Design of Uniform Hollow Carbon Nanoarchitectures: Different Capacitive Deionization between the Hollow Shell Thickness and Cavity Size

Abstract Carbon‐based materials with high capacitance ability and fast electrosorption rate are ideal electrode materials in capacitive deionization (CDI). However, traditional carbon materials have structural limitations in electrochemical and desalination performance due to the low capacitance and poor transmission channel of the prepared electrodes. Therefore, reasonable design of electrode material structure is of great importance for achieving excellent CDI properties. Here, uniform hollow carbon materials with different morphologies (hollow carbon nanospheres, hollow carbon nanorods, hollow carbon nano‐pseudoboxes, hollow carbon nano‐ellipsoids, hollow carbon nano‐capsules, and hollow carbon nano‐peanuts) are reasonably designed through multi‐step template method and calcination of polymer precursors. Hollow carbon nanospheres and hollow carbon nano‐pseudoboxes exhibit better capacitance and higher salt adsorption capacity (SAC) due to their stable carbonaceous structure during calcination. Moreover, the effects of the thickness of the shell and the size of the cavity on the CDI performance are also studied. HCNSs‐0.8 with thicker shell (≈20 nm) and larger cavity (≈320 nm) shows the best SAC value of 23.01 mg g−1 due to its large specific surface area (1083.20 m2 g−1) and rich pore size distribution. These uniform hollow carbon nanoarchitectures with functional properties have potential applications in electrochemistry related fields.

wt %) under vigorous stirring at room temperature. After 15 min, 0.4 g of resorcinol (R) and 0.56 mL of formaldehyde (F) were added to the solution, and the system was kept under vigorous stirring for 24 h at room temperature. The precipitates were separated by centrifugation, washed with deionized water and ethanol, and then dried at 50 o C overnight. Second, the precipitates were calcinated under N2 at 700 o C for 2 h to obtain the SiO2@CNSs. The obtained SiO2@CNSs were etched in 3M NaOH solution to obtain HCNSs. This sample is also named HCNSs-0.4. Same as the above steps, in order to prepare HCNSs with different shell thicknesses, the amount of R and F is changed to 0.2 g, 0.28 mL (HCNSs-0.2); 0.6 g, 0.84 mL (HCNSs-0.6). Similarly, in order to prepare HCNSs with different cavity sizes, the quantities of TEOS, R and F are changed to 0.865 mL, 0.1 g, 0.14 mL (HCNSs-0.1); 6.92 mL, 0.8 g, 1.12 mL (HCNSs-0.8).

Synthesis of hollow carbon nanorods (HCNRs)
MnOx nanowires were fabricated according to previous report. Briefly, 190 mg of KMnO4 and 100 mg of PVP were dissolved in 80 mL of H2O. After vigorous magnetic stirring at room temperature for 30 min, the resulting solution was transferred into a 100 mL Teflon-lined autoclave, which was sealed and maintained at

Synthesis of hollow carbon nano-pseudoboxes (HCNBs)
Pseudocubic Fe2O3 was fabricated according to previous report. A NaOH solution (90 mL, 6.0 M) was added to 100 mL of well-stirred 2.0 M FeCl3·6H2O in a 250 mL Pyrex bottle for 30 min. The tightly stoppered bottle containing the Fe (OH)3 gel was placed in a laboratory oven preheated to 100 °C, and the gel was aged for 8 days.
After the treatment, red products were collected by filtration and washed three times with deionized water and ethanol before drying at 50 °C overnight.

Synthesis of hollow carbon nano-ellipsoids (HCNEs)
The process of preparing ellipsoidal Fe2O3 is similar to that of preparing pseudocubic The tightly stoppered bottle containing the Fe (OH)3 gel was placed in a laboratory oven preheated to 100 °C, and the gel was aged for 8 days. After the treatment, red products were collected by filtration and washed three times with deionized water and ethanol before drying at 50 °C overnight. Later, the process of preparing HCNEs is the similar to that of preparing HCNBs. The only difference is that the ellipsoidal Fe2O3 template replaces the pseudocubic Fe2O3 template.

Synthesis of hollow carbon nano-capsules (HCNCs)
The process of preparing rodlike Fe2O3 is similar to that of preparing ellipsoidal Fe2O3. The only difference is that the added Na2SO4 solution concentration is 0.60 M.
Later, the process of preparing HCNCs is the similar to that of preparing HCNBs. The only difference is that the rodlike Fe2O3 template replaces the pseudocubic Fe2O3 template.

Synthesis of hollow carbon nano-peanuts (HCNPs)
The process of preparing peanut like Fe2O3 is similar to that of preparing ellipsoidal Fe2O3. The only difference is that the added Na2SO4 solution concentration is 1.00 M.
Later, the process of preparing HCNPs is the similar to that of preparing HCNBs. The only difference is that the peanut like Fe2O3 template replaces the pseudocubic Fe2O3 template.
Raman spectroscopy was obtained by using Renishaw InVia Reflex (514 nm laser).
The chemical states are measured using an Axis Ultra X-ray photoelectron spectroscope (XPS, Kratos Analytical Ltd., UK) equipped with a standard monochromatic Al-Kα source (hv = 1486.6 eV). Nitrogen sorption isotherms were carried out using a BELSORP-mini (BEL, Japan). The specific surface area (SSA) was analyzed by Multipoint Brunauer-Emmett-Teller (BET) technique.

Electrochemical performance measurements
The electrode ink was prepared by mixing 80 wt% active material with 10 wt% Vulcan XC 72 and 10 wt% PVDF in NMP solvent under ultrasonication for 30 min. A certain volume of the ink was dropped onto the graphite paper with a thickness of 0.5 mm (area of 1 × 1 cm 2 ) and dried at 60 °C for 12 h. The potential sweep cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) measurements were conducted by using an electrochemical workstation (CHI-760E) with three-electrode configuration in 1.0 M NaCl electrolyte. The Ag/AgCl electrode and platinum (Pt) wire were used as reference and counter electrodes, respectively. Cyclic voltammetry (CV) and gravimetric charge-discharge (GCD) measurements were carried out in the potential range of -1 to -0.1 V.
The specific capacitance of electrodes was calculated using Eq. (S1), where Cm (F g -1 ) represents for the specific capacitance of the electrode, I (A) for the current density, ∆V for the voltage change, v (mV s -1 ) for the scan rate and m (g) for the mass of the working electrode.

CDI performance measurements
Membrane assisted CDI (MCDI) unit-cell was constructed with two pairs of identical electrodes, ion exchange membrane for anion and cation, and spacer. Anion-and cation exchange membranes were used to alleviate the co-ion effect. Each individual CDI carbon electrode was composed of active material, Vulcan XC 72, and PVDF in the ratio of 8:1:1, and prepared on the titanium plate of 2×2 cm 2 (thickness: 1 mm) as current collector. Before assembling the MCDI unit-cell, the electrodes were immersed in 584 mg/L of saline water for 24 h to completely wet the surface. The CDI tests were conducted using a batch-mode with a continuous recycling system, which includes a CDI cell, a peristaltic pump, a power source, and a fluid reservoir.
The ambient temperature and the total volume of the NaCl solution in the desalination experiment were maintained at 298 K and 50 mL, respectively. In the CDI desalination process，the saline water was desalinated through MCDI unit-cell and recycled in a closed circuit. The real-time change of the brackish water concentration was measured by a conductivity probe which was connected to the CDI system. The correlation between conductivity and concentration was achieved based on a calibration table prepared before the test (Figure S31). The original concentration of NaCl aqueous solution used in the desalination system is 10 mM (corresponding to 6 adsorption capacity (SAC, mg g -1 ) and average salt adsorption rates (ASAR, mg g -1 min -1 ) at t min were calculated as follows: Eq. (S2)

Eq. (S3)
where C0 and Ct are the NaCl concentrations at initial stage and t min, respectively