# Fileset

[s41467-025-56389-6.pdf](https://mdr.nims.go.jp/filesets/dabc93a2-b606-4205-b430-ec66e27b32fd/download)

## Creator

[Akira Takakura](https://orcid.org/0009-0004-3639-2330), [Taishi Nishihara](https://orcid.org/0000-0001-6973-2005), [Koji Harano](https://orcid.org/0000-0001-6800-8023), [Ovidiu Cretu](https://orcid.org/0000-0002-1822-8172), [Takeshi Tanaka](https://orcid.org/0000-0001-7547-7928), [Hiromichi Kataura](https://orcid.org/0000-0002-4777-0622), [Yuhei Miyauchi](https://orcid.org/0000-0002-0945-0265)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Coalescence of carbon nanotubes while preserving the chiral angles](https://mdr.nims.go.jp/datasets/d44f2589-e40f-4299-a208-b432bbd56dd6)

## Fulltext

Coalescence of carbon nanotubes while preserving the chiral anglesArticle https://doi.org/10.1038/s41467-025-56389-6Coalescence of carbon nanotubes whilepreserving the chiral anglesAkira Takakura 1, Taishi Nishihara 1, Koji Harano 2,3, Ovidiu Cretu2,Takeshi Tanaka 4, Hiromichi Kataura 4 & Yuhei Miyauchi 1Atomically precise coalescence of graphitic nanocarbon molecules is one ofthemost challenging reactions in sp2 carbon chemistry. Here, we demonstratethat two carbon nanotubes with the same chiral indices (n, m) are efficientlycoalesced into a single (2n, 2m) nanotubewith preserved chiral angles via heattreatment at less than 1000 °C. The (2n, 2m) nanotubes constitute up to ≈20%–40% of the final sample in the most efficient case. Additional opticalabsorption peaks of the (2n, 2m) nanotubes emerge, indicating that thereaction occurs over the entire sample. The reaction efficiency stronglydepends on the chiral angle, implying that C–C bond cleavage and recombi-nation occurs sequentially. Furthermore, the reaction occurs efficiently evenat 600 °C in an atmosphere containing trace amounts of oxygen. These find-ings offer routes for the structure-selective synthesis of large-diameter nano-tubes and modification of the properties of nanotube assemblies viapostprocessing.Joining or fusing graphitic nanocarbon materials comprising a hex-agonal sp2 carbon network in a well-controlled manner is among themost challenging issues in sp2 carbon chemistry as it requires thecleavage of numerous C–C bonds and recombination into sp2 bondswith atomic precision. The conversion of large nanocarbon macro-molecules such as fullerenes and carbon nanotubes into large, coa-lesced molecules1–9 is a well-known conundrum (Fig. 1a, b). Since theelectronic and optical properties of these nanocarbon materials aredominated by the geometry and topology of π electrons delocalizedon the sp2 carbon network10, the coalescence of these nanocarbonmaterials can drastically modify their properties. The seamless coa-lescence of two single-walled carbon nanotubes (hereafter referred toas nanotubes) into one thick nanotube is of particular interest (Fig. 1b).Because electronic11–13, optical14,15, chemical16,17, thermal18,19, andmechanical20,21 properties of nanotubes strongly depend on theirstructure defined by their diameter and chiral angle (or chiral indices(n,m)) (Fig. 1c), considerable effort hasbeendevoted to achieving fullystructure-controlled synthesis22,23 or structural separation24–31 ofnanotubes. However, structure-controlled synthesis or separationmethods have remained limited to (n, m) species with relatively smalldiameters (≈ 1 nm or less), and precisely structure-selective synthesisor sorting of thick nanotubes with diameters ofmore than ≈ 1.3 nm hasnot been realized because of the great variety of geometrically possi-ble structures of nanotubes with similar diameters and properties.Atomically precise nanotube coalescencemay address this issue. If theefficient coalescence reaction of small-diameter (n, m) nanotubes forwhichhigh-purity samples are available24–31 can be triggered, structure-selective synthesis of thick (2n, 2m) nanotubes can be achieved fromthe (n, m) nanotubes as precursors (Fig. 1b). Furthermore, the mac-roscopic properties of bulk nanotube aggregates, including theirelectronic and thermal conductivities, can be dramaticallymodified byforming covalent bonds between carbon nanotubes in the aggregateusing partial coalescence reactions, and may provide alternativeways for modifying the properties of nanotube assemblies viapostprocessing.However, the coalescence of nanotubes remains challenging.Although the connection of two fullerene molecules via solid-statereaction2 and the conversion of fullerene peapods into double-walledReceived: 29 April 2024Accepted: 16 January 2025Check for updates1Institute of Advanced Energy, Kyoto University, Uji, Kyoto 611-0011, Japan. 2Center for Basic Research on Materials, National Institute for Materials Science,Tsukuba, Ibaraki 305-0044, Japan. 3Research Center for Autonomous Systems Materialogy (ASMat), Institute of Integrated Research, Institute of ScienceTokyo, Yokohama, Kanagawa 226–8501, Japan. 4Nanomaterials Research Institute, AIST, Tsukuba, Ibaraki 305-8565, Japan.e-mail: miyauchi@iae.kyoto-u.ac.jpNature Communications |         (2025) 16:1093 11234567890():,;1234567890():,;http://orcid.org/0009-0004-3639-2330http://orcid.org/0009-0004-3639-2330http://orcid.org/0009-0004-3639-2330http://orcid.org/0009-0004-3639-2330http://orcid.org/0009-0004-3639-2330http://orcid.org/0000-0001-6973-2005http://orcid.org/0000-0001-6973-2005http://orcid.org/0000-0001-6973-2005http://orcid.org/0000-0001-6973-2005http://orcid.org/0000-0001-6973-2005http://orcid.org/0000-0001-6800-8023http://orcid.org/0000-0001-6800-8023http://orcid.org/0000-0001-6800-8023http://orcid.org/0000-0001-6800-8023http://orcid.org/0000-0001-6800-8023http://orcid.org/0000-0001-7547-7928http://orcid.org/0000-0001-7547-7928http://orcid.org/0000-0001-7547-7928http://orcid.org/0000-0001-7547-7928http://orcid.org/0000-0001-7547-7928http://orcid.org/0000-0002-4777-0622http://orcid.org/0000-0002-4777-0622http://orcid.org/0000-0002-4777-0622http://orcid.org/0000-0002-4777-0622http://orcid.org/0000-0002-4777-0622http://orcid.org/0000-0002-0945-0265http://orcid.org/0000-0002-0945-0265http://orcid.org/0000-0002-0945-0265http://orcid.org/0000-0002-0945-0265http://orcid.org/0000-0002-0945-0265http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-56389-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-56389-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-56389-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-025-56389-6&domain=pdfmailto:miyauchi@iae.kyoto-u.ac.jpwww.nature.com/naturecommunicationscarbon nanotubes at temperatures higher than 1200 °C6 have beenreported, the number of sp2 C–C bonds that must be cleaved andreconstructed for thenanotube coalescence ismuch larger than that infullerenes. In addition, the chemical and thermal stability of nanotubesis much higher than that of fullerenes. Previous studies have shownthat partial removal of carbon atoms by electron irradiation at 800 °Cin transmission electron microscopy (TEM) could initiate the coales-cence reaction of nanotubes4. The existence of small amounts ofcoalesced nanotubes after heat treatment of nanotube aggregations atmore than 1700 °Cwas also found by TEMobservations5. However, therealization of efficient coalescence all over the macroscopic nanotubeaggregations remains an open issue.Here, we report efficient coalescence of (n,m) carbon nanotubesinto doubled (2n, 2m) nanotubes enabled using heat treatment. Afterheat treatment at 900–1000 °C under low pressure, micro-Ramanspectroscopy measurements and aberration-corrected TEM observa-tions were used to verify the chiral-angle-preserving coalescence of (n,m) nanotubes into (2n, 2m) nanotubes. Optical absorption peaks ofthe exciton resonance for the (2n, 2m) structures emerged after thereaction, whereas the absorption intensity of the original (n, m)structures decreased. Moreover, the efficiency of the coalescencestrongly depended on the chiral angle of the nanotubes; only armchairn =m (θ = 30°) and near-armchair n ≈m (θ ≈ 30°) types showed effi-cient coalescence. The content of coalescence-derived (2n, 2m)nanotubes in the final product reached 20%–40% for the armchair andnear-armchair cases, as evaluated using optical absorption spectro-scopy. In contrast, near-zigzag (n, m) nanotubes with n >> m showedlow coalescence efficiency. These results can be explained based onthe chirality dependence of energy costs for the coalescence reactionoriginating from geometric factors. Furthermore, we found that thecoalescence reaction occurs efficiently even at 600 °C when traceamounts of oxygen is introduced into the reaction chamber. Thesefindings offer an approach for synthesizing chiral-structure-controlledcarbon nanotubes with large diameters as well as an opportunity tomodify the physical properties of carbon nanotube aggregates usingpostprocessing, which may be useful for their use as bulk materials invarious applications.Results and discussionCoalescence reaction of (6,5) nanotubesFirst, we demonstrate that efficient coalescence of (n, m) = (6,5) near-armchair carbon nanotubes into doubled (2n, 2m) = (12,10) nanotubesis feasible. One possible reason for the low efficiency of the coales-cence reaction in previous studies may be the use of mixed nanotubeswith various chiral structures specified by either the chiral indices(n, m) or the chiral angle and diameter (θ, d) (Fig. 1c), which hardlycoalesce geometrically. Therefore, in this study, we examined thecoalescence reaction using the aggregation of carbon nanotubes withthe same (n,m) structures.We fabricated nanotubemembranes via thefiltration of (6,5)-enricheddispersions32 (Fig. 2a) preparedusing the gelchromatography separationmethod29. Themembranewas transferredonto the sapphire substrate (Fig. 2b), followed by heat treatment in avacuum using an electric furnace (see Methods for the details of thesample preparation and the heating experiment). Figure 2c, d showsthe optical absorption of (6,5)-enriched nanotube membranes beforeand after heat treatment at 1000 °C, measured using optical trans-mission spectroscopy (see Supplementary Fig. 1 for the broadbandspectra up to 3 eV). In thesemeasurements, the optical absorptionwasmeasured using an incident light beam covering the entire substrate.Thus, the absorption spectra represent the ensemble-averaged opticalproperties of themembranes. The strong absorption peak arising fromthe first sub-band exciton (S11) for (6,5) nanotubes was observed at1.21 eV before and after heat treatment. After heat treatment, theintensity of this exciton peak decreased, indicating the decreasednumber of (6,5) nanotubes in the membrane. In addition, a distinctabsorption peak emerged at 0.67 eV after the heat treatment (theasterisk in Fig. 2d). According to the table of the excitation photonenergy for each (n, m) nanotube type, namely the Kataura plot14,33, theS11 exciton absorption peak energies of doubled (2n, 2m) species areexpected to be approximately half of those of the (n, m) nanotubes,while their S22 exciton absorption peak energies should be nearlythe same as the S11 peak energies of the original (n, m) species. Thus,the emergence of this additional low energy peak originating from theexciton resonance of nanotubes thicker than the original (6,5) speciesis the signature of efficient nanotube coalescence reaction throughoutthe membrane. The reaction occurred in the entire sample is alsosupported by the macroscopic changes in transport properties beforeand after the nanotube coalescence reaction (see Methods and Sup-plementary Fig. 2). The insets in Fig. 2c, d show typical radial-breathing-mode (RBM) Raman spectra of the sample before and afterheat treatment, respectively. Since the RBM Raman shift is inverselyproportional to the nanotube diameter34, the appearance of anabd d 2dθθθa2a1θθ(n,m)(2n,2m)cFig. 1 | Coalescence of nanocarbon molecules. Coalescence of fullerenes (a) andcarbon nanotubes (b) into large coalesced molecules. θ and d in (b) are the chiralangle and diameter, respectively. (c) Structure of carbon nanotubes specified bythe chiral indices (n,m), which represent the coordinates of the chiral vector (bluesolid arrow) with respect to the basis a1 and a2. The chiral vector (Ch) connects twocarbon atoms (open and filled circles) on a graphene plane, and a nanotube isformed by rolling to connect them (gray and blue solid lines). The diameter d isdescribed using Ch����=π. Geometrically, if two nanotubes with identical (n, m)structures (the chiral vectors are indicated by the blue solid and dotted arrows) arecoalesced into one thick nanotube, the chiral angle θ of the nanotubes before andafter coalescence can be preserved, and the diameter d doubles (chiral vectorindicated as a red arrow)4,36; thus, the chiral indices (n, m) after the coalescencereaction are expected to be (2n, 2m). In (b), the coalescence reaction of two (6,5)nanotubes into a (12,10) nanotube is shown.Article https://doi.org/10.1038/s41467-025-56389-6Nature Communications |         (2025) 16:1093 2www.nature.com/naturecommunicationsadditional RBM peak (indicated by an asterisk) at approximately halfthe RBM Raman shift of the (6,5) nanotube (indicated by a gray-filledcircle) also supports the formation of nanotubes with twice thediameter.To confirm that the emergence of the additional absorption peakwas due to the nanotube coalescence more directly, we performedTEM observations using aberration-corrected TEM on thin (6,5)membranes before and after heat treatment (see Methods for details).Because the nanotube membranes prepared for optical transmissionmeasurements were too thick, we prepared thinner membranes forTEMobservations. After heat treatment, someparts of themembraneswere broken and missing, but we managed to find isolated or smallbundled nanotubes that allowed TEM observations. In the repre-sentative TEM images of the (6,5) nanotube membrane before heattreatment, we observed only small-diameter (d ≈ 0.75 nm) nanotubes(Supplementary Fig. 3a). Although some damage was found on thenanotube sidewalls presumably because of the electron beam irra-diation during the TEM observation, the observed diameter(d ≈0.75 nm) and chiral angle (θ ≈27°) were consistent with those of(6,5). In contrast, in the membrane after heat treatment (Fig. 2e), weobserved many nanotubes with diameters of ≈1.5 nm that is nearlytwice the diameter of the (6,5) nanotubes. Figure 2f compares theobserved high-resolution TEM image of a nanotube formed after heattreatment with a diameter of ≈1.5 nm to the simulated TEM image of a(12,10) nanotube (seeMethods for details of the simulation). Themoirépattern, caused by the interference of the hexagonal carbon networktextures on the top and bottom surfaces of the nanotube, is consistentwith that expected for a (12,10) nanotube.We also confirmed the chiralangle θ ≈27° through the nanobeam diffraction pattern as shown inFig. 2g (see Methods for details on the diffraction measurements andchiral angle determination). These observations indicate that the ori-ginal (6,5) nanotubes coalesced, with the original chiral angle pre-served after the coalescence of the nanotubes. We also observed asmall number of thicker nanotubes with nearly triple the diameter(Fig. 2e and Supplementary Fig. 3b), suggesting that the coalescence ofthree semiconducting (6,5) nanotubes into a single nanotube withtriple the diameter can also occur.Chirality dependenceFurther, we studied the chirality dependence of the nanotube coales-cence reaction efficiency. We prepared single-chirality nanotubemembranes in which the major components were (10,0), (9,1), (9,2),(8,3), (6,5), and (6,6) (see Methods). Among them, only (6,6) was themetallic type, and the others were semiconducting. These nanotubeshad similar diameters (≈0.7–0.8 nm) with different chiral angles, asshown in Fig. 3a. Therefore, we focused on the dependence of thecoalescence efficiency on the chiral angle in this experiment. As shownin the photographs in Fig. 3a, the nanotube dispersions exhibited avariety of colors, reflecting the chirality-dependent resonanceenergiesof the second sub-band excitons32,35. Figure 3b shows the opticalabsorption of each (n, m) membrane after heat treatment at 1000 °C(those before heat treatment are shown in Supplementary Fig. 4). Theemergence of distinct absorption peaks (indicated by the asterisks) at≈0.5–0.6 times the photon energy of the first sub-band exciton reso-nance (indicated by the filled circles) was observed for near-armchair(6,5) and armchair (6,6) species, both with large chiral angles of ≈30°.These results are in stark contrast to those of the near-zigzag andzigzag species. Only a very small peak [(8,3), (9,2)] or almost no peak[(10,0), (9,1)] was observed for these nanotube types with chiral anglesof ≈0°. Figure 3c shows the coalescence efficiency as a function of(6,5) VacuumfiltrationDispersion0.30.20.1logT1.61.20.80.4Photon energy (eV)2001000I (arb. u.) 300200100R (cm 1)After**RBM**0.60.40.20logT1.61.20.80.4Photon energy (eV)2001000I (arb. u.) 300200100R (cm 1)BeforeRBMc dMembraneVacuum heating900–1000 ºCBefore Afterba3 nm3 nm 3 nm0.75 nm 1.50 nm 2.25 nme1 nm 1 nmObservation(12,10)Simulationf g2 nm−1Fig. 2 | Evidenceof efficient carbonnanotube coalescencewithpreservedchiralangles. a Schematic of the (6,5) carbon nanotube (left) and a photograph of itsdispersion (right). b Photographs of the (6,5) nanotubemembrane before and aftervacuum heat treatment at 1000 °C. Scale bar, 2mm. Optical absorption (−logT,whereT is transmittance) spectra of (6,5) nanotubemembranes before (c) and after(d) the vacuum heat treatment. The insets show the RBM features of the Ramanspectra. I and ωR are intensity and Raman shift. The gray filled circles indicate thepeaks corresponding to the (6,5) nanotubes. e Transmission electron microscope(TEM) images obtained for the samples after vacuumheat treatment at900 °C.Barswith different colors indicates lengths corresponding to diameters of typicalnanotubes found in the sample. f Comparison of the observed TEM image of thegenerated nanotube with d ≈ 1.5 nm and simulated one for a (12,10) nanotube. g Ananobeam diffraction pattern for the newly generated nanotube with d ≈ 1.5 nm.Two sets of diffraction spots are positioned at the apices of regular hexagons(yellow and blue). The gray arrow indicates the direction of the CNT axis. Insetindicates the real space TEM image of the observed nanotube. Scale bar, 2 nm.Source data are provided as a Source Data file.Article https://doi.org/10.1038/s41467-025-56389-6Nature Communications |         (2025) 16:1093 3www.nature.com/naturecommunicationschiral angle. The efficiency for each (n, m) type, ηnm, was calculatedusing the integrated peak area of the original S11 of (n, m) nanotubesbefore heat treatment ðAbeforenm Þ and that of emerged (2n, 2m) nano-tubes after heat treatment ðA2n2mÞ (see Methods and SupplementaryFigs. 5 and 6 for details) as ηnm =A2n2m=Abeforenm . Under the currentexperimental conditions, some portions of the nanotubes were lostafter heat treatment, but the remainingnanotubeswereconverted intothe doubled (2n, 2m) species. For the (near) armchair (6,5) and (6,6)nanotubes, the efficiency reached ≈ 10%–20%. Regarding the contentratio of the emerged (2n, 2m) species in the final sample, defined asA2n2m=ðAafternm +A2n2mÞ, where Aafternm is the integrated peak areas of theoriginal S11 of (n, m) nanotubes after heat treatment, they were ≈20%–40% for the (6,5) and (6,6) species. It should be noted that thesevalues represent an estimated lower limit of the content ratio, as thesmall contribution of the (2n, 2m) S22 exciton absorption expected atalmost the same photon energy is neglected. The content ratio of theemerged species is considered to be sufficiently high to not onlymodify the macroscopic optical properties of nanotube aggregatesbut also considerably modify their electric transport properties (seeMethods and Supplementary Fig. 2).Coalescence of the enantiomersMoreover, we examined the possibility of coalescence between the(6,5) and (5,6) nanotubes (equivalent to (11, − 5)), which areenantiomers of each other, because, in addition to the (n,m) + (n,m)→(2n, 2m) case confirmed above, (n,m) + (m, n)→ (n +m, n +m) (Fig. 3d)may also be geometrically allowed. The results indicate that doubled(12,10) and (10,12) species aremuchmore abundant than (11,11) speciesin the sample after heat treatment, as follows. We adjusted the nano-tube dispersion to contain equal amounts of enantiomers of the (6,5)and (5,6) nanotubes (see Methods) and examined the possibility ofgenerating (12,10) and (11,11) nanotubes. If the possibility of coales-cence is perfectly equal among all possible cases, the probability ofgenerating either (12,10) or (10,12) is equal to that of producing (11,11)nanotubes. The (12,10) and (10,12) nanotubes should show the sameexciton absorption peak at the photon energy of 0.67 eV, and anexciton peak M11 for (11,11) species of comparable intensity must alsobe observed at a photon energy of ≈1.8 eV14,33. Figure 3e, f shows theabsorption spectrum after the reaction. As a result, the M11 excitonabsorption peak for (11,11) metallic nanotubes was not as clearlyobserved as that for (12,10) species in the optical absorption spectra.Although a small peak seemingly exists at around 1.7–1.8 eV, as indi-cated by the open triangles in Fig. 3f, this peak also exists before heattreatment; thus, it cannot be assigned solely to (11,11).Mechanism of the coalescenceLet us nowdiscuss themechanismof the efficient coalescence reactionthat allowed for only nanotubes with relatively large chiral angles (i.e.,aa2a1(10,0)(9,2)(9,1)(8,3)(6,6)(6,5)(10,0) (9,1) (9,2)(8,3) (6,5)(6,6)Zigzag (θ = 0º)Armchair (θ = 30º)b0.160.10logT1.20.80.4EP (eV)0.130.07logT1.61.00.4EP (eV)0.240.10logT1.30.90.5EP (eV)*0.350.13logT1.61.00.4EP (eV)0.350.15logT3.62.41.2EP (eV)0.300.10logT1.61.00.4EP (eV)***(6,6)θ = 30º(6,5)θ = 27º(8,3)θ = 15.3º(9,2)θ = 9.8º(9,1)θ = 5.2º(10,0)θ = 0ºc20100Coalescence efficiency (%)302520151050Chiral angle (°)d ef1.00.80.60.40.20logT0.60.50.40.30.2logT1.91.51.10.70.3Photon energy (eV)(6,5) & (5,6)BeforeAfter(12,10) & (10,12)Enantiomers(6,5) (5,6)(11,11)?Fig. 3 | Chirality dependence of the coalescence reaction. a Map of (n, m)nanotubes examined in this studywith optical images of the dispersions of each (n,m) carbonnanotube.bOptical absorption (−logT, whereT is transmittance) spectraof the membranes after vacuum heat treatment at 1000 °C (the one for (6,5) is thesame as that in Fig. 2d). EP, photon energy. The filled circles and asterisks indicatethe lowest energy excitonpeaksof (n,m) and (2n, 2m) nanotubes, respectively. Thesolid arrows indicate the photon energy of the lowest energy exciton peak of (2n,2m) nanotubes, but no peak appears. The insets show the schematics of nanotubeswith two lines showing the chiral angles. c Efficiency of the coalescence reaction asa function of chiral angle. The error bar for the data at θ = 30° represents thestandard error obtained by performing multi-peak fitting to the spectral data of asingle representative sample, using Igor Pro® software. These errors for other chiralangles are negligible and not shown. d Schematic depicting the coalescence of theenantiomers of (6,5) and (5,6) nanotubes. Optical absorption spectrum of (6,5) and(5,6) nanotube membranes before (e) and after (f) heat treatment. The open tri-angles indicate the expected photon energy corresponding to the first sub-bandexciton of (11,11) nanotubes. Source data are provided as a Source Data file.Article https://doi.org/10.1038/s41467-025-56389-6Nature Communications |         (2025) 16:1093 4www.nature.com/naturecommunicationsarmchair and near-armchair types). According to previous theoreticalstudies on nanotube coalescence36, from a geometric point of view,any (n,m) structure is allowed to coalesce into the (2n, 2m) structure,and the coalescence between (n,m) and its enantiomer (m, n)may alsooccur. However, the experimental results contradict this prediction.Therefore, there must be an unknown mechanism that drives thestriking difference in the coalescence efficiency depending on thechiral angle. Here, we propose a sequential bond cleavage andrecombination (SBCR) mechanism to understand the observed chiral-angle dependence of the coalescence efficiency (Fig. 4). In the SBCRscenario, once a few C–C bonds are connected between two (n, m)nanotubes (yellow highlighted C–C bonds in Fig. 4a, b), the next can-didate C–C bond to be cleaved and recombined is the one next to thejust-connected bond (the red and blue highlighted C–C bonds (solidlineswith scissors to be cleaved, anddashed lines to be recombined) inFig. 4b) because of the highest strain on it among all C–Cbonds due tostress concentration. The high strain on the next bond drives thecleavage of the bond in the original (n, m) nanotubes and the forma-tion of C–Cbonds between the two (n,m) nanotubes. The coalescenceof the entire nanotube can be achieved if this reaction sequentiallyoccurs as a chain reaction.Figure 4c, d shows the geometry of the C–C bonds (red and bluehighlighted C–C bonds) to be cleaved and recombined in the SBCRscenario for zigzag (10,0) and near-armchair (6,5) nanotubes, respec-tively. As shown in Fig. 4c, the distance between the candidate carbonatoms of the two (10,0) nanotubes that must be connected increaseswhen the axes of the original nanotubes should be kept parallel (theright panel in Fig. 4c); this constraint is natural for the nanotubes in abundle. The situation is similarbetween the enantiomers (6,5) and (5,6)(Fig. 4e). In contrast, those between two (6,5) nanotubes are muchcloser than those in the above cases, and the sequential reaction seemsreadily possible (Fig. 4d). Because of this geometrical difference, underthe SBCR scenario, only large-chiral-angle (near-armchair and arm-chair) (n, m) nanotubes smoothly coalesce to form long (2n, 2m)nanotubes. The cylindrical cross sections of the coalesced (2n, 2m)structure and the two original tubes could not form perfect circles butbecame considerably distorted owing to the strain induced by theconversion junction among one (2n, 2m) structure and two (n, m)structures. The slight redshifts observed in the RBM of the original (n,m) types after coalescence (Supplementary Fig. 7) can be attributed tocircumferential stress resulting from cross-sectional distortion37 inpartially coalesced nanotubes.Figure 4f shows the calculated energy change of partially coa-lesced nanotubes along the SBCR scenario for the coalescence reac-tion of three typical nanotube structures, (6,5) with θ = 27°, (8,3) with θ= 15°, and (10,0) with θ = 0°, and (6,5) and its enantiomer counterpart(5,6), as a function of the number of cleaved and newly formed bondsalong the zigzag direction (see Methods for details). For clarity, weplotted the results up to 12 fused bonds (see Supplementary Fig. 8 forthe results up to all bonds combined, where all the (2n, 2m) nanotubeshave lower energy than the initial two (n,m) nanotubes, ultimately). Inthe simulation, no constraint was imposed on the axis direction of thetwo original nanotubes for simplicity. The energy change required toincrease the number of C–C bonds in the (2n, 2m) structures indicatesthat the coalescence reaction of zigzag (10,0) nanotubes hardly occursbecause of the increasing energy cost to form the next C–C bonds dueto the high strain. There is no energy stabilization for the (8,3) and(6,5) + (5,6) cases either. In contrast, for the (6,5) case, after severalC–C bonds are formed, the energy change due to the formation of onemoreC–Cbond is negative; thus, the reaction could sequentially occuras a chain reaction. These results are consistent with the experimentalresults and support the proposed SBCR mechanism.Finally, wediscuss the effect of gas environment conditions on thecoalescence reaction.Whenwe placed a (6,5) nanotubemembrane in aabf 3210-1-2E / NC (kJ mol1 )12840Number of fused bonds(10,0) (6,5)/(5,6) (8,3) (6,5)d (6,5) (6,5)c (10,0)(10,0)e (6,5) (5,6)Fig. 4 | Mechanism of coalescence with chiral-angle dependence. Schematic ofthe sequential bond cleavage and recombination (SBCR) mechanism from the side(a) and the bottom (b). C-C bonds highlighted in red and blue are those that breakand recombine. In particular, the red and blue highlighted in (b) will now break andrecombine (broken lines). Schematics of the three-dimensional arrangement of thezigzag direction of two (10,0) nanotubes (c), two (6,5) nanotubes (d), and enan-tiomers of (6,5) and (5,6) nanotubes (e) contributing to cleavage and recombina-tion for the coalescence (red and blue lines in each left panel). Each photograph onthe right shows a coalescence reaction in progress demonstrated using amolecularstructure model kit (MOL-TALOU®). f Energy difference ΔE of partially coalescednanotubes relative to that of the initial two nanotubes normalizedby the number ofcarbon atoms NC in the calculation, plotted as a function of the number of cleavedand newly formed bonds along the zigzag direction, as calculated using molecularmechanics simulations. Source data are provided as a Source Data file.Article https://doi.org/10.1038/s41467-025-56389-6Nature Communications |         (2025) 16:1093 5www.nature.com/naturecommunicationshigh-vacuum-sealed quartz tube for heat treatment, interestingly, nosignature of nanotube coalescence could be observed in the opticalabsorption and RBM Raman results even after heat treatment at1000 °C (see Supplementary Fig. 9). The coalescence reaction occur-red only when (6,5) nanotube membranes were placed in a relativelylarge quartz tube, where the vacuum level was maintained at ≈10−4 Pa,but a trace amount of residual and evolved gases similar to the com-position of air was continuously supplied during heat treatment (seeMethods and Supplementary Fig. 10). This result implies that a smallnumber of residual gas molecules promotes the reaction3. Therefore,supplying a small amount of reactive gas molecules, such as oxygen,may enable them to play a role similar to electron irradiation duringTEM observations4, namely, breaking C–C bonds or removing C atomsto initiate andpromote the nanotube coalescence reaction. To test thishypothesis, we conducted heat treatment experiments in an argonatmosphere (1 kPa) while controlling the partial pressure of oxygen gasat 10 Pa and below 10-4Pa, respectively. To prevent excessive oxida-tion, the reaction temperature was set to 600 °C. Figure 5 comparesthe absorption spectra after heat treatment using oxygen gas (Fig. 5a)and that without oxygen gas (Fig. 5b). A distinct exciton absorptionpeakof (12,10) nanotubes appearedonly in the 10 Paoxygen condition.Raman spectroscopy further confirmed that nanotubes with twice theoriginal diameter formed exclusively after reactions conducted in thepresence of oxygen gas (see inset). These findings demonstrate thatnanotube coalescence via heat treatment can occur efficiently evenat much lower temperature (600 °C) than previously reported(≈1700 °C)5 when an appropriate assisting gas is used under anappropriate condition. The previous theoretical study on the reactionof oxygen and nanotubes17 suggests that oxygen atoms have strongenergetic favorability to adsorb on small-diameter carbon nanotubesand form unzipped C-O-C epoxy chains along a direction of minimumangle to the tube axis. If this reaction occurs between two adjacentnanotubes with same chirality, it might help cleaving C-C bonds alongthe red and blue lines in each left panel in Fig. 4c–e, and preferentiallypromote the coalescence reaction via the SBCR mechanism. Furtherdetailed mechanisms of the chemical reaction with oxygen and theoptimal reaction conditions remain to be clarified in the future studies.In conclusion,wedemonstrated the efficient coalescence of (n,m)carbon nanotubes into (2n, 2m) nanotubeswith doubled diameter andthe same chiral angle via heat treatment. The coalesced nanotubesexhibited distinct exciton resonance peaks in the absorption spectra,indicating that a large number of coalesced nanotubes is generated allover the sample with an electronic structure and optical propertiesinherent to the (2n, 2m) nanotubes. In addition to optical absorptionspectroscopy, the doubled diameters of the coalesced nanotubeswereconfirmed via the TEM measurements and the emergence of theadditional RBM peak at half the wavenumber of that of the original (n,m) nanotubes. The preservation of the chiral angle was also directlyobserved in the TEM images. A distinct chiral-angle dependence of thecoalescence efficiency was observed, and an SBCR mechanism wasproposed based on the results. Furthermore, the presence of traceamounts of oxygenwas found to enable the coalescence reaction evenat more than 1000 °C lower temperature than previously reported.These findings provide routes for the structure-selective synthesis oflarge-diameter single-walled carbon nanotubes with well-definedchirality from structure-purified small-diameter nanotubes and forfabricating covalently connected macroscopic nanotube aggregateswith modified optical properties, electric conductivity, thermal con-ductivity, and mechanical strength via postprocessing.MethodsFabrication of carbon nanotube membranesSingle-chirality carbon nanotube membranes were fabricated using avacuum filtration method32. Single-chirality nanotubes with (n, m) =(10,0), (9,1), (9,2), (8,3), (6,5), and (6,6) were separated from thestarting materials of HiPco (NanoIntegris) and CoMoCAT (SG65 orSG65i, Sigma-Aldrich) samples using gel column chromatography29,31.The separated nanotubes were dispersed in pure water with varioussurfactants, including sodium deoxycholate (DOC), sodium cholate(SC), sodium dodecyl sulfate (SDS), and sodium lithocholate (LC). Weprepared dispersions of (10,0) nanotubes (NanoIntegris HiPco,0.12μgmL−1 in 0.3% SC+0.9% SDS + 0.08% LC)31, (9,1) nanotubes(NanoIntegris HiPco, 0.12μgmL−1 in 0.3% SC+0.9% SDS +0.06% LC)31,(9,2) nanotubes (NanoIntegris HiPco, 0.09μgmL−1 in 0.3% SC+0.9%SDS +0.12% LC)31, (8,3) nanotubes (NanoIntegrisHiPco, 0.12μgmL−1 in0.3% SC +0.9% SDS +0.1% LC)31, and (6,5) nanotubes (CoMoCAT SG65,1.14μgmL−1 in 0.5% SC+0.5% SDS +0.03% DOC)29. The enantiomers,(5,6) (equivalent to (11, − 5)) and (6,5) nanotubes were prepared asdescribed in Ref. 30. For the experiments using the enantiomers, wemixed (5,6) nanotubes (CoMoCAT SG65, 1.14μgmL−1 in 0.5% SC+0.5%SDS +0.028% DOC) and (6,5) nanotubes (CoMoCAT SG65, 1.21μgmL−1in 0.5% SC+0.5%SDS +0.026%DOC) to control the enantiomeric ratioin the sample to 1:1. As for metallic (6,6) nanotubes, at first metallicnanotube mixture was obtained as unadsorbed fraction in 0.5% SC +0.5% SDS elution29, then the metallic nanotube fraction was adsorbedto the gel at 26 °C and pH 10.1 adjusted with NaOH, and the adsorbedmetallic nanotubes were eluted by stepwise increase of DOC con-centration (0.005% step in 0.5%SC and0.5% SDS). The (6,6) nanotubeswere obtained at 0.045% DOC (CoMoCAT SG65i, 2.4μgmL−1 in 0.5%SC+0.5% SDS + 0.045% DOC). To confirm that the coalescence reac-tion occurs regardless of the original material, we also prepared (6,5)nanotubes from a different original material (Nopo HiPco)29. Theoptical absorption spectra indicating the coalescence of this materialare shown in Supplementary Fig. 11. In the vacuum filtration process, apolycarbonate membrane filter with a pore size of 100 nm (MERCK,VCTP02500) and a filter holder with an effective filtration area of2.1 cm2 (ADVANTECH, KGS-25) were used. Each dispersed nanotubesolutionwas diluted to below the criticalmicelle concentration of eachsurfactant (0.08%–0.25% (w/v) for DOC, 0.39%–0.65% (w/v) for SC, and0.20%–0.29% (w/v) for SDS), and it was filtered at 50–80 kPa for ≈30min. Following the nanotube solution, hot water (5mL) was pouredinto the filtering system to remove excess surfactants. Then, themembrane filter with the nanotubes was dried in air at 1–3 kPa for30min. After cutting the obtained nanotubemembrane on the filter toa size suitable for transfer onto sapphire substrates, it was immersed inchloroform for 15min to dissolve the filter. The nanotube membranefloating on chloroform was scooped using a sapphire substrate andcleaned using chloroform, ethanol, and acetone in the same order.Coalescence reactionsCoalescence reactions were examined by heating single-chiralitynanotube membranes on a sapphire substrate in a quartz tube usingan electric furnace (Fig. 2b). For high temperature (900–1000 °C) heattreatments, the pressure in the quartz tube was reduced to ≈5.3 × 10−4Pa. Initially, the furnace was heated to 300 °C and kept for 10min toremove residual molecules other than nanotubes in the membrane.Then, the temperature of the furnacewas increased to 900 or 1000 °Cfor 15min for the coalescence reaction. For the low-temperature(600 °C) heat treatment (Fig. 5), the (6,5) nanotube membrane onsapphire substrate was heated to 600 °C and kept for 15min in an Ar(99.9999% purity, (Fig. 5a) or containing 1% oxygen (Fig. 5b)) atmo-sphere controlled at 1 kPa in the quartz tube.Optical spectroscopyOptical absorption spectra weremeasured using the combination of aFourier-transform infrared spectrometer (JASCO, FT/IR-6600) and aUV spectrophotometer (SHIMADZU, UV-1800). Raman spectroscopywas conducted using a micro-Raman setup (Nanophoton, RAMAN-touch and Renishaw, inVia confocal Raman microscope). The excita-tion laser wavelengths for (6,5) and (6,6) species were 532, 488, andArticle https://doi.org/10.1038/s41467-025-56389-6Nature Communications |         (2025) 16:1093 6www.nature.com/naturecommunications785 nm, respectively. In the Raman spectra of (6,5) and (6,6) nano-tubes, we observed not only the RBM peak of the major (n, m) nano-tubes but also those of minority nanotubes included in the membraneowing to incomplete separation. This is because RBM observationrelies on resonant Raman scattering34, and even when a minor com-ponent is resonant to the excitation laser, its Raman signal is stronglyenhanced.Electrical conductivity measurementsTo examine the macroscopic changes in transport properties after thenanotube coalescence reaction throughout the membrane, we per-formed four-terminal electrical resistance measurements of (6,5) and(6,6) membranes before and after the reaction at 900 °C (Supple-mentary Fig. 2). The (6,5) or (6,6) nanotubemembranes were preparedon 5 × 5mm2 sapphire substrates. Gold electrodes were deposited onthe nanotube membranes, and their current–voltage characteristicswere investigated via four-terminal sensing using a source meter(KEITHLEY, 2636B). We found that the direction of the change inresistance was opposite for semiconducting (6,5) and metallic (6,6)nanotubes. The resistance of semiconducting (6,5) nanotubesdecreased after heat treatment (Supplementary Fig. 2a), whereas theresistance of metallic (6,6) nanotubes increased (SupplementaryFig. 2b). This striking difference between the semiconducting andmetallic nanotubes indicates that the observed change was notmerelya consequence of the degradation of the material. The change inconductivity might be attributed to sidewall functionalization or theadsorption of residual gas molecules, including oxygen, which caninduce p-type doping. Larger diameter nanotubes formed via coales-cence have lower band gaps than the original ones and may be morereadily doped. The existence of the small amount of tripled (3n, 3m)nanotubes that are always metallic regardless of the (n, m) of the ori-ginal nanotubes may also affect the result. Although the detailedmechanism of the observed opposite trend is unclear at this stage,these results indicate that the coalescence reaction can be used tocontrol the electrical transport properties of the membranes viapostprocessing.Transmission electron microscopyFor TEM observations, thin nanotube membranes were prepared on amolybdenum grid. The membranes transferred on the grids weretreated at 300 °C in a vacuum for 10min to remove residualmoleculeson the nanotubes and heat-treated at 900 °C for 15min for the coa-lescence reaction. TEM observation was carried out on an aberration-corrected TEM instrument (FEI TitanCubed) at an acceleration voltageof 80 kV under 4 × 10–6 Pa in the specimen column using a mono-chromator for the incident electron beam (ΔE =0.15 eV). Images werecaptured and processed on a CMOS camera (Gatan OneView, Dmode,4096 × 4096 pixels) operated in the binning 2 mode. The images wererecorded at under-focus conditions (defocus value: ca –6 nm) at anelectron dose rate of ca 1 × 106 e– nm–2 s–1, and the exposure time wasset to 1.0 s. TEM image simulationwasperformedby using amulti-sliceprocedure implemented in Elbis software38 using the experimentalobservation conditions described above. Diffraction patterns of indi-vidual nanotubes were obtained using the same microscope at 80 kVby forming a parallel electron beam with a diameter of ≈ 5 nm. Dif-fraction patterns were processed on DifPACK module of DigitalMi-crograph software (Gatan, Inc.) to determine a chiral angle of anindividual nanotube.Estimation of coalescence efficiencyCoalescence efficiencies were calculated based on the total area of theabsorption of the first sub-band optical transition obtained throughfitting procedures (see Supplementary Fig. 5). The absorption spectrawere fitted with a sum of a linear baseline (gray solid line) and Lorentzfunctions for the first sub-band optical transition (red filled areas) andthe phonon sideband (blue filled areas). Supplementary Fig. 6 displaysthe fitting procedure of the absorption spectrumwith the exclusion ofthe background baseline.Molecular mechanics simulationsAvogadro (ver. 1.2.0)39 was used to perform molecular mechanicscalculations to evaluate changes in the total energy as a function of thenumber of connected C–C bonds between two (n, m) nanotubes.Calculations were performed for (10,0) (zigzag, length 38 Å), (8,3)(chiral, length 42 Å), (6,5) (near-armchair, length 42 Å), and (6,5) + (5,6)(near-armchair enantiomers, length 42 Å) nanotubes using MMFF94and MMFF94s as force fields. According to the SBCR mechanismproposed in the main text, C–C bonds were cleaved and recombinedalong the zigzag direction from the center of the nanotube, and theenergy for each step was obtained after structural relaxation.Q-mass analysis and gas compositionResidual and evolved gases during the coalescence reactions weredetected using a quadrupole mass spectrometer (CANON ANELVACORPORATION,M201QA-TDM), whichwas connected to a quartz tubein an electric furnace. The result is shown in Supplementary Fig. 10.SchematicsSchematics of carbon nanotubes, fullerenes, and their coalescedmolecules were drawn using VESTA 340.10NormalizedIntensity300200100Raman shift (cm-1)0.150.100.050logT1.61.41.21.00.80.6Photon energy (eV)10NormalizedIntensity300200100Raman shift (cm-1)0.80.60.40.20logT1.61.41.21.00.80.6Photon energy (eV)****abFig. 5 | Effect of the addition of oxygen gas.Optical absorption (−logT, where T istransmittance) spectra of (6,5) nanotube membranes after heat treatment at600 °C in an 1 kPa argon atmosphere (a) with (10 Pa) and (b) without (<10−4Pa)oxygen gas. The filled circles and asterisks indicate the lowest energy exciton peaksof (6,5) and (12,10) nanotubes, respectively. Insets in (a,b) are Raman spectra of theRBM. The filled circles and asterisks indicate the RBM peaks for (6,5) and (12,10)species, respectively. Source data are provided as a Source Data file.Article https://doi.org/10.1038/s41467-025-56389-6Nature Communications |         (2025) 16:1093 7www.nature.com/naturecommunicationsData availabilityThe data that support the findings of this study are available from thecorresponding author upon request. Source data are provided withthis paper.References1. Yeretzian, C., Hansen, K., Diederichi, F. & Whetten, R. L. Coales-cence reactions of fullerenes. Nature 359, 44–47 (1992).2. Wang,G.-W., Komatsu, K.,Murata, Y. &Shiro,M. Synthesis andX-raystructure of dumb-bell-shaped C120. Nature 387, 583–586 (1997).3. Nikolaev, P., Thess, A., Rinzler, A. G., Colbert, D. T. & Smalley, R. E.Diameter doubling of single-wall nanotubes.Chem. Phys. Lett. 266,422–426 (1997).4. Terrones, M., Terrones, H., Banhart, F., Charlier, J.-C. & Ajayan, P. M.Coalescence of single-walled carbon nanotubes. Science 288,1226–1229 (2000).5. Yudasaka, M. et al. Diameter enlargement of HiPco single-wallcarbon nanotubes by heat treatment. Nano Lett. 1, 487–489 (2001).6. Bandow, S., Takizawa, M., Hirahara, K., Yudasaka, M. & Iijima, S.Raman scattering study of double-wall carbon nanotubes derivedfrom the chains of fullerenes in single-wall carbon nanotubes.Chem. Phys. Lett. 337, 48–54 (2001).7. Nasibulin, A. G. et al. A novel hybrid carbon material. Nat. Nano-technol. 2, 156–161 (2007).8. Wang, M. & Li, C. M. Magnetic properties of all-carbon graphene -fullerene nanobuds. Phys. Chem.Chem. Phys. 13, 5945–5951 (2011).9. Shimizu, T., Lungerich, D., Harano, K. & Nakamura, E. Time-resolvedimaging of stochastic cascade reactions over a submillisecond tosecond time range at the angstrom level. J. Am. Chem. Soc. 144,9797–9805 (2022).10. Georgakilas, V., Perman, J. A., Tucek, J. & Zboril, R. Broad familyof carbon nanoallotropes: classification, chemistry, and applica-tions of fullerenes, carbon dots, nanotubes, graphene, nanodia-monds, and combined superstructures.Chem. Rev. 115, 4744–4822(2015).11. Saito, R., Fujita, M., Dresselhaus, G. & Dresselhaus, M. S. Electronicstructure of chiral graphene tubules. Appl. Phys. Lett. 60,2204–2206 (1992).12. Wildöer, J. W. G., Venema, L. C., Rinzler, A. G., Smalley, R. E. &Dekker, C. Electronic structure of atomically resolved carbonnanotubes. Nature 391, 59–62 (1998).13. Odom, T. W., Huang, J.-L., Kim, P. & Lieber, C. M. Atomic structureand electronic properties of single-walled carbon nanotubes. Nat-ure 391, 62–64 (1998).14. Kataura, H. et al. Optical properties of single-wall carbon nano-tubes. Synth. Met. 103, 2555–2558 (1999).15. Saito, R., Dresselhaus, G. & Dresselhaus, M. S. Trigonal warpingeffect of carbon nanotubes. Phys. Rev. B 61, 2981–2990 (2000).16. Tanaka, T., Jin, H., Miyata, Y. & Kataura, H. High-yield separation ofmetallic and semiconducting single-wall carbon nanotubes byagarose gel electrophoresis. Appl. Phys. Express 1, 114001 (2008).17. Guo, Y., Jiang, L. & Guo, W. Opening carbon nanotubes into zigzaggraphene nanoribbons by energy-optimum oxidation. Phys. Rev. B82, 115440 (2010).18. Yamamoto, T., Watanabe, S. & Watanabe, K. Universal features ofquantized thermal conductance of carbon nanotubes. Phys. Rev.Lett. 92, 075502 (2004).19. Nishihara, T., Takakura, A., Miyauchi, Y. & Itami, K. Ultra-narrow-band near-infrared thermal exciton radiation in intrinsic one-dimensional semiconductors. Nat. Commun. 9, 3144 (2018).20. Dumitrica, T., Hua, M. & Yakobson, B. I. Symmetry-, time-, andtemperature-dependent strength of carbon nanotubes. Proc. Natl.Acad. Sci. USA 103, 6105–6109 (2006).21. Takakura, A. et al. Strength of carbon nanotubes depends on theirchemical structures. Nat. Commun. 10, 3040 (2019).22. Yang, F. et al. Chirality-specific growth of single-walled carbonnanotubes on solid alloy catalysts. Nature 510, 522–524 (2014).23. Sanchez-Valencia, J. R. et al. Controlled synthesis of single-chiralitycarbon nanotubes. Nature 512, 61–64 (2014).24. Arnold, M. S., Stupp, S. I. & Hersam, M. C. Enrichment of Single-walled carbon nanotubes by diameter in density gradients. NanoLett 5, 713–718 (2005).25. Ghosh, S., Bachilo, S. M. & Weisman, R. B. Advanced sorting ofsingle-walled carbon nanotubes by nonlinear density-gradientultracentrifugation. Nat. Nanotechnol. 5, 443–450 (2010).26. Zheng, M. et al. DNA-assisted dispersion and separation of carbonnanotubes. Nat. Mater. 2, 338–342 (2003).27. Tu, X., Manohar, S., Jagota, A. & Zheng,M. DNA sequencemotifs forstructure-specific recognition and separation of carbon nanotubes.Nature 460, 250–253 (2009).28. Liu, H., Nishide, D., Tanaka, T. & Kataura, H. Large-scale single-chirality separation of single-wall carbon nanotubes by simple gelchromatography. Nat. Commun. 2, 309 (2011).29. Yomogida, Y. et al. Industrial-scale separation of high-purity single-chirality single-wall carbon nanotubes for biological imaging. Nat.Commun. 7, 12056 (2016).30. Wei, X. et al. Experimental determination of excitonic band struc-tures of single-walled carbon nanotubes using circular dichroismspectra. Nat. Commun. 7, 12899 (2016).31. Yomogida, Y., Tanaka, T., Tsuzuki, M., Wei, X. & Kataura, H. Auto-matic sorting of single-chirality single-wall carbon nanotubes usinghydrophobic cholates: implications for multicolor near-infraredoptical technologies. ACS Appl. NanoMater. 3, 11289–11297 (2020).32. Nishihara, T. et al. Empirical formulation of broadband complexrefractive index spectra of single-chirality carbon nanotubeassembly. Nanophotonics 11, 1011–1020 (2022).33. Sato, K., Saito, R., Jiang, J., Dresselhaus, G. & Dresselhaus, M. S.Discontinuity in the family pattern of single-wall carbon nanotubes.Phys. Rev. B 76, 195446 (2007).34. Dresselhaus, M. S., Dresselhaus, G., Saito, R. & Jorio, A. Ramanspectroscopy of carbon nanotubes. Phys. Rep. 409, 47–99 (2005).35. Wei, N. et al. Colors of single-wall carbon nanotubes. Adv. Mater.33, 2006395 (2021).36. Kawai, T., Miyamoto, Y., Sugino, O. & Koga, Y. General sum rule forchiral index of coalescing ultrathin nanotubes. Phys. Rev. Lett. 89,085901 (2002).37. Chang, T. Radial breathingmode frequencyof single-walled carbonnanotubes under strain. Appl. Phys. Lett. 93, 061901 (2008).38. Hosokawa, F., Shinkawa, T., Arai, Y. & Sannomiya, T. Benchmark testof accelerated multi-slice simulation by GPGPU. Ultramicroscopy158, 56–64 (2015).39. Hanwell, M. D. et al. Avogadro: an advanced semantic chemicaleditor, visualization, and analysis platform. J. Cheminform. 4,17 (2012).40. Momma, K. & Izumi, F. VESTA 3 for three-dimensional visualizationof crystal, volumetric andmorphologydata. J. Appl. Crystallogr.44,1272–1276 (2011).AcknowledgementsThis work was supported by JST CREST Grant Number JPMJCR18I5(Y.M.), JSPS KAKENHI Grant Numbers JP22K18287 (Y.M.), JP24H00044(Y.M.), JP23H01791 (T.N.), JP23H04874 (K.H.), and JP24K08253 (O.C.),and JST FOREST Grant Number JPMJFR222N (T.N.). The TEM observa-tions were supported by Kyoto University Nano Technology Hub in“Advanced Research Infrastructure forMaterials andNanotechnology inJapan (ARIM)” sponsored by the Ministry of Education, Culture, Sports,Science and Technology (MEXT), Japan. We thank Hiroki Kurata, AtsushiYamaguchi, and Tsutomu Kiyomura for their assistance with TEMobservations shown in Supplementary Fig. 3, and Koichi Okudaira for hisassistance with TEM observations in the main text.Article https://doi.org/10.1038/s41467-025-56389-6Nature Communications |         (2025) 16:1093 8www.nature.com/naturecommunicationsAuthor contributionsY.M. directed the project and wrote the manuscript. Y.M., A.T., and T.N.conceived theconcept. A.T. prepared thenanotubemembranes, carriedout all the experiments and computational simulations except for thepreparation of single chirality enriched nanotube samples and TEMobservations. K.H. and O.C. contributed to TEM observations andsimulations. T.T. and H.K. prepared single chirality enriched nanotubesamples. T.N. supported optical measurements, figure preparation, andwriting themanuscript. All authors contributed to the preparation of themanuscript.Competing interestsThe authors declare no competing interests.Additional informationSupplementary information The online version containssupplementary material available athttps://doi.org/10.1038/s41467-025-56389-6.Correspondence and requests for materials should be addressed toYuhei Miyauchi.Peer review informationNature Communications thanks Ye Fan and theother, anonymous, reviewer(s) for their contribution to thepeer reviewofthis work. A peer review file is available.Reprints and permissions information is available athttp://www.nature.com/reprintsPublisher’s note Springer Nature remains neutral with regard to jur-isdictional claims in published maps and institutional affiliations.Open Access This article is licensed under a Creative CommonsAttribution 4.0 International License, which permits use, sharing,adaptation, distribution and reproduction in any medium or format, aslong as you give appropriate credit to the original author(s) and thesource, provide a link to the Creative Commons licence, and indicate ifchanges were made. The images or other third party material in thisarticle are included in the article’s Creative Commons licence, unlessindicated otherwise in a credit line to the material. If material is notincluded in the article’s Creative Commons licence and your intendeduse is not permitted by statutory regulation or exceeds the permitteduse, you will need to obtain permission directly from the copyrightholder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.© The Author(s) 2025Article https://doi.org/10.1038/s41467-025-56389-6Nature Communications |         (2025) 16:1093 9https://doi.org/10.1038/s41467-025-56389-6http://www.nature.com/reprintshttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/www.nature.com/naturecommunications Coalescence of carbon nanotubes while preserving the chiral angles Results and discussion Coalescence reaction of (6,5) nanotubes Chirality dependence Coalescence of the enantiomers Mechanism of the coalescence Methods Fabrication of carbon nanotube membranes Coalescence reactions Optical spectroscopy Electrical conductivity measurements Transmission electron microscopy Estimation of coalescence efficiency Molecular mechanics simulations Q-mass analysis and gas composition Schematics Data availability References Acknowledgements Author contributions Competing interests Additional information