*2.1. Electrolytic Conditions to Synthesize High-Purity, High-Yield CNTs from CO2*

The first part of this study systematically explores electrochemical parameters to reveal a wide variety of conditions that yield a high-purity, high-yield CNT product by means of electrolysis of CO2 in 770 ◦C lithium carbonate. An in depth look at the material composition and morphologies of the products is conducted, particularly around the transition metal nucleation zone of CNT growth. The latter part of this study reveals molten electrochemical conditions that produce macroscopic assemblies of CNTs. This study also serves as a sister study [51] in which small electrolytic changes in the 770 ◦C molten Li2CO3 yield major changes to the product consisting of new, non-CNT nanocarbon allotropes.

Previously, we showed that the high production rate (using a high electrolysis current density, J, of 0.6 A/cm2) electrolytic splitting of CO2 in molten Li2CO3 electrolyte using a Muntz Brass cathode (60% Cu and 40% Zn) and a Nichrome C (60% Ni, 24% Fe and 16% Cr) anode produced a high-quality (97% purity), high-aspect-ratio carbon nanotube (CNT) product with the addition, to the electrolyte, of either 0.1 wt% Fe2O3 [43] or 2.0 wt% Li2O [41,44]. The addition of higher concentrations of either iron or lithium oxide to the electrolyte increased the formation of defects in the CNTs, as measured by Raman spectroscopy, which at this higher current density induced spiraling of the CNT during growth and the controlled growth of a variety of helical carbon nano-allotropes including single- and double-braided helices was observed, as well as flat, spiral morphologies [43,44].

Here, conditions related to the high-purity CNT synthesis were systematically varied to determine other electrochemical conditions that support the high-purity, low-defect formation of straight (non-helical) CNTs. Examples of the conditions that were varied include the composition of the cathode and anode, the additives to the lithium carbonate electrolyte, and the current density and time of the electrolysis. Variations of the electrodes include the use of cathode metal electrodes such as Muntz brass Monel, or Nichrome alloys. Anode variations include noble anodes such as iridium, various nickel-containing anodes including nickel, Nichrome A or C, Inconel 600, 625, or 718, or specific layered combinations of these metals. Electrolyte additives that were varied include Fe2O3, and nickel or chromium powder, and electrolyses were varied over a wide range of electrolysis current densities. Several electrolyses studied here, which yielded high-purity, high-yield carbon nanotubes, are described in Table 1. Scanning electron microscopy (SEM) of the products of a variety of those CNT syntheses was conducted by CO2 electrolysis in molten Li2CO3 at 770 ◦C, and the results are presented in Figure 2.

**Table 1.** A variety of Electrolytic CO2 splitting conditions in 770 ◦C Li2CO3 producing a high yield of carbon nanotubes.


**Figure 2.** SEM of the synthesis product of high-purity, high-yield carbon nanotubes under a variety of electrochemical conditions by electrolytic splitting of CO2 in 770 ◦C Li2CO3. The washed product was collected from the cathode subsequent to the electrolysis described in Table 1. Moving left to right in the panels, the product was analyzed by SEM with increasing magnification. Scale bars in panels (starting from left) are for panels A: 100, 50, and 10 μm; for panels B: 100, 20, and 5 μm; for panels C: 40, 5, and 2 μm; for panels E: 200, 40, and 10 μm.

For Electrolysis A, the top row of Table 1 presents the electrochemical conditions, and the top row of Figure 2 presents the SEM results of the product of a repeat of the electrochemical conditions of the described 0.1 wt% Fe2O3 electrolysis (the same lithium carbonate electrolyte, the same Muntz Brass cathode and Nichrome C anode, and the same 0.6 A/cm2 current density and 30 min electrolysis duration), but used a simpler (from a material perspective) alumina (ceramic Al2O3), rather than stainless steel 304 electrolysis cell casing. Use of the alumina casing in this study limited the pathways for metals to enter, thereby the reducing parameters to evaluate, and possibly effect, the electrolytic system. Note, however, that the stainless steel 304 was not observed to corrode, and the switch

from stainless to alumina was not observed to materially affect the electrolysis product. The CNT product was again 97% purity, and the coulombic efficiency was 99%, which quantified the measured available charge (current multiplied by the electrolysis time) to the measured number of four electrons per equivalent of C in the product, and the carbon nanotube length was 50 to 100 μm.

The second row of Figure 2 (panels B) represents a change only in the current density, which was lowered to 0.15 A/cm2, and the electrolysis time, which was increased to 4 h, and the result was a decrease in product purity to 94%, a decrease in CNT length to 20–80 μm, and a modest decrease in coulombic efficiency to 98%. At this current density, as observed in the third row of Figure 2, panels C, the addition of 0.1 wt% Ni along with the 0.1 wt% Fe2O3 resulted in 96% purity, zigzag-patterned and twisted, rather than straight CNTs. These twists could be induced by over-nucleation decreasing control of the CNT linear growth. In the most magnified of these product images (right side of the Figure, 2 μm bar resolution), evidence of the over-nucleation can be observed in the larger nodules visible at the CNT tips and joints.

At a low current density of 0.08 A/cm2, with an electrolyte additive of 0.1 wt% Fe2O3, the conventional Muntz Brass and Nichrome electrodes exhibited a significant drop in CNT product purity to 70%. Coulombic efficiency tended to drop off with current density, and, in this case, the coulombic efficiency of the synthesis was 82%. Product purity could be increased by refining the mix of transition metals available to the electrolytes or increasing the surface area. The alloy compositions of the metals used as electrodes are presented in Table 2. The metal variation was further refined by combining the metals in Table 2 as anodes; for example, using a solid sheet of one Inconel alloy, layered with a screen or screens of another Inconel alloy. This approach was utilized to obtain the results shown in the lowest row of Figure 2 (panels E); an anode of Inconel 625 with three layers of (spot welded) 100-mesh Inconel 600 screen, a return to a single electrolyte additive (0.1 wt% Fe2O3), and a very low current density of 0.08 A/cm2 were utilized. As seen in panels E of the figure, the product was high purity (97%) and consisted of 20–50 μm length CNTs, and the coulombic efficiency was 75%. Not shown, but included in Table 1 (Electrolysis G), are the results obtained under the same electrode, and the same 0.08 A/cm<sup>2</sup> electrolysis conditions. However, with the electrolyte addition of both 0.1 wt% Fe2O3 and 0.1 wt% Ni at J = 0.15 A/cm2, the product was twisted CNTs as shown in Figure 2 panels C, the purity was 96%, and the coulombic efficiency was 80%.


**Table 2.** Compositions of various alloys used (weight percentage).
