, 44.4° and 78.2° result from diffraction at the (002), (101) and (110) planes of graphite, indicating graphitization of nanocages is high and removal of ferrous cores is complete.

Fig. 2

	Figure Option ViewDownloadNew Window
	Fig. 2 TEM image (a) and XRD pattern (b) of the S-doped Fe@C core-shell nanoparticles, HRTEM image (c) and XRD pattern (d) of the S-doped GNCs prepared from removing the ferrous cores by acid-treatment

XPS is employed to analyzed structure change of sulfur-doping. Before removing the S-doping structure, the acid-treated GNCs shown in Fig. 2(c) have been ana lyzed by XPS to measure content and structure of doping (Fig. 3(a)), which indicates that the sulfur content of the nanocages is 3.3at% and the rest is carbon. The S2p peak (Fig. 3(a)) of sulfur is related to S-doping in graphitic structure rather than FeS cores, and the peaks located between 163 and 164 eV might be attributed to S-C bending^{[ 17]}, while the peak located at ~167.1 eV might indicate existence of sulfonate species^{[ 17, 18]}, due to acid treating Fe@C nanoparticles might oxidize some S-doping structure. To removal of S-doping structure, air-oxidization is more efficient way than heating in vacuum (Fig. 3(a)), as only sulfonate species might transform into SO₂ but some S-C bending is very stable in the rear condition (Fig. 3(b)). When oxidized in air, all S-doping structure can be oxidized to SO₂, leading to complete removal of S-doping (Fig. 3(a)).

Fig. 3

	Figure Option ViewDownloadNew Window
	Fig. 3 (a) XPS spectra of S-doped GNCs before (acid- treating) and after (vacuum-heating or air-oxidization) removing doping structure, (b) XPS S2p spectra of GNCs before and after removing doping structure (S-doping structure of air-oxidized GNCs is completely removed and therefore their S2p spectra is not distinguished)

The specific surface area ( S_BET) and pore volume of graphitic nanocages have been calculated from N₂ adsorption-desorption isotherms (Fig. 4(a)), which shows a higher uptake in the region of low nitrogen pressure (<0.05 P/ P₀) by the air-oxidized nanocages than that by the nanocages without removal of S-doping structure, indicating the air-oxidized nanocages have more micropores and higher S_BET. Furthermore, the air-oxidized nanocages sample has a very pronounced adsorption hysteresis from P/ P₀ = 0.4 to ~0.9 (Fig. 4(a)), suggesting that this sample consist of more mesopores than acid-treated nanocages (Fig. 4(b)). This observation suggests that removal of S-doping structure form lots of nanopores in the shells of the nanocages, leading to their specific surface area increases from 540 cm³/g to 850 cm³/g and their mesopore volume rises from 0.44 cm³/g to 0.9 cm³/g (Fig. 4(a) and (b)).

Fig. 4

Figure Option ViewDownloadNew Window

Fig. 4 (a) Nitrogen adsorption/desorption isotherms showing a much higher uptake of N2 in doping-removed GNCs (air-oxidized GNCs) than S-doped GNCs (Fig. 2(c)). (b) Mesopore size distributions (BJH) showing that there is a much higher population of mesopores in the air-oxidized GNCs than in other ones. (c) Raman spectra of GNCs before and after removing S-doping structure. (d) HRTEM image of the GNCs after removing S-doping structure (nanopores are arrowed)

To investigate the change of graphitic structure, the graphitic nanocages with and without doping structure are analyzed by Raman spectrum (Fig. 4(c)). Both two samples have peaks located near 1324 and 1580 cm^-1, Showing good graphitization. However, the peaks of the graphitic nanocages oxidized in air are sharper than those of the ones without air-oxidization. This is attribute to air-oxidizing S-doping structure to remove such structure, which is one kind of defects in the graphitic layers (Fig. 4(c)).

The HRTEM image of the air-oxidized nanocages shows a lot of nanopores exist in their shells arrowed in Fig. 4(d). According to HRTEM image and mesopore size distribution determined by Barrett-Joyner-Halenda (BJH) method (Fig. 4(b)), lots of mesopores (from 0.44 cm³/g to 0.9 cm³/g) have formed from removal the S-doping template. However, those mesopores are much bigger than the size of single S atom, which is very likely caused by air-oxidation of S-template: the oxidation of the S atoms heats up C atoms located closely to the S atoms and then those C atoms are burned up in air to create mesopores in the graphitic shells. Compared with previous N-template approach, the present S-template approach performs more effective towards enhancing specific surface area (595 m²/g vs 850 m²/g) and mesopore volumes (0.72 cm³/g vs 0.9 cm³/g)^{[ 9]}.

The present GNCs prepared with high specific surface area, mesopore volume and good graphitization can be applied in wide areas as adsorbents, electrode materials, and energy storage media^{[ 1, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14]}. With a bigger inner hollow space, the GNCs can be used as gas or radiopharmaceuticals carriers^{[ 8]}. For instance, as a catalyst support, this material has a better graphitic structure, leading to fast electron-transfer and high stability, and its higher mesopore volume provides better efficient diffusion channels for high utilization of catalysts.