*2.1. Electrolytic Conditions Varied to Synthesize New Nanocarbon Allotropes from CO2*

This study varies conditions in which small electrolytic changes in a 770 ◦C molten Li2CO3 yield major changes to the product consisting of new, non-CNT nanocarbon allotropes. In a parallel study to this new carbon allotropes study, different systematic changes to these same electrolytic parameters have revealed a wide range of electrochemical conditions that synthesize only a high-purity, high-yield product consisting only of various carbon nanotube morphologies. That sister study focuses on the transition metal nucleation zone of CNT growth and also reveals molten electrochemical conditions which produce assemblies of CNTs [29]. Straight CNT type carbon allotropes tended to dominant when either 0.1% wt iron oxide was added or the anode contained a high amount of Fe (Nichrome C: 24%, Inconel 718 18.5%, or Inconel 718 at 18.5%). The electrochemical conditions varied here are composition and/or architecture of the cathode and anode, additives and their concentrations to the Li2CO3 electrolyte, and current density and time of the electrolysis. Electrolyte additives that are varied include Fe2O3 and nickel or chromium powder. Electrolyses are varied over a range of electrolysis current densities. 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. Alloy composition of the metals used as electrodes is presented in Table 1. Metal variation was further refined by combining the metals in Table 1 as anodes, for example using a solid sheet of one Inconel alloy, layered with a screen or screens of another Inconel alloy, such as an anode of Inconel 625 with 3 layers of (spot welded) 100 mesh Inconel 600 screen.


**Table 1.** Compositions of various alloys used (weight percentage).

*2.2. Electrochemical Conditions to Synthesize Bamboo and Pearl Nanocarbon Allotropes from CO2*

We have conducted several thousands of different electrolyses to split CO2 in molten lithium carbonate. A fascinating, but rarely observed, product which occurred in less than 30 of those many electrolyses had nano-morphology analogous to the macro-structure of bamboo, but had been only observed as a low fraction of the total product. Table 2 summarizes the systematic optimization of electrolysis conditions in 770 ◦C Li2CO3 to optimize and maximize the electrolytic formation of this nano-bamboo. A few prior electrolyses producing nano-bamboo were associated with nickel electrodes, or started with ramping up of the current to encourage nucleation. Experiment Electrolysis I in the top row of Table 2 includes both these features including nickel as both the cathode and anode, and (not delineated in the table) an initial 10 min electrolysis at a constant 0.01 and then 0.02 A/cm2, followed by 5 min at 0.04 and then 0.08 A/cm2, after which the constant current electrolysis was conducted at the tabulated 0.2 A/cm2. Nano-bamboo was evident in the product SEM, but constituted a minority (30 wt%) of the total product. As seen in Electrolysis II in Table 2, an increase in the nano-bamboo product is achieved with the direct addition of Ni and Cr powders to the electrolyte, and the anode is replaced by a noble metal (iridium) accompanied by a 5-fold decrease in current density. As noted in the table, this Electrolysis II has the first majority, 60 wt%, of the nano-bamboo product. Coulombic efficiency quantifies the measured available charge (current multiplied by the electrolysis time) to the measured number of four electrons per equivalent of C in the product. Coulombic efficiency tends to drop off with current density, and in this case the coulombic efficiency of the synthesis was 79%.


**Table 2.** Systematic variation of CO2 splitting conditions in 770 ◦C Li2CO3 to optimize formation of nano-bamboo and nano-pearl carbon allotropes.

The low current ramping, pre-electrolysis conditions can have benefits and disadvantages. (1) Low current conditions may support the reduction and deposition of initial graphene layers to facilitate ongoing reduction and growth. In addition, lower current can favor transition metal deposition at the cathode and formation of nucleation sites, at low concentrations compared to carbonate (from CO2) in the electrolyte. The analysis of bound versus free metal cations in the molten electrolyte for a reduction potential calculation has been a challenge. Without Nernst activity and temperature correction, the reduction rest potentials of Ni, Fe, Cr and Cu and CO2 at room temperature are CO2(IV/0) = −1.02, Cr(III/0) = −0.74, Ni(II/0) = −0.25, Fe(III/0) = −0.04, and Cu(II/0) = 0.34. Note, however, that the free activity of tetravalent carbon as carbonate C(IV)O3 <sup>2</sup><sup>−</sup> formed by the reaction of C(IV)O2 with electrolytic oxide in pure molten carbonate solutions is many orders of magnitude higher than the dissolved transition metal ion activity in the electrolysis electrolyte. This helps favor the thermodynamic and kinetic reduction of the tetravalaent carbon, over metal deposition at the cathode. The practical observation is that, for the majority of molten carbonate CO2 electrolyses we have studied, the initial low current ramping is not observed to promote highest purity carbon deposition.

The first row of Figure 2 presents the product SEM of Electrolysis III, which continues to use a low current density, continues to exhibit a coulombic efficiency of 78%, focuses on a Ni powder addition to electrolyte, and refines the anode to Inconel 718 with two layers of Inconel 600, with an increase to 89 wt% the nano-bamboo product. Additionally, this electrolysis used an "aged" electrolyte (not delineated in the table). The freshly molten electrolyte requires time (up to 24 h) to reach a steady state equilibrium (pre-equilibration step) [16,20]. For Electrolysis III, the electrolyte was aged 24 h prior to melting and prior to immersion of the electrodes. However, it was observed that the aging is disadvantageous towards maximizing the nano-bamboo yield. A final refinement, in immediate use of the freshly melted electrolyte (elimination of the aging step), increases the nano-bamboo product to 90 wt% of the product (row 2 in Figure 2, and Electrolysis IV, and repeated as V in Table 2). Interesting, the 6% non-bamboo product in Electrolyses IV and V appears to be conical carbon nano-fiber (CNF) morphology, with its distinctive triangular-shaped voids in the morphology as seen in the second row of Figure 2. A simplified electrolysis eliminates observed CNF impurities resulting in 95% of the nano-bamboo allotorope. This Electrolysis VI is conducted without the current ramp activation at a high 0.4 A/cm2 current density, and exhibits a 99.7% coulombic efficiency. This electrolysis was tailored to have a purposeful excess of nucleation metals accomplished both with the use of Nichrome C electrodes, which contain Ni, Fe and Cr (Table 1), and through the direct addition of Ni and Cr powders to the electrolyte.

**Figure 2.** SEM of the synthesis product of nano-bamboo and nano-pearl allotropes of carbon by electrolytic splitting of CO2 in 770 ◦C Li2CO3. Moving left to right in the panels, the product is analyzed by SEM with increasing magnification. Scale bars in panels (starting from left) are for panels U: 100, 10, 3 μm (different electrolysis) and 2 μm; for panels T: 5, 2, 1 and 1 μm; for panels 11: 50, 30, 20 and 15 μm; for panels X: 50 μm 10, 1 and 2 μm.

The continued use of high concentrations of added transition metal powder to the electrolyte and low current density, but a change of electrodes, yields another distinct nanocarbon allotrope termed here as "hollow nano-onions". Specifically, in Electrolysis VII in Table 2 and Figure 2, the same concentration of Ni powder that had been used as in Electrolyses VI and V was used, and again the electrolyte was not aged nor were ramped initiation currents applied. However, a Monel cathode and Nichrome C anode were used, resulting in a 95 wt% of the product having a distinctive hollow nano-onions morphology. The hollow nature of the nano-onions will be revealed by TEM, but their spheroid character is seen by SEM in the third row of Figure 2. When the pure nickel electrolyte additive was changed to half nickel and half chromium powder, as summarized in Table 2 for Electrolyses

VIII and IX, the product has a distinctive "nano-pearl" morphology with its similarity to a beaded necklace. Here, the product fraction increased to 97% of this nano-pearl carbon and is seen by SEM in the bottom row of Figure 2. Electolyses VII–IX conducted have low J = 0.0 A/cm2 and exhibit a diminished coulombic efficiency of 79 to 80%.

Figure 3 compares TEM of the new nano-bamboo, nano-pearl and conical CNF nanocarbon allotropes synthesized by molten carbonate electrolysis. As seen in the top left panel of the figure, the conical carbon nanofibers (CNFs) exhibit conical voids typical of this CNF structure. Growth of the nano-bamboo is seen in the left middle of the figure and is nucleation driven, and the nucleation region appears to change shape, moving from tip to interior of the structure. We hypothesize that the lateral walls forming the bamboo "knobs" may be related to a periodic depletion of the carbon building leading walls. The walls of the nano-bamboo and nano-pearl allotropes exhibit graphene walls characterized by the typical intergraphene wall separation of 0.33 to 0.34 nm, as noted, and as measured by the observed separation between dense carbon planes in the TEM. The lower left of the figure shows the lateral multiple graphene layers separating the "knobs" of the nano-bamboo structure. The lower right of the figure shows the curved multiple graphene layers comprising the walls of the individual "beads" of the nano-pearl structure.

Figure 4 probes the elemental composition by HAADF (High Angle Annular Dark-Field TEM) and compares TEM of the new nano-bamboo and nano-pearl nanocarbon allotropes synthesized by molten carbonate electrolysis. As seen from the HAADF, the nano-bamboo product is pure carbon. That is with the exception of the presence of copper that, as shown in the lower left corner of the top left panel, is pervasively distributed at low concentration throughout, and likely originates from the grid mount of the product sample. HAADF probes two nano-pearl samples. The first exhibits a high or 100% concentration of carbon (the noise level is high) and little or no Ni, Cr or Fe. The second probes for carbon at higher resolutions and the rise and fall of carbon levels is evident as the probe moves from left to right over two separate nano-pearl structures.

The conical CNF, nano-bamboo and nano-pearl are new and unusual high-yield carbon allotropes as synthesized by molten electrolysis. Similar CVD synthesized morphologies have been synthesized by CVD. In particular, the CVD conical CNF structure has been widely characterized as shown in the upper row of Figure 5 [53–56]. In that figure, it is proposed that the morphology in CVD is due to repeated stress-induced deformation of the shape of the nucleating (Ni) metal, which causes the metal particles to jump and form the observed lateral graphene separation bridging the allotrope walls. Globular spaced nano-bamboo and nano-pearl allotropes are less common in CVD but have been observed. An example is shown in the lower left row of Figure 5, whose structures were attributed to the periodic formation of pores in the structure due to defects on the outer layers [57–59]. One specific application of bamboo CVD CNTs is as platforms for building layer by layer based biosensors [60]. Generally, carbon fibers are categorized as amorphous, or as shown on the lower right side of Figure 5, as built from graphene platelets, carbon nanotubes or conical type structures [61,62].

Figure 6 probes the TEM and the elemental composition by HAADF of the new hollow nano-onion nanocarbon allotropes synthesized by molten carbonate electrolysis. As seen, some of the nano-onion inner cores contain metal while others are void. The walls of the hollow nano-onions are composed of graphene layers as characterized by the typical intergraphene wall separation of 0.33 to 0.34 nm, as noted in Figure 6 and as measured by the observed separation between dense TEM carbon planes. As seen in the HAADF of the figure, when the core is vacant, the nano-onion is pure carbon, and when the core contains metal, the metal is either nickel or a mix of nickel and iron.

**Figure 3.** TEM of new molten carbonate synthesized carbon allotropes: Comparison of nano-bamboo, nano-pearl and conical CNF. (**A**): CNF; (**B**): Nano-bambpo; (**C**): Nano-pearl; (**D**), D-1, D-2 Nanobamboo; (**E**), E-1, E-2, E-1-1; E1-2 & E1-3 Nano-bamboo including measured graphene layer thickness; (**F**–**H**) Nano-bamboo knobs, (**I**,**J**) Nano-pearl; J-1, J-2 Nano-pearl with measured graphene layer thickness.

**Figure 4.** High angle annular dark-field TEM (HAADF) elemental analysis of nano-bamboo (panel A) and nano-pearl (panels B and C and HAADF elemental profiles) carbon allotropes synthesized by molten carbonate electrolysis. (**A**): Nano-bamboo with elemental analysis; (**B**): Nano-pearl with Elemental profile; (**C**): Nanopearl with Elemental profile.

**Figure 5.** Top row: Conical variations of bamboo carbon nanofibers, and their proposed mechanism of growth, as formed by nickel nucleated CVD using methane and hydrogen. Reproduced open access from Reference [53] Left bottom: Knotty bamboo nanocarbon variations by CVD, and their proposed mechanism of growth. Modified from Reference [58] Right bottom: General graphene layer conformations occurring in carbon nanofibers. Modified from Reference [61].

**Figure 6.** TEM and HAADF elemental analysis of the hollow nano-onion carbon allotrope synthesized by molten carbonate electrolysis. (**A**) and A-1: Hollow nano-onions with and without trapped metal. (**B**) and B-1: Hollow nano-onion without trapped metal. Note, measured graphene layer thickness is part of sub-figure B-1. (**C**) and C-1: Hollow nano-onion with trapped metal. Note, measured graphene layer thickncess and the elemental profile are part of sub-figure (**C**). (**D**): Hollow nano-onion with and without trapped metal. Note, measured elemental HAADF and elemental analysis are part of sub-figure (**D**).
