Switching of alternative electrochemical charging mechanism inside single-walled carbon nanotubes: a quartz crystal microbalance study

We probed electrochemical ion storage in single-walled carbon nanotubes (SWCNTs) of different diameters in two different organic electrolytes using electrochemical quartz crystal microbalance (EQCM) tracking. The measurements showed that charge storage probed by cyclic voltammetry did not deteriorate when steric effects seemed to hinder the accessibility of counter-ions into SWCNTs, and instead proceeded predominantly by co-ion desorption, as was shown by the decrease in the electrode mass probed by EQCM. The dominant mechanism correlated with the SWCNT diameter/ion size ratio; counter-ion adsorption dominated in the whole potential range when the diameter of SWCNTs was comparable to the size of the largest ion, whereas for larger diameters the charge increase coincided with a decrease in the electrode mass, indicating the dominance of co-ion desorption. The dominance of co-ion desorption was not observed in activated carbon, nor was it previously reported for other carbon materials, and is likely switched on because the carrier density of SWCNT increases with applied potential, and maintains the electrode capacity by co-ion desorption to overcome the steric hindrances to counter-ion adsorption.

Nitrogen adsorption isotherm measurements (SHIMADZU Gemini2375 analyzer) were used to determine the Brunauer-Emmett-Teller specific surface area (S BET ) of the samples, and to confirm the success of the decapping treatment. The S BET values obtained before and after the decapping treatment (Table S1) show the increase in the surface area upon decapping.
The increase in the surface area can be deemed to originate solely from the opening of the end caps if the acid and heat treatments do not lead to the creation of defects on the walls of SWCNTs.
This can be judged from the Raman spectra of the samples before and after decapping. Raman spectroscopy was performed using a JASCO NRS-3300 spectrometer with 532 nm-wavelength Nd:YAG excitation source. The obtained spectra are shown in Fig. S1. The peak around 1600 cm -1 is known as the G band, which is used as an indicator on the degree of graphitization of the carbon network of the SWCNT wall. The peak around 1350 cm -1 is the disorder-induced D band, whose intensity increases with the increase in the defects, impurities, and domain boundaries in the graphitic network. 1 All the spectra show very low D-band intensity that either remained unchanged or improved (for SWCNT2.5), which means that the samples had high crystallinity that they maintained at the end of the treatments.
The low frequency region (the left panel) of Fig. S1 shows the peak known as the radial breathing mode (RBM), whose vibration frequency is known to correlate with the diameter of the SWCNTs being probed. We estimated the diameter range of SWCNTs in the three samples from the RBM peaks using the following equation: Here, ω RBM is the frequency of the RBM peak, A = 234 cm -1 , and B is the upshift in ω RBM caused by tube-tube interactions, and has the value 10 cm -1 . 2 The peak location values were obtained from fitting the spectra with a Lorentzian function using Fityk software. The instrumental limitations of the spectrometer used in our study prevented the extension of the measureable Raman shift below 100 cm -1 , and that caused the spectrum for the sample SWCNT2.5 to show only a partial peak shape in the measurable RBM region. Therefore, the diameter values obtained for that sample were estimated from the extrapolation of the partial peak shape through the fitting process. The calculated diameter ranges are given in Table S1.  (111) double-crystal monochromator. XRD diffraction data and simulations were used to estimate the mean tube diameter and interstitial radius for each sample assuming the triangular bundle lattice shown in Fig. S2. 3 The obtained values are given in Table S1.  As example on the unique in electronic structure of SWCNTs, we obtained candidate chiralities corresponding to the mean tube diameter of each sample, and plotted their electronic density of states in Fig. S3, Fig. S4, and Fig. S5. The quartz crystal microbalance system (AT-cut QCM922A microbalance from SEIKO EG&G Co., Ltd.; basic frequency 9 MHz) was used in parallel with cyclic voltammetry (SP-50CN potentiostat, Bio-Logic Science Instruments). To prepare the electrodes, the sample under evaluation was dispersed in ethanol, then drop-cast on a 5-mm diameter platinum-coated quartz substrate using a micropipette at the rate of 0.5 -1 μL. The substrate was placed on a hotplate set to 90 °C to allow the ethanol to evaporate. The weight for each sample is given in Table S1. The SWCNT-coated substrate was used as the working electrode in a three-electrode polyether ether ketone (PEEK) cell (QA-CL4PK, SEIKO EG&G Co., Ltd), with a platinum mesh as the counter electrode and Ag/Ag + as the reference electrode. The cell was left overnight after assembly to ensure good wetting of the electrode materials. The cell was connected to the EQCM system and the potentiostat simultaneously to measure the frequency shift during cyclic voltammetry in the potential range between -2.0 and +1.0 V at a scan rate of 10 mV/s. The measurement setup is shown in Fig. S7.
The change in the electrode mass was calculated using the Sauerbrey equation: … (S2) ∆ =-2 2 0 ∆ In the equation above, ∆F is the change in frequency in Hz, F 0 is the initial frequency, A E is the area of the electrode in (A=0.196 cm 2 ), μ is the crystal modulus of elasticity (2.947×1011 g/cm.s 2 ), and ρ is the crystal density (2.648 g/cm 3 ).

Fig. S7
Cell configuration and experimental setup for simultaneous CV and EQCM measurements