Open AccessCCS ChemistryRESEARCH ARTICLE1 Feb 2022Successive Free-Radical C(sp2)–C(sp2) Coupling Reactions to Form Graphene Huaqiang Cao†, Cheng Wang, Baojun Li, Tianyu Chen, Peng Han, Yan Zhang, Haijun Yang, Qunyang Li and Anthony K. Cheetham† Huaqiang Cao† *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Department of Chemistry, Tsinghua University, Beijing 100084 †H. Cao and A. K. Cheetham contributed equally to this work.Google Scholar More articles by this author , Cheng Wang Department of Chemistry, Tsinghua University, Beijing 100084 Google Scholar More articles by this author , Baojun Li Department of Chemistry, Tsinghua University, Beijing 100084 Google Scholar More articles by this author , Tianyu Chen Beijing Advanced Innovation Center for Imaging Technology and Key Laboratory of Terahertz Optoelectronics (MoE), Department of Physics, Capital Normal University, Beijing 100048 Google Scholar More articles by this author , Peng Han Beijing Advanced Innovation Center for Imaging Technology and Key Laboratory of Terahertz Optoelectronics (MoE), Department of Physics, Capital Normal University, Beijing 100048 Google Scholar More articles by this author , Yan Zhang Beijing Advanced Innovation Center for Imaging Technology and Key Laboratory of Terahertz Optoelectronics (MoE), Department of Physics, Capital Normal University, Beijing 100048 Google Scholar More articles by this author , Haijun Yang Department of Chemistry, Tsinghua University, Beijing 100084 Google Scholar More articles by this author , Qunyang Li Applied Mechanics Laboratory, Department of Engineering Mechanics, School of Aerospace Engineering, Tsinghua University, Beijing 100084 Google Scholar More articles by this author and Anthony K. Cheetham† *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Department of Materials Science and Metallurgy, University of Cambridge, Cambridge CB3 0FS Materials Research Laboratory, University of California, Santa Barbara, Santa Barbara, CA 93106 †H. Cao and A. K. Cheetham contributed equally to this work.Google Scholar More articles by this author https://doi.org/10.31635/ccschem.021.202100919 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Graphene is of great interest because of its exciting properties and potential applications, but its production on a large-scale still presents considerable challenges. Herein, we report the synthesis of predominately few-layer graphene, due to π–π stacking, and single-layer graphene from reaction between hexabromobenzene and Na metal, followed by annealing to improve crystallinity. The reaction proceeds via a free-radical C(sp2)–C(sp2) coupling mechanism, which is supported by theoretical calculations. The graphene can host unpaired spin electrons, leading to a short acquisition time for a solid-state nuclear magnetic resonance 13C spectrum from unlabeled graphene, which is ascribed to the very short spin-lattice relaxation time. High catalytic activity for transforming amine to imine with a conversion of >99% and a yield of ∼97% is demonstrated, and high electronic conductivity of ∼105 S·m−1 is found by terahertz spectroscopy. The reaction delivers a method for synthesizing graphene with a high spin concentration from perbrominated benzene molecules by using an active metallic agent, such as Na, Li, or Mg. Download figure Download PowerPoint Introduction Graphene, in the strictest term, is a single-layer of carbon atoms tightly packed into a two-dimensional (2D) honeycomb lattice; in a broader sense, graphene can be distinguished as three different types of 2D graphene materials, that is, single-, double-, and few-layer (3–10 layers) samples.1 Many methods have been reported for the synthesis of graphene,2 including bottom-up approaches such as solvothermal synthesis and sonication based on reactions between ethanol and sodium in solution,3 surface-assisted coupling and cyclodehydrogenation gas-phase reactions on metal gold or silver surfaces through a radical addition process (albeit with technological limitations substrate extension),4 and on Cu surfaces through radical-coupling reactions.5 Graphene films have also been synthesized by chemical vapor deposition techniques using a mixture of methane and hydrogen at 1000 °C on Cu foils.6 Several top-down approaches have also been explored, including the exfoliation–reintercalation–expansion of graphite to generate single-layer graphene sheets ∼250 nm in size7; nonchemical, solution-phase exfoliation of graphite in organic solvents to form monolayer and few-layer graphene with a lateral size of a few micrometers8; and exfoliation of graphite by brief (60 s) heating to 1000 °C in forming gas (3% hydrogen in argon) to generate sub-10-nanometer graphene nanoribbons.9 Recently, atomic-scale structural defects in graphene,10–20 which can host unpaired spins,10–12 has aroused great interest. Such defect engineering21–26 can be used to produce materials that have applications in quantum information processing,27 spins,28 photons,29 and so on. Defects in graphene are usually generated during synthesis, device preparation, or transfer procedures, and include point defects,11 line defects containing pentagon–heptagon pairs,13 holes and edge structures,14 tilt grain boundaries,15 dislocations,16 topological defects,17 atomic-scale lattice defects,18 extended defects,19 and so on; these can greatly alter graphene’s physical and chemical properties. Herein, we report a simple method, and the associated free-radical C(sp2)–C(sp2) coupling reaction mechanism, for the synthesis of graphene from the reaction of the hexabromobenzene (HBB), with metallic Na, Li, or Mg. Although the synthesis of graphene quantum dots from HBB and Na has been reported,30 the detailed mechanism of this C(sp2)–C(sp2) coupling reaction has not been described previously.30,31 Our free-radical C–C coupling reaction to form graphene is neither a Wurtz reaction [coupling between alkyl halide-C(sp3) and alkyl halide-C(sp3)], nor a Wurtz–Fittig reaction [coupling between alkyl halide-C(sp3) and aryl halide-C(sp2)].32 Neither is it an Ullmann reaction (the coupling of aryl halides with copper undergoing an oxidative addition step),32 a free-radical polymerization reaction without the chair propagation step,33 nor a general inorganic nanoparticle growth mechanism.34,35 It is a free-radical aryl halide C(sp2)–C(sp2) coupling reaction, but not a free-radical chain reaction.36 It has been stated that “the coupling of two aryl halides with sodium is impractical,”30,32 but biaryl compounds can be formed via the Ullmann reaction.32 Our mechanism has similarities to the classical mechanisms, but there are also differences which deserve to be understood and elaborated upon. Our synthetic method is also different from a free-radical polymerization involving successive additions of free-radical building blocks33 and from the formation of inorganic nanocrystals via nucleation and subsequent growth (the Ostwald ripening mechanism34 or oriented attachment mechanism35). However, as described later, it progresses from an organic molecular HBB to a “macromolecule” with fused six-membered rings—“graphene molecules”—to graphene via successive C–C coupling steps in a radical step-growth manner. The as-synthesized graphene has been characterized by many physical and chemical techniques, which clearly demonstrate the formation of the 2D nanostructured carbon product and shed light on the underlying reaction mechanism. In addition to establishing the reaction mechanism, we show that our graphene contains stable free radicals and can function as a metal-free catalyst for an amine to imine oxidation reaction with a conversion of >99% and a yield of ∼97%. Experimental Methods Materials All chemical reagents, such as hexabromobenzene [HBB, C6Br6, Analytical Reagent (AR), 99.0%] purchased from J&K Scientific Ltd. (Beijing, China) and Na (AR, >99.8%) from Acros Organics (France), were used without purification unless otherwise noted. Toluene (AR, >99.5%; Beijing Chemical Works) was treated with metal Na to remove traces of water. Synthesis of graphene In a typical synthesis, 1000 mg of HBB powder was dissolved in 400 mL of dehydrated toluene with stirring to form a solution in a Teflon-lined stainless steel autoclave (500 mL in capacity), followed by rapid addition of 600 mg of metal Na in small pieces into the above autoclave with stirring. Then the mixture was sealed, heated to 180 °C, and maintained at this temperature for 50 h. After the autoclave was allowed to cool to room temperature, the products were collected and treated with absolute alcohol to deplete unreacted Na, followed by washing with deionized water. After drying under vacuum, the as-obtained solid product was placed into a porcelain crucible, which was placed inside a tube furnace for annealing. Argon gas was first fed into the tube furnace at room temperature for 1 h to ensure that the furnace was full of inert gas. Then, the tube furnace was heated from room temperature to 450 °C [we also selected temperatures of 500 and 550 °C for Brunauer–Emmett–Teller (BET) measurements] with a heating rate of 10 °C·min−1, followed by cooling to room temperature and then holding at room temperature for 3 h. After the above annealing procedure, the final product was collected. Also, the same reaction system was heated at 180 °C under atmospheric pressure and reflux conditions for 15 h, while all the other reaction parameters remained unchanged. In addition, metallic Li, Mg, Fe, Co, Ni, Cu, and Zn were independently selected as a reagent in place of Na for this reaction, while all the other reaction parameters remained unchanged. Also, 4,4′-dibromobiphenyl and 1,3,5-trichlorobenzene, in place of HBB, were treated with Na, respectively, while all the other reaction parameters remained unchanged. The resultant product was washed by absolute alcohol and deionized water alternately, using centrifugation at 10,000 rpm (revolution per minute) for 5 min several times until the suspension was colorless. The first deionized-water-washing supernatant was retained for consecutive 72 h-dialysis process, which was used for transmission electron microscopy (TEM) observation, while the solid product was dried in a vacuum oven at 60 °C for 24 h for other characterization. The yield is affected by the size of the Na pieces in the air, which in turn is affected by the humidity in the air. Results and Discussion Characterization of graphene Representative TEM images and the corresponding electron diffraction (ED) patterns, atomic force microscopy (AFM), and optical microscopy images of the graphene samples are shown in Figures 1a and 1b and Supporting Information Figures S1–S5. The maximum observed size of the graphene nanosheet is ∼80 μm × 80 μm ( Supporting Information Figure S2b). Annealing the sample generates single-crystalline areas with lattice fringe spacing of ∼0.38 nm, corresponding to the (002) ( Supporting Information Figure S1). This interlayer spacing is larger than the distance between adjacent graphene layers in AB stacked graphite, 0.335 nm, which is due to the oxygen functional groups, generated by O2 or CO2 in the termination step of the C–C coupling (detailed discussion in Supporting Information), on the edge or plane of graphene. Also, the lattice fringes of ∼0.38 nm agrees with the X-ray diffraction (XRD) result (d(002) = 0.381 nm). When different regions of the as-synthesized sample are inspected, well-defined ED spots instead of ring patterns, are always observed ( Supporting Information Figure S1b), indicating the crystallinity of all regions examined.37 Figure 1c clearly shows the short-range ordering of nanosheets, demonstrated by both XRD and Raman results. Figure 1 | Physical characterization of the as-synthesized graphene. (a) TEM image. (b) AFM images of the samples of the first four largest number of layers deposited on an oxidized silicon wafer substrate. (c) Raman spectrum (532 nm) of the samples deposited on an oxidized silicon wafer substrate. (d) XPS image of the sample. Download figure Download PowerPoint The step height between the Si wafer (SiO2/Si) substrate surface and the thinnest graphene surface is ∼0.53 nm, as measured by AFM. This is important evidence that we can synthesize graphene only a single-atom-layer thick.2 According to the AFM thickness statistics, the thickness distribution of graphene is 15% for single-layer, 22.5% for bi-layer, and 62.5% for few-layers (3–10 layers), respectively ( Supporting Information Figures S4 and S5 and Table S1). The XRD pattern of the powder sample is shown in Supporting Information Figure S6. For the peak related to any stacking, we calculated the interlayer distance as 0.381 nm (corresponding to 2θ = 23.3°) based on Bragg’s law, 2dsinθ = λ (λ =1.5406 Å), which corresponds to the (002) crystallographic plane of stacked graphene sheets.38 The carbon (002) diffraction peak is extremely broad, which may stem from any short-range ordering of layers or the existence of defects in the sample39; note the successful removal of NaBr by washing with water ( Supporting Information Figure S6a). Micro-Raman spectroscopy characterization presents two intense peaks, the G-peak at ∼1578 cm−1 and the D-peak at ∼1361 cm−1, as well as three weak peaks at ∼2867 cm−1 (2D-peak) (historically called G′, related to the second order of zone-boundary phonons, but having nothing to do with the G-peak), ∼3002 cm−1 (D +D′-peak, which is caused by a two-phonon defect-assisted process; this peak is often labeled as D + G in the literature), and ∼3171 cm−1 (2D′-peak, an overtone of D′), respectively (Figure 1c and Supporting Information Figures S7 and S8). The large overlap between the D- and the G-peaks indicates that the material has a partially disordered structure. The Raman mapping IG/ID ratios at the edges and in the inner regions are approximately 1.27 and 1.67 ( Supporting Information Figure S8), respectively, indicating that the peripheral edges have more defects than the interior edges (detailed discussion in Supporting Information). X-ray photoelectron spectroscopy (XPS) analysis (Figure 1d and Supporting Information Figure S9 and Tables S2–S5) reveals the predominant C1s peak (∼97.2 atomic percentage (at. %)) and an O1s peak (∼2. 8 at. %). Peaks for Br and Na were not found, indicating that NaBr had been removed and that no C–Br appears in the final product, which is further supported by inductively coupled plasma mass spectrometry (ICP-MS) results ( Supporting Information Figure S10 and Table S6). The high-resolution XPS data ( Supporting Information Figure S9 and Tables S2 and S3) reveal the existence of sp2 aromatic carbon, carbon-containing functional groups (C–OH, C=O, –COOH), sp3-carbon, and π–π*-carbon in the final product, among which ∼91.8 at. % is for sp2-carbon, ∼2.7 at. % is for sp3-carbon, and ∼2.7 at. % is for other carbon covalently bound to oxygen.40,41 The nitrogen adsorption–desorption isotherms and Barret–Joyner–Halender (BJH) pore size distribution curves of the powder sample ( Supporting Information Figure S11 and Table S7) indicate aggregates of plate-like particles impart slit-shaped pores ( Supporting Information Figures S11a, S11c, S11e, and S11g).42 The BET specific surface area (SSA) value for our sample is 443.6 m2·g−1 ( Supporting Information Figures S11b, S11d, S11f, and S11h), and the average pore diameters are ∼1.43 (for DBJH) and ∼2.65 nm (for DBET) from the corresponding pore diameter distributions ( Supporting Information Figures S11a and S11b; detailed discussion in Supporting Information). The electron spin resonance (ESR) spin trapping technique is a general approach that can provide direct evidence for the presence of free radicals.43 For the detection of free radicals in this work, 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) and phenyl-N-tert-butylnitrone (PBN) spin traps were employed. Both can confirm the existence of free-radical intermediates and differentiate between carbon- and oxygen-centered radicals (Figure 2a and Supporting Information Figures S12–S15; detailed discussion in Supporting Information). In addition, we have measured the ESR spectrum for a powder sample ( Supporting Information Figures S16 and S17 and Tables S8 and S9), which shows a strong peak with a g-factor close to 2.003 and a linewidth <1 mT ( Supporting Information Table S8). This indicates that it is incompatible with metal ions ( Supporting Information Figure S16 and Table S8).21 Figure 2 | Chemical characterization of the as-synthesized graphene. (a) First-derivative ESR of the reaction solution after 30 min. The ESR spectra for the HBB radical produced by metal Na in toluene trapped by DMPO (upper left) and PBN (upper right). (b) 1D MAS ssNMR spectrum of the unlabeled graphene. (c) Matrix-free LDI-TOF-MS spectrum in the negative-ion mode. (d) FT-IR spectroscopy spectra of the toluene solvent and the as-synthesized graphene. Download figure Download PowerPoint Solid-state nuclear magnetic resonance (ssNMR) can provide important structural insights into carbon-based materials, and high-resolution ssNMR using magic-angle spinning (MAS) is a primary technique for the characterization of graphene-related structures such as graphite oxide,44 graphene oxide (GO),45 and reduced GO (rGO) electrodes.46 The 13C MAS ssNMR spectrum reveals only one peak at a chemical shift of δ ∼128.6 ppm with full width at half maxima of approximately 17 ppm, which is characteristic of sp2 aromatic carbon atoms (Figure 2b); no peaks for other carbon-containing functional groups, such epoxide C (δ 59.7 ppm) or alcohol C–OH (δ 69.6 ppm),44 or the methyl carbon of toluene (δ ∼20 ppm)47,48 are observed. The ssNMR peak is wider than that obtained with liquid-state NMR due to anisotropic effects that are caused by restricted molecular motion in the solid state (detailed discussion in Supporting Information). The slightly asymmetric character of our peak reflects the presence of other unresolved components, but our one-dimensional (1D) ssNMR approach cannot distinguish these sp2-carbons. 2D ssNMR for 13C-unlabeled samples would require 108-fold more acquisition time due to the low natural abundance of 13C, whereas the use of 13C-labeling is prohibitively expensive. Thus, it is currently difficult to resolve the components of the peak at 128.6 ppm. Our ssNMR observations further demonstrate that the as-synthesized sample is overwhelmingly composed of an sp2-carbon network (i.e., graphene). To confirm that the chemical shift at δ ∼128.6 ppm comes from graphene and not from other organic aromatic molecules,49 the spin-lattice relaxation time (T1 = 4.29 s) and solid-state ESR (ssESR) spectra were measured. We find that the short measurement time of 16.5 min for an ssNMR experiment can be ascribed to the very short T1 caused by paramagnetic relaxation enhancement effects ( Supporting Information Figures S18 and S19) because of the paramagnetic centers that are observed in the ESR spectra ( Supporting Information Figures S16 and S17 and Tables S8 and S9, detailed discussion in Supporting Information). Thus, these data show that the ∼128.6 ppm peak in the ssNMR cannot be derived from sp2-carbons in aromatic molecules, such as the HBB reagent or the toluene solvent ( Supporting Information Table S10), but must stem from sp2-carbon in graphene. The 1H ssNMR measurement reveals a strong peak at δ ∼6.8 ppm accompanied by a weak peak at δ ∼2.5 ppm ( Supporting Information Figure S20). The peak at ∼6.8 ppm is attributed to hydroxyl hydrogen attached to the edge of graphene (formation of oxygen-containing functional groups discussed in Supporting Information Section 3.8).50,51 Because of the defects of the graphene detected by high electron spin concentration from the ESR data ( Supporting Information Figure S16 and Table S9) and Raman mapping ( Supporting Information Figure S8), the weak peak at 2.5 ppm may be attributed to a shielded hydroxyl hydrogen, induced by the adjacent defect electron structures, near the defect edge of graphene.52,53 To understand the reaction pathways and intermediates, we carried out laser desorption ionization time-of-flight mass spectrometry (LDI-TOF-MS) measurements on the reaction solution after 30 min at 180 °C (Figure 2c and Supporting Information Figures S21–S24 and Tables S11–S13). Although, we do not expect to be able to explain every peak in the mass spectrum,50 we can detect the intermediate pentabromobenzene radical, [C6Br5]• ( 2), at m/z of 466.6810, 468.6300, 470.6322, 472.6304, 474.6316, and 476.6042; bromobiphenyl derivatives (similar to 3), for example, [C16H8O12Br2]− with m/z of 549.5606, 551.5568, and 553.5274; terphenyl derivatives (similar to 7), for example, [C18H12O12Br]− with m/z of 498.0877 and 500.1276; and the bromotriphenyl derivative (similar to 8), for example, [C22H11O13Br]−m/z of 562.0644 and 564.0542. These appear to be the four smallest structures generated by C–C coupling from the HBB precursor molecules, and this result helps us to understand the proposed reaction mechanism, which is shown in Figure 3. The peaks from five bromine-free aromatic derivatives containing oxygen can also be monitored (Figure 2c and Supporting Information Figure S24 and Table S13), demonstrating that these oxygen-containing structures are also formed by the inhibitor O2 or CO2 ( Supporting Information) when the bromine-containing radical concentration is low. These intermediates are easily dissolved in water due to their oxygen-containing groups. They can be removed from the final products after multiple washing processes, thereby reducing the oxygen concentration in the final products as observed in the XPS data (Figure 1d and Supporting Information Table S2). More fragmentation peaks can also be observed because LDI is a harder ionization method than matrix-assisted laser desorption ionization. We focused on molecular masses below 600 (m/z) because the key intermediates obtained by coupling two or three benzene rings from the first two steps of the radical reaction are below 600 (m/z), such as triphenylene, the largest intermediate in the first two reaction, with relative molecular mass of 228.3. Moreover, the peaks above 600 (m/z) in matrix-free LDI-TOF-MS spectrum, which are “fused graphene molecules” of different sizes, are too numerous and weak to distinguish. Figure 3 | Overall synthetic mechanism of graphene from HBB (C6Br6) under solvothermal conditions with three possible basic routes to generate the key intermediate triphenylene. Route I: (Step 1) Initiation step, HBB free-radical anions (1) formed via electron transfer, reduction by Na, and fragmentation of HBB free-radical anions to neutral C6Br5• pentabromobenzene radicals (2); (Step 2) Coupling step, C–C coupling of (2) to form perbromo-1,1′-biphenyl (3); (Step 3) Generation of 2′,3,3′,4,4′,5,5′,6,6′-nonabromo-[1,1′-biphenyl]-2-radical (5); (Step 4) Coupling between 5 and 2, followed by reduction by Na to form 3,3′,3″,4,4′,4″,5,5′,5″,6,6′,6″-dodecabromo-[1,1′:2′,1″-terphenyl]-2,2″-biradical (7); (Step 5): Coupling within 7 to generate 2,3,6,7,10,11-hexabromotriphenylene hexaradical (8), due to steric hindrance. Route II: (Step 1) Initiation step, the same as Step 1 in Route I; (Step 2) Coupling step, the same as Step 2 in Route I; (Step 3) Generation of 3,3′,4,4′,5,5′,6,6′-octabromo-[1,1′-biphenyl]-2,2′-biradical (6) from 3;(Step 4) Coupling between 6 and 4 to form 8. Route III: (Step 1) Initiation step, the same as Step 1 in Route I; (Step 2) Coupling step, the same as Step 2 in Route I; (Step 3) Generation of 3,4,5,6-tetrabromobenzene-1,2-biradical C6Br4˙˙ (4) from 2; (Step 4): Coupling by three radicals (4) to form 8. Repeat of above C(sp2)–C(sp2) coupling steps to produce graphene. Termination step: radicals react with O2 or CO2 to generate oxygen-containing structures (9, 10, 11, 12, 13, 14, etc.). Download figure Download PowerPoint To further understand the reaction mechanism, we used Fourier transform infrared (FT-IR) spectroscopy to study the intermediate products after a reaction time of 30 min (Figure 2d).50,51 In comparison with the solvent toluene, we observe a newly formed peak at 3279 cm−1 in the reaction system, which is attributed to O–H stretching vibrations (νO–H) from intramolecularly H-bonded O–H groups in carboxylic acids or hydroxybenzenes. It does not correspond to water because the νO–H bond for water should appear at ∼3400 cm−1 (detailed discussion in Supporting Information). Reaction mechanism ESR analysis (Figure 2a and Supporting Information Table S8) offers direct evidence for the presence of radicals in the reaction systems, which is incompatible with metal ions ( Supporting Information Figure S10 and Table S8).21 The overall reaction of the synthesis of graphene is shown in eq 1, and proceeds by a step-growth radical carbon–carbon (C–C) coupling mechanism (Figure 3) with various key intermediate structures and an average yield of >23%. ((1)) The nonchain reaction mechanism can be broken down into radical initiation, coupling, and termination step. Initiation step Typically, a radical anion will be generated after adding an electron to a precursor molecule, and then the radical anion will fragment to form a neutral free radical and an anion. The reaction of the aromatic aryl halides (here C6Br6, i.e., HBB) by metallic Na in toluene is an electron transfer reaction, which follows a stepwise mechanism,54 and generates the radical anion C6Br6 ˙ − ( 1) (eq 2).55 The low, negative, standard reduction potential of sodium [E⊖(Na+/Na) = −2.714 V] indicates that it is a strong reducing agent. The π* orbital accepts the electron56 associated with an aromatic ring, and after the electron is accepted the product is radical anion 1.57 The resonance structures of 1, comprising approximately 18 different structures (shown in eq 3) contain a pair of electrons and an unpaired electron which are localized on two carbon atoms of the π system, and the three electrons can be distributed on any two ortho or para atoms of the ring.57 Because the “extra” electron in the intermediate radical anion C6Br6 ˙ − will occupy a high-energy antibonding orbital, this will weaken the structure. Thus, it will cleave the C–Br bond to give the most stable neutral radical C6Br5• ( 2) and NaBr, eq 2(ii).57,58 ((2)) ((3)) Radical coupling and termination steps Usually, two radicals can react with one another in one of two ways: radical–radical combination or disproportionation.57 Radical–radical combination reactions are usually very fast and favorable, since they are highly exothermic,57,58 but radical concentration plays an important role in such reactions. At high radical concentrations, the radicals will prefer to undergo rapid combination reactions to yield more stable, nonradical species rather than a radical product.58 The C–C coupling of radical 2 can thus yield perbromo-1,1′-biphenyl 3, (eq 4), which can be stabilized by twisting the C–C single bond between two benzene rings. ((4)) Subsequently, radicals are regenerated following the above stepwise mechanism, like the reaction in eq 2, to generate the 3,4,5,6-tetrabromobenzene-1,2-biradical C6Br4•• 4 from radical C6Br5• 2 (eq 5). Of course, other kinds of debrominated phenyl radicals can be obtained by loss of bromine atoms from different positions in the phenyl ring. In addition, 2′,3,3′,4,4′,5,5′,6,6′-nonabromo-[1,1′-biphenyl]-2-radical ( 5) (eq 6) and 3,3′,4,4′,5,5′,6,6′-octabromo-[1,1′-biphenyl]-2,2′-biradical ( 6) (eq 7) can also be generated from 3, respectively, and other debromination biphenyl radicals generated by Na reduction can also be formed in a similar manner. ((5)) ((6)) ((7)) Radical coupling of 5 with 2, followed by Na reduction to generate 3,3′,3″,4,4′,4″,5,5′,5″,6,6′,6″-dodecabromo-[1,1′:2′,1″-terphenyl]-2,2″-biradical ( 7) (eq 8) can form 2,3,6,7,10,11-hexabromotriphenylene hexaradical ( 8) (eq 9) (Route I). The triphenylene structure 8 is the key intermediate for formation of graphene. Other debrominated triphenyl skeleton free radicals (the number of bromine atoms on three outer benzene rings, x, y, and z = 0–3, respectively, due to the steric hindrance effect) can also be formed. After repeating the above steps, fused