*3.3. Product Characterization*

The raw product was collected from the cathode after the experiment and cool down, followed by an aqueous wash procedure that removed electrolyte congealed with the product as the cathode cooled. The washed carbon product was separated by vacuum filtration. The washed carbon product was dried overnight in a 60 ◦C oven, yielding a black powder product.

The coulombic efficiency of electrolysis is the percent of applied, constant current charge that was converted to carbon, determined as:

$$100\% \times \text{C}\_{\text{expperimmental}} / \text{C}\_{\text{thecovritical}} \tag{4}$$

This was measured by the mass of washed carbon product removed from the cathode, Cexperimental, and calculated from the theoretical mass, Ctheoretical = (Q/nF) × (12.01 g C mol<sup>−</sup>1), which was determined from Q, the time-integrated charged passed during the electrolysis, F, the Faraday (96485 As mol−<sup>1</sup> e−), and the n = 4 e<sup>−</sup> mol−<sup>1</sup> reduction of tetravalent carbon, consistent with Equation (2).

Characterization: The carbon product was washed, and analyzed by PHENOM Pro Pro-X SEM (with EDX), FEI Teneo LV SEM, and by FEI Teneo Talos F200X TEM (with EDX). XRD powder diffraction analyses were conducted with a Rigaku D = Max 2200 XRD diffractometer and analyzed with the Jade software package. Raman spectra were collected with a LabRAM HR800 Raman microscope (HORIBA). This Raman spectrometer/microscope used an incident laser light with a high resolution of 0.6 cm−<sup>1</sup> at 532.14 nm wavelength.

#### **4. Conclusions**

Molten carbonate electrolysis of CO2 provides an effective path for the C2CNT synthesis of CNTs and macroscopic CNT assemblies. This study explored a variety of electrochemical configurations, systematically varying the electrode composition, electrode current density, electrolysis time, current ramping initiation, and electrolyte additives and their concentrations. The highest observed CNT purity synthesis (97%) utilized a specialized anode consisting of two layers of high-surface-area Inconel 600 (screen) on Inconel 718, a Muntz Brass cathode, with an 0.1 wt% Fe2CO3 additive to the 770 ◦C Li2CO3 electrolyte, and with the electrolysis current conducted for 4 h at an intermediate current density (without current ramping) of 0.15 mA/cm2. The product, as analyzed by SEM, was aligned CNTs with a length of 100 to 500 μm and an aspect ratio of over 1000. The anode, cathode, and electrolyte additive choice were found to be effective for controlling the transition metal nucleation, which is critical to high-purity electrolytic CNT growth.

In a sister paper [51], slight variations of the synthesis parameters led to the formation of a variety of new, high-purity, non-CNT nanocarbon allotropes. All syntheses in the present study produced a majority of CNTs, but the morphology of the CNT product changed widely with the synthesis conditions. Depending on the synthesis conditions, alternate CNT products that were as short as 10 to 30 μm, and curled, rather than straight, or mixed with carbon nano-onions were observed. The high-purity product exhibited a sharp XRD graphic peak, and a low Raman ID/IG ratio, which was indicative of low defects in the carbon structure. The XRD also contained iron carbide, and nickel and chromium lithium oxides, which, based on TEM, were found to be located within the CNT.

TEM HAADF showed that the inner core of the CNT length was generally free of metals (void, with 100% carbon walls), but in other areas, the void was filled with transition metal. As they were produced by molten carbonate electrolysis, the CNT walls were conclusively shown to be comprised of highly uniform concentric, cylindrical graphene layers with graphene characteristics and inter-layer spacing of 0.33 to 0.34 nm. When the internal transition metal was within the CNT tip, the layered CNT graphene walls were observed to bend in a highly spherical fashion around the metal supporting the transition metal nucleated CNT growth mechanism. Several syntheses had unusual nodules, many of them highly spherical, on the CNT, generally comprising carbon and containing a low level of internal transition metal. Generally, intersecting CNTs did not merge, but in a few cases, graphene layers bent to become part of the CNT intersection, which was consistent with the occurrence of the occasional, related growth of intersecting CNTs, such as branching.

The study also demonstrates new syntheses of assemblies of CNTs by means of the C2CNT process, with structural implications towards their potential applications for nanofiltration and neural nets and demonstrated pores sizes ranging from 50 nm to 1 μm.

**Author Contributions:** Conceptualization, S.L. and G.L.; methodology, S.L., X.L., G.L. and X.W.; writing S.L. and G.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** C2CNT LLC funded this research through the C2CNT LLC XPrize support funding.

**Data Availability Statement:** The authors confirm that the data supporting the findings of this study are available within the article.

**Conflicts of Interest:** The authors declare no conflict of interest.
