**4. Conclusions**

Known conventional carbon allotropes include diamond and graphite, and more recently buckyballs, graphene, carbon nanofibers and CNTs. In this study, the range of carbon allotropes and morphologies, and in particular nanocarbon morphologies, that can be synthesized by the molten carbonate electrolysis of CO2 has been greatly expanded. Fascinating high purity morphologies that have been obtained in this study by the systematic variation of electrolysis conditions are conical CNF, nano-bamboo, nano-pearl, Ni-coated CNT, nano-flower, nano-dragon, nano-rod, nano-belt, nano-onion (also previously synthesized by an alternative methodology in [25]), hollow nano-onions and nano-trees. Each of these CNMs have their unusual and distinctive morphologies, such as the nano-trees with their branching CNT structure, or the nano-bamboo and -pearl with their different, but repeated knob or bulb shapes. These distinctive morphologies may lead to unusual physical chemical properties with implications useful to applications, such as those utilizing the high

strength, high thermal, magnetic, electronic, piezoelectronic, tribological characteristics of graphene-based materials, but which distribute these properties differently throughout the unusual geometries of this novel allotropes. For example, alternative applications such as high-capacity Li-anodes, unusual electronics, EMF shielding, improved lubricants, and new structural or polymer composites may be anticipated.

This study has explored a variety of electrochemical configurations, systematically varying electrode composition, current density and electrolysis time, current ramping initiation, and variation of electrolyte additives and their concentrations. The observed nanocarbon structures were analyzed by SEM, TEM, including with HAADF, Raman and XRD. With the exception of the nano-rod structure, each of the structures is graphitic in nature, containing graphene layers arranged in a variety of geometries. The graphene layers exhibit the characteristic inter-layer spacing of 0.33 to 0.34 nm. Except for the presence of Ni, Fe, Cr and occasionally Cu, which may serve as nucleating growth sites, each of the structures is pure carbon. Generally, intersecting graphene layers did not merge, but in the nano-tree, the graphene layers bend at intersections leading to the observed branching.

Many of the structures including nano-bamboo, nano-pearl, Ni-coated CNTs and conical CNFs exhibit walls containing concentric graphene layers. The nano-dragon and nano-belt structures include layered planer or planar-twisted graphene layers. Several of the observed structures, including nano-trees, and hollow and filled nano-onions, exhibit concentric, highly spherical graphene layers generally composed of carbon and containing a low level of internal transition metal. A new pathway to the formation of nano-onions via phosphate addition to the electrolyte is demonstrated, and the hypothesis that phosphate selectively binds transition metal ions should be pursued.

Each of the syntheses were conducted in a 770 ◦C Li2CO3 electrolyte with or without various additives, on a variety of metal or metal alloy electrodes, and with a range of current densities. In a sister paper [29], slight variations of these same synthesis parameters form high-purity carbon but only with the CNT structure. The varied anode and cathodes contained either pure Ni, or mixes also including Fe and Cr, or various mixes of an extended variety of transition metals. However, the changing conditions led to variations of the CNT morphology (length, diameter, curled or straight, added defects etc.). All syntheses in the study from Electrolysis IV onward produced a high-purity product of the stated structure, with the exception of the conical CNFs that were a minority (6%) within a majority of nano-bamboo carbon, and the moderate purity (85%) nano-belt carbon product. Coulombic efficiency of the electrolyses ranged from 79 to 80% at lower current densities of 0.08 A/cm2, to over 99% at current densities of 0.2 A/cm2 or higher. The high purity products each exhibited sharp XRD graphic peaks, and a moderate (0.3 to 1.3) Raman ID/IG ratio indicative of a moderate level of defects in the structure. In addition to a majority of pure, graphitic carbon, the XRD also exhibited different singular or mixed transition metal salts of either iron carbide, or nickel, chromium or copper lithiated oxides.

TEM HAADF of the new nanostructures shows that their inner core is generally metal-free (void, with the walls 100% carbon), but in other areas, the void is filled with transition metals: Ni, Fe and/or Cr. Except for the nano-rod product, each of the structures included distinct graphene layers with a graphene characteristic, inter-layer spacing of 0.33–0.34 nm. Depending on the nanostructure, adjacent graphene layers were organized either in a planer, cylindrical or spherical geometry. When the internal transition metal is in the tip, the layered graphene walls are observed to bend in a highly spherical fashion around the metal supporting the transition metal nucleated CNT growth mechanism. The use of a Ni-anode, or an excess of added Ni to the electrolyte, leads to Ni-coated CNTs when stainless steel is used as the cathode. Generally, intersecting graphene layers did not merge, but in the nano-tree allotrope, graphene layers bend to become part of a CNT intersection consistent with branching.

Molten carbonate electrolysis of CO2 provides an effective path for the synthesis of a portfolio of unusual, valuable nanocarbon allotropes. Mass production of these structures from CO2 will provide a valuable incentive to consume this greenhouse gas. Such structures are rare, or were previously non-existent, and are not generally commercially available. However, those that are in use, such as nano-onions made by pyrolysis of nanodiamonds [35], or by CVD, have high carbon footprints and associated costs at over USD 1 million/ton. CNT production by the molten carbonate electrolysis of CO2, the C2CNT process, is an inexpensive synthesis comparable to the cost of aluminum oxide splitting in the industrial production of aluminum [14]. The scale-up of this process was awarded the 2021 Carbon XPrize XFactor award for producing the most valuable product from CO2 [31,32]. The new synthesis conditions consist of small variations of the scaled C2CNT process with a comparable, straightforward path to scale up to contribute to consumption of CO2 for climate change mitigation.

**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.
