Solid-state structures of peapod bearings composedof finite single-wall carbon nanotube andfullerene moleculesSota Satoa,b,c, Takashi Yamasakic, and Hiroyuki Isobea,b,c,1

aJapan Science and Technology Agency, Exploratory Research for Advanced Technology, Isobe Degenerate π-Integration Project and bAdvanced Institute forMaterials Research, Aoba-ku, Sendai 980-8577, Japan; and cDepartment of Chemistry, Tohoku University, Aoba-ku, Sendai 980-8578, Japan

Edited by Jerrold Meinwald, Cornell University, Ithaca, NY, and approved May 8, 2014 (received for review April 11, 2014)

A supramolecular combination of carbon nanotube and fullerene,so-called a peapod, has attracted much interest, not solely becauseof its physical properties but also for its unique assembled struc-tures of carbonaceous entities. However, the detailed structuralinformation available was not sufficient for in-depth understand-ing of its structural chemistry or for exploratory research inspiredby novel physical phenomena, mainly because of the severelyinhomogeneous nature of currently available carbon nanotubes.We herein report solid-state structures of a molecular peapod. Thisstructure, solved with a belt-persistent finite carbon nanotubemolecule at the atomic level by synchrotron X-ray diffraction,revealed the presence of a smooth, inflection-free Hirshfeld sur-face inside the tube, and the smoothness permitted dynamicmotion of the C60 guest molecule even in the solid state. This pre-cise structural information may inspire the molecular design ofcarbonaceous machines assembled purely through van der Waalscontacts between two neutral molecules.

molecular bearing | crystal structure | dynamic solid-state structure |host–guest complex

Acarbonaceous supramolecular system called a peapod, i.e.,a host–guest composite of a single-wall carbon nanotube

(SWNT) and fullerene, is attracting considerable interest invarious fields due to its unique electronic and molecular struc-tures (1). Although interesting physical phenomena of peapodsare being discovered, especially in solid-state physics (2–5), littlefundamental and in-depth understanding of peapods has beenaccumulated at the molecular or atomic levels until quite re-cently. The first reports of structural chemistry related to pea-pods appeared through the studies of [10]cycloparaphenylene([10]CPP) (6): Yamago and coworkers (7) first reported a mod-erate level of association (association constant Ka ∼ 103 M ino-dichlorobenzene) with C60 in the solution phase, and solid-statecrystal structures were reported with C60 and C70 by Jasti and co-workers (8) and Yamago and coworkers (9). Although this mod-erate level of association in [10]CPP raised a question regardingthe stability of peapods in general (5, 10), we recently showed thatthe association of belt-persistent tubular molecules, [4]cyclo-2,8-chrysenylenes ([4]CC2,8) (11–13), with C60 was much higher andrecorded a 106- and 109-fold higher association constant in the samemedium (Ka ∼ 109 M) and in benzene (Ka ∼ 1012 M), respectively(Fig. 1A) (14, 15). The level of association in this molecular peapodwas comparable to the one expected from theoretical studies withinfinite SWNT peapods (10) and, to the best of our knowledge, washighest among host–guest complexes in organic media to date. Theuniqueness of molecular recognition in the curved π-systems wasfurther accentuated by the fact that this tight association does nothamper dynamic rolling motions of the guest, providing an in-triguing possibility as a molecular bearing (16). To deepen theunderstanding of tightest host–guest complex composed of twoapolar and neutral components and also to accelerate the de-velopment of carbonaceous molecular machines (17), the struc-tural information of this molecular peapod, especially at the atomic

level, is indispensable. We herein report the solid-state structuresof the peapod bearing. We show that, even in the solid state, thebelt-persistent tubular molecule allows the dynamic motion of theencapsulated C60 molecule. An inflection-free, smooth surface in-side the tube was revealed by a combination of diffraction analysisusing a high-flux X-ray beam (18) and graphical inspection usingthe Hirshfeld surface of the encapsulated C60 probe (19). Theatomic-level structural information at the tube–sphere interfaceshould be valuable and useful for the in-depth understanding ofcurved π-systems, for the discussion of peapods in the solid state,and for the design of peapod molecular machines in the future.

ResultsSolid-State NMR Analysis. We first investigated the solid-statestructure of the peapod bearing (M)-(12,8)-[4]CC2,8⊃C60 throughsolid-state NMR analysis. The encapsulation in [4]CC resulted inan up-field shift of the 13C resonance of C60 (Fig. 1B and SI Ap-pendix, Fig. S1): under magic-angle spinning (MAS) conditions of20 kHz, the resonance of naked C60 appeared at 143 ppm, whereasthe resonance of (M)-(12,8)-[4]CC2,8⊃C60 appeared at 140 ppm.The 3-ppm difference was unequivocally ascribed to the encap-sulation by recording two separate resonances with a mixture ofC60 and (M)-(12,8)-[4]CC2,8⊃C60. Although the resonances from[4]CC under the MAS conditions were not intense, they wereenhanced under the cross-polarization MAS conditions to showthe corresponding resonances (SI Appendix, Fig. S2).

Significance

Carbonaceous entities possessing tubular and spherical shapesspontaneously form a host–guest complex. This supramolecu-lar complex, so-called a peapod, is unique among host–guestpairs in that it is assembled solely by van der Waals interac-tions at the concave–convex interface of sp2-carbon networks.Recently, a molecular version of this supramolecular systemrevealed the presence of the extremely tight association con-comitantly with the dynamic motions of the guest in apolarmedia. In this paper, an atomic-level structure of the molecularpeapod is revealed by a crystallographic method to show thepresence of an inflection-free surface inside the tubular mole-cule. Enjoying rotational freedom at this smooth surface, theguest fullerene molecule rolls dynamically even in the solid state.

Author contributions: S.S. and H.I. designed research; S.S., T.Y., and H.I. performed re-search; T.Y. contributed new reagents/analytic tools; S.S., T.Y., and H.I. analyzed data; andH.I. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data deposition: The atomic coordinates and structure factors have been deposited in theCambridge Structural Database, Cambridge Crystallographic Data Centre, Cambridge CB21EZ, United Kingdom (CSD reference nos. CCDC 993074 and CCDC 993075).1To whom correspondence should be addressed. E-mail: isobe@m.tohoku.ac.jp.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1406518111/-/DCSupplemental.

8374–8379 | PNAS | June 10, 2014 | vol. 111 | no. 23 www.pnas.org/cgi/doi/10.1073/pnas.1406518111

Dow

nloa

ded

by g

uest

on

June

25,

202

0

As demonstrated previously with naked C60, the solid-stateNMR spectra provide the evidence for dynamic motion in thesolid state. Due to the random orientations of the moleculesfixed in the solid state against the external magnetic field, ordi-nary molecules affords a broad powder pattern of resonancesunder static NMR conditions without MAS (20), whereas C60affords a narrow single resonance even under the static NMRconditions by cancelling out the chemical shift anisotropy throughthe dynamic rolling motion in the solid state (21–23). When weconducted the solid-state NMR analysis of (M)-(12,8)-[4]CC2,8⊃C60under static conditions (Fig. 1C and SI Appendix, Fig. S2), we ob-served a narrow single resonance of C60 at 140 ppm along with thebroad powder-pattern resonances of the encapsulating [4]CC host.This observation showed that, even in the solid state, C60 in [4]CCrolled rapidly on the NMR timescale. When we conducted variable-temperature (VT) NMR analysis, the resonance maintained itssymmetric peak shape, and the half-width of the C60 resonance in[4]CC was maintained at 798 ± 53 Hz (4.5 ± 0.3 ppm, 176 MHz)throughout the accessible temperature range of our conventionalinstrument (from 70 °C to –30 °C; SI Appendix, Fig. S3). Thisresult indicated that the dynamic motion of C60 at the lowestreachable temperature (–30 °C) was rapid enough to averageaway the chemical shift anisotropy. The detailed kinetics of the

motion should be further investigated, for instance, by a wide-boreNMR instrument specialized for the ultralow temperature analysisof a 13C-enriched specimen (24). It is also important to note that thehalf-width of the C60 resonance is much smaller than in the previousdata with infinite SWNT peapods (3,500 Hz, 35 ppm; 100 MHz)(24). The broadening effect of the previous investigation should beattributable to the presence of various SWNT structures, which, inturn, confirmed the importance of discrete molecular structures forthe precise structural analysis.

Molecular and Packing Structures of Vacant and Filled Tubes. Wethen conducted structural analysis using crystallographic meth-ods. The crystal structure of vacant (12,8)-[4]CC2,8 is describedfirst. A single crystal of (12,8)-[4]CC2,8 was obtained from aracemic mixture and revealed the tubular molecular structurethrough diffraction analysis with monochromated X-rays (BL-1Abeamline; Photon Factory) (25). The average dihedral angle atthe single-bond linkages was 18.48 ± 0.16°, which described thesp2-carbon atoms on a curved plane along the cylindrical axis(Fig. 2A) (13). The average diameter of the tube was measuredat the carbon atoms closest to the equator (2- and 8-positions)and was 14.03 ± 0.04 (14.101 × 2 and 13.951 × 2) Å (13, 26). Themolecules were packed in a thread-in-bead entanglement be-tween enantiomers, a packing motif similar to the packing of the(16,0)-isomer (12), but the molecules of (12,8)-isomer were fur-ther knitted to form a 2D network of the molecules by accom-modating hexyl chains of two different enantiomeric molecules inthe tube. Each of the enantiomers was separately stacked throughan interdigitated entanglement of hexyl chains to form homochiralcolumns of (P)- and (M)-structures, respectively (26).The crystal structure of (M)-(12,8)-[4]CC2,8⊃C60 from X-ray

diffraction analysis is described next (Fig. 2B). Although severedisorders inherent to the dynamic bearing system hampered theanalysis with ordinary X-ray beams, a synchrotron macromolec-ular crystallography beamline (BL41XU; SPring-8) (18) allowedus to solve the complex structure of the peapod at the atomiclevel. In the presence of chlorine atoms of CH2Cl2 molecules ina single crystal of the chiral trigonal P32 space group at –173 °C,we unequivocally assigned the absolute configuration of (12,8)-[4]CC2,8 as (M)-helicity with a reliable Flack parameter of 0.13(13). Note that this assignment finally confirmed the previousconclusion from the spectral and theoretical investigations (11).Upon the encapsulation of C60, the average dihedral angle atthe single-bond linkages was reduced to 11.64 ± 2.18°, and thesmoothness of the curved π-system was further emphasized. Theaverage diameter was 13.95 ± 0.01 (13.976, 13.969, 13.943, and13.925) Å and did not deviate much from the diameter found atthe same position of the vacant system (see above). The smalldeviation indicated that the tubular structure of the vacant hostwas ideally preorganized for the C60 guest (14, 27). As shownin Fig. 2B, the (M)-(12,8)-[4]CC2,8⊃C60 molecules were alignedin a columnar assembly again by the interdigitated hexyl sub-stituents (26). The encapsulation of C60 thus disturbed only thethread-in-bead entanglement without affecting the interdigitatedentanglement and resulted in a similar periodic spacing of [4]CCmolecules, whether vacant or filled, in the homochiral column [13.5Å for (12,8)-[4]CC2,8 and 13.3 Å for (M)-(12,8)-[4]CC2,8⊃C60].

Disorders and Hirshfeld Surfaces of Encapsulated C60 Molecules. Oneof the most remarkable features of the crystal structure was thepresence of disordered C60 molecules. As shown in Fig. 3, weidentified as many as four disordered structures of C60 with anidentical position of the center of gravity to minimize the Rfactors [R1 (observed data) = 0.1111, wR2 (all data) = 0.2999].The severe disorders among four C60 molecules may indicate asmall energy difference among these structures, which should bebeneficial for the dynamic motion in the solid state. Unexpect-edly, close examination of the disorders revealed the presence of

160 140 120 170 140 110ppmppm

C (M)-(12,8)-[4]CC

70 °C

30 °C

–30 °C

B

A

C60

C60

(M)-(12,8)-[4]CC

(M)-(12,8)-[4]CC

(M)-(12,8)-[4]CC

+

R

R

R

R

R

R

R

R

1

34

5

7

9 10 11 28 R= hexyl= C60

Fig. 1. Solid-state NMR analysis of peapod bearing. (A) Chemical structureof (M)-(12,8)-[4]CC2,8⊃C60. (B) Spectra of C60, (M)-(12,8)-[4]CC2,8⊃C60 and amixture of C60 and (M)-(12,8)-[4]CC2,8⊃C60 at 25 °C under MAS conditions.See SI Appendix, Fig. S1, for the whole region. (C) VT NMR spectra of(M)-(12,8)-[4]CC2,8⊃C60 under static conditions without MAS. See SI Appendix,Fig. S3, for all of the data.

Sato et al. PNAS | June 10, 2014 | vol. 111 | no. 23 | 8375

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

June

25,

202

0

anomalous carbon atoms: Two carbon atoms at the opposingcorners of C60 were commonly located at almost identical posi-tions in the crystal, and, as a result, the sum of their occupancy atthis position accumulated to 100%. As a result of columnar as-sembly of the molecules, the carbon anomalies were furtheraligned in the crystal. Although we do not fully understand theorigin of this anomaly at this stage, this observation may indicatethat a subtle difference in the shape of the tube results in biasedorientations of the C60 molecules (28, 29): The two anomalouscarbon atoms define the circumscribing circle of C60 by locatingat the opposite sites, and the anomalous locations let the longestdiameter of the guest escape from the direct contact with thetube wall.Analysis of the crystal structure with the Hirshfeld surface

revealed the details of the peapod assembly (19, 30). By parti-tioning the space of the crystal into nonoverlapping volumes ofmolecules, the Hirshfeld surface of a molecule graphically pro-vides various information, for instance, about the shape of thespace dominated by an electron distribution of the molecule orabout the environment surrounding the molecule (19). TheHirshfeld surface of the encapsulated C60 in the molecular peapodtherefore serves as an inspection probe for the inner space: Itshould provide information about the encapsulating space of thehost as well as the interacting contacts at the surface (31).Reflecting the void space surrounded by hexyl chains over the

peapod as well as the close contacting areas surrounded by [4]CCaround the belt region, the Hirshfeld surface of C60 appeared asa football-shaped surface (Fig. 4; see also SI Appendix, Fig. S5,

and Movies S1–S3 for the additional details). By coloring thesurface based on the curvedness, shape index, and de mappings,we abstracted information on the surrounding environments, i.e.,the tubular inner space. The three color mappings of curvedness,shape index, and de, in short, show the geometric inflection ofthe surface, the convex and concave areas of the surface, and thedistance from the surface to the external atoms, respectively. Thecurvedness mapping on the Hirshfeld surface showed no divid-ing nodes in the region wrapped by [4]CC and indicated thesmoothness of the inner tubular surface without geometric in-flection. The shape index mapping with blue and green areasshowed the presence of concave and flat surfaces of the tube,respectively, and should represent a generally typical shapecharacter of the inner surface of SWNT. The green lines of flatareas that appeared helically around the belt region are so sig-nificant that shows the presence of a chiral surface inside the he-lical tube (32). The green dots on the de mapping appeared undermost of the sp2-carbon atoms of [4]CC and showed the presenceof efficient Chost–Cguest contacts distributed evenly over the tube.This observation confirmed that the tightest association of the host–guest complex to date has been achieved purely by nondirectionalvan der Waals interactions between two neutral molecules.A comparison with the relevant molecular peapod (7, 8),

[10]CPP⊃C60, further revealed structural features that should beimportant for tight association and dynamic structures. Althoughthe Hirshfeld surface revealed the presence of an inflection-freecomplementary surface at the interface of (M)-(12,8)-[4]CC2,8⊃C60,slipped π–π stack motifs between planar hexagons of the host and

C2

C2

C2

C2C8

C8

C8

C8

A B

C2

C2

C2

C2C8

C8

C8

C8

Fig. 2. Molecular structures from synchrotron X-ray diffraction analysis of a single crystal. For molecular structures viewed from the top, the hexyl chains andC60 are shown as wireframe diagrams, and the chrysenylenes are shown in ORTEP diagrams with thermal ellipsoids at the 30% level. The numberings ofcarbon atoms at the 2- and 8-positions are also shown. For the packing structures viewed from the side, the (P)- and (M)-structures are colored in red and blue,respectively. Solvent molecules with disorders are omitted for clarity. (A) Molecular and packing structures of (12,8)-[4]CC2,8. For the molecular structure, the(M)-structure is shown. Disordered hexyl chains are found, and one representative structure is shown. See CCDC 993075 for the crystal data. (B) Molecular andpacking structures of (M)-(12,8)-[4]CC2,8⊃C60. One of the representative structures for four disordered C60 molecules (25% occupancy) and hexyl chains areshown. See Fig. 3 for the details of the C60 disorders and CCDC 993074 for the crystal data.

8376 | www.pnas.org/cgi/doi/10.1073/pnas.1406518111 Sato et al.

Dow

nloa

ded

by g

uest

on

June

25,

202

0

guest were observed for the crystal structure of [10]CPP⊃C60 (8).As shown in SI Appendix, Fig. S5, the Hirshfeld surface analysisof [10]CPP⊃C60 indeed showed apparent signatures of π–πstacking as the result of severe deformation with dividing nodesof inflections. The absence of disordered C60 molecules in [10]CPP⊃C60 also corresponds well to the planar facial recognitionbetween two polygons (8). The difference in molecular recogni-tion, i.e., tube–sphere vs. planar π–π recognitions, may thus haveled to the great difference in the association (7, 14). The structuralcomparison disclosed a distinct difference, albeit subtle at firstglance of chemical structures, in the smoothness of inner surfaceand demonstrated a unique nature of van der Waals interface thatis highly sensitive to molecular shapes.

DiscussionSolid-state dynamic motions of C60 in a finite SWNT moleculewere revealed by NMR analysis. The rapid bearing motion in thesolid state suggests interesting physical phenomena to be ex-plored in solid-state materials science, and the detailed kineticsof the motion is of immediate interest. A smooth, inflection-freeinner surface of the tubular molecule was disclosed by the crystalstructure of the molecular peapod, which should be a distinctivestructural feature of peapods in general. This structural in-formation on a smoothly curved π-tube shed the first light on the“mysterious world that exists inside the carbon nanotube” (4)at the atomic level by clarifying the structural origins of uniquemolecular recognition at the interfaces of curved π-systems. A

Fig. 3. Disorder breakdown with different colors on disordered C60 molecules in (M)-(12,8)-[4]CC2,8⊃C60. (A) Molecular structure showing the disordered C60

molecules with highlights of carbon anomalies shown in ball models. (B) Molecular structures with C60 carbon atoms in ORTEP diagrams with thermal ellipsoidat the 20% level. Although one might postulate that two anomalous atoms could function as an axis for shaft motions, the directions of the thermal ellipsoidsin the disordered structures do not support such dynamic single-axis motions. The disordered structures provided static snapshots at –173 °C, and the biasedorientations with anomalous positions may correspond to the presence of a local energy minimum around this location. (C) Columnar packing structuresshowing the positions of anomalous carbon atoms with sphere models. For the structure viewed from the side, a transparent van der Waals surface of(M)-(12,8)-[4]CC2,8⊃C60 is also shown.

Sato et al. PNAS | June 10, 2014 | vol. 111 | no. 23 | 8377

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

June

25,

202

0

recent polar host–guest complex of cucurbit[7]uril and ferro-cenes demonstrated a unique synthetic complex assembled witha delicate match between hydrophobicity, charge, and size inaqueous media to rival the tightest complex in Nature (33, 34),and this peapod complex now added another novel entry ofsynthetic design for neutral host–guest complex in apolar organicmedia to be assembled, tightly and solely, by van der Waalsinteractions at the convex–concave sp2-carbon surfaces. The com-bination of the complementary tube–sphere shapes and non-directional van der Waals interactions further plays a key role inthe dynamic motion of the guest and should be explored forfriction-free molecular machines (14–16). The concomitantpresence of the tight binding and the dynamic supramolecularcomplex should be of great interest for the theoretical studies,and the helical environment found on the Hirshfeld surface is ofspecial interest for its effect on the dynamic motion. Further-more, because the peapod bearing of discrete molecules providesaccess to a homochiral columnar assembly in the crystalline solidstate, an interesting possibility of synchronized orientations ofsingle-axis motion of the peapod bearing may be exploited (35,36). The manipulation of the periodicity of C60 sites in the tu-bular column through designing the interdigitated chains asa spacer is also of great interest (37).

Materials and MethodsSynthesis. The synthesis of [4]CC and (M)-(12,8)-[4]CC2,8⊃C60 was conductedby the methods reported in the literatures (11, 14). The racemic mixture and(M)-isomer of (12,8)-[4]CC2,8 was obtained by preparative HPLC using ODScolumns (20ϕ × 250 + 250 + 250 mm; Kanto Chemical; eluent, 50% (vol/vol)methanol/CH2Cl2) and cholester columns (20ϕ × 250 + 250 mm; NacalaiTesque; eluent, 40% (vol/vol) methanol/CH2Cl2), respectively.

Solid-State NMR Analysis. A specimen in powder form was loaded in a 2.5ϕ ×3.5-mm rotor tube and analyzed on Bruker AVANCE III 700 spectrometer(176 MHz for 13C). The chemical shift was externally referenced to the res-onance of glycine at 176.03 ppm.

Crystallographic Analysis of (12,8)-[4]CC. A single crystal of racemic (12,8)-[4]CCwas obtained from a solution in methanol/CH2Cl2 [∼1:1 (vol/vol)] at 25 °C in a-loosely sealed vial. The single crystal was mounted on a thin polymer tip withcryoprotectantoil and frozenat –178 °Cwith flash-cooling. Thediffractionanalysisof a single crystal with a synchrotron X-ray sourcewas conducted at –178 °C at theBL-1A beamline at KEK PF (25) with a diffractometer equipped with a DectrisPILATUS 2M-F PAD detector. The collected diffraction data were processed withtheHKL2000 softwareprogram (38). The structurewas solvedby a charge flipping

method (39) and refined by full-matrix least-squares on F2 using the SHELX pro-gram suite (40) running on the Yadokari-XG 2009 software program (41). Geo-metrical restraints on the alkyl chains and the solvent, i.e., DFIX, SADI, and SIMU,were used in the refinements. The nonhydrogen atoms in the aromatic part andthe solventwereanalyzedanisotropically,whereas thenonhydrogenatoms in thedisordered alkyl chains were partially analyzed isotropically. Hydrogen atomswere input at calculated positions and refined with a ridingmodel. The details ofthe crystal data are summarized in SI Appendix, Table S1. Two level-A alerts aresuggested by the PLATON/CIF check program: Despite the use of a synchrotronsourceand trials for several crystals, thedataqualitywasnothighenough toavoidthe alerts, and datawere only collected to a resolution of∼1.2Å. The lack of high-angle diffraction data can be attributed to the solvent disorder.

Crystallographic Analysis of (M)-(12,8)-[4]CC2,8⊃C60. A single crystal of(M)-(12,8)-[4]CC⊃C60 was obtained from a solution in methanol/CH2Cl2 [∼1:1(vol/vol)] at 25 °C in a loosely sealed vial. The single crystal was mounted ona thin polymer tip with cryoprotectant oil and frozen at –258 °C with flash-cooling, and the temperature was gradually raised to –173 °C. The diffrac-tion analysis of a single crystal with a synchrotron X-ray source was con-ducted at –173 °C at the BL41XU beamline at SPring-8 (18) with adiffractometer equipped with a Rayonix MX225HE CCD detector. Thecollected diffraction data were processed with the HKL2000 software pro-gram (38). The structure was solved by a direct method (42) and refined byfull-matrix least-squares on F2 using the SHELX program suite (40) runningon the Yadokari-XG 2009 software program (41). Geometrical restraints, i.e.,DFIX, DANG, and SIMU on the alkyl chains and the solvent and SIMU on thefour C60 molecules with 25% occupancy for each C60 modeled as rigid bodies,were used in the refinements. All of the nonhydrogen atoms were analyzedanisotropically. Hydrogen atoms were input at calculated positions and re-fined with a riding model. Due to the severe disorder and fractional occu-pancy, the electron density attributed to some solvent molecules was notproperly modeled, and the structures were refined without these solvents bythe PLATON Squeeze technique (43, 44). The details of the crystal data aresummarized in SI Appendix, Table S2. Two level-A alerts are suggested by thePLATON/CIF check program: Despite the use of a synchrotron source and trialsfor several crystals, the data quality was not high enough to avoid the alerts,and data were only collected to a resolution of ∼0.95 Å. The lack of high-anglediffraction data can be attributed to the presence of the solvent disorder.

ACKNOWLEDGMENTS. We thank Dr. Y. Nishiyama (JEOL) for helpful dis-cussion, Dr. H. Sato (Rigaku) for his support in the diffraction analysis,KEK PF (Research 2013G640) and SPring-8 (Research 2013B0042) for the useof the X-ray diffraction instruments, Central Glass Company for the gift ofhexafluoroisopropanol for the synthesis, Ms. A. Yoshii (Tohoku University)for the preparation of starting materials, and the Analytical Center forGiant Molecules (Tohoku University) for the use of the solid-state NMRinstruments. This study was partly supported by Grant-in-Aid for ScientificResearch, KAKENHI (24241036, 25107708, and 25102007).

1. Smith BW, Monthioux M, Luzzi DE (1998) Encapsulated C60 in carbon nanotubes.Nature 396(6709):323–324.

2. Krive IV, Shekhter RI, Jonson M (2006) Carbon “peapods”—a new tunable nanoscalegraphitic structure. Low Temp Phys 32(10):1171–1194.

3. Monthioux M (2002) Filling single-wall carbon nanotubes. Carbon 40(10):1809–1823.

4. Iijima S (2002) Carbon nanotubes: Past, present, and future. Physica B 323(1-4):1–5.

5. de Juan A, Pérez EM (2013) Getting tubed: Mechanical bond in endohedral de-rivatives of carbon nanotubes? Nanoscale 5(16):7141–7148.

6. Iwamoto T, Watanabe Y, Sakamoto Y, Suzuki T, Yamago S (2011) Selective and randomsyntheses of [n]cycloparaphenylenes (n = 8-13) and size dependence of their electronicproperties. J Am Chem Soc 133(21):8354–8361.

7. Iwamoto T, Watanabe Y, Sadahiro T, Haino T, Yamago S (2011) Size-selective en-capsulation of C60 by [10]cycloparaphenylene: Formation of the shortest fullerene-peapod. Angew Chem Int Ed Engl 50(36):8342–8344.

curvedness

–1.2 0.4

shape index

–1.0(concave)

1.0(convex)

de

1.0 Å 2.0 Å

Fig. 4. Hirshfeld surface of (M)-(12,8)-[4]CC2,8⊃C60 with disordered structures. Solvent molecules with disorders are omitted for clarity. Curvedness, shape index, andde are mapped in colors over the Hirshfeld surfaces. See SI Appendix, Fig. S4, for the complementary surfaces of [4]CC andMovies S1–S3 for more detailed inspection.

8378 | www.pnas.org/cgi/doi/10.1073/pnas.1406518111 Sato et al.

Dow

nloa

ded

by g

uest

on

June

25,

202

0

8. Xia J, Bacon JW, Jasti R (2012) Gram-scale synthesis and crystal structures of [8]- and[10]CPP, and the solid-state structure of C60@[10]CPP. Chem. Sci 3(10):3018–3021.

9. Iwamoto T, et al. (2013) Size- and orientation-selective encapsulation of C70 by cy-cloparaphenylenes. Chemistry 19(42):14061–14068.

10. Okada S, Saito S, Oshiyama A (2001) Energetics and electronic structures of encap-sulated C60 in a carbon nanotube. Phys Rev Lett 86(17):3835–3838.

11. Hitosugi S, Nakanishi W, Yamasaki T, Isobe H (2011) Bottom-up synthesis of finitemodels of helical (n,m)-single-wall carbon nanotubes. Nat Commun 2(10):492.

12. Hitosugi S, Yamasaki T, Isobe H (2012) Bottom-up synthesis and thread-in-beadstructures of finite (n,0)-zigzag single-wall carbon nanotubes. J Am Chem Soc 134(30):12442–12445.

13. Matsuno T, et al. (2014) Geometric measures of finite carbon nanotube molecules:A proposal for length index and filling indexes. Pure Appl Chem 86(4):489–495.

14. Isobe H, Hitosugi S, Yamasaki T, Iizuka R (2013) Molecular bearing of finite carbonnanotube and fullerene in ensemble rolling motion. Chem Sci 4(3):1293–1297.

15. Hitosugi S, et al. (2013) Assessment of fullerene derivatives as rolling journals in a fi-nite carbon nanotube bearing. Org Lett 15(13):3199–3201.

16. Cumings J, Zettl A (2000) Low-friction nanoscale linear bearing realized from multi-wall carbon nanotubes. Science 289(5479):602–604.

17. Kay ER, Leigh DA, Zerbetto F (2007) Synthetic molecular motors and mechanicalmachines. Angew Chem Int Ed Engl 46(1-2):72–191.

18. Hasegawa K, et al. (2013) SPring-8 BL41XU, a high-flux macromolecular crystallog-raphy beamline. J Synchrotron Radiat 20(Pt 6):910–913.

19. McKinnon JJ, Spackman MA, Mitchell AS (2004) Novel tools for visualizing and exploringintermolecular interactions in molecular crystals. Acta Crystallogr B 60(Pt 6):627–668.

20. Mehring M (1983) Principles of High Resolution NMR in Solids (Springer, Berlin), 2nd Ed.21. Tycko R, et al. (1991) Molecular dynamics and the phase transition in solid C60. Phys

Rev Lett 67(14):1886–1889.22. Yannoni CS, Johnson RD, Meijer G, Bethune DS, Salem JR (1991) 13C NMR study of the

C60 cluster in the solid state: Molecular motion and carbon chemical shift anisotropy.J Phys Chem 95(1):9–10.

23. Johnson RD, Yannoni CS, Dorn HC, Salem JR, Bethune DS (1992) C60 rotation in thesolid state: Dynamics of a faceted spherical top. Science 255(5049):1235–1238.

24. Matsuda K, Maniwa Y, Kataura H (2008) Highly rotational C60 dynamics inside single-walled carbon nanotubes: NMR observations. Phys Rev B 77(7):075421.

25. Yamada Y, et al. (2013) Data management system at the photon factory macromo-lecular crystallography beamline. J Phys Conf Ser 425(1):012017.

26. Matsuno T, Kamata S, Hitosugi S, Isobe H (2013) Bottom-up synthesis and structures ofπ-lengthened tubular macrocycles. Chem Sci 4(8):3179–3183.

27. Wittenberg JB, Issacs L (2012) Supramolecular Chemistry: From Molecules to Nano-materials (Wiley, Chichester, UK), Vol 1, pp 25–43.

28. Koshino M, Solin N, Tanaka T, Isobe H, Nakamura E (2008) Imaging the passage of

a single hydrocarbon chain through a nanopore. Nat Nanotechnol 3(10):595–597.29. Solin N, et al. (2007) Imaging of aromatic amide molecules in motion. Chem Lett

36(10):1208–1209.30. Wolff SK, et al. (2012) CrystalExplorer (University of Western Australia, Crawley, WA,

Australia), Version 3.1.1.31. Makha M, McKinnon JJ, Sobolev AN, Spackman MA, Raston CL (2007) Controlling the

confinement and alignment of fullerene C70 in para-substituted calix[5]arenes. Chemistry

13(14):3907–3912.32. Hitosugi S, et al. (2014) Asymmetric autocatalysis initiated by finite single-wall carbon

nanotube molecules with helical chirality. Org Lett 16(3):645–647.33. Jeon WS, et al. (2005) Complexation of ferrocene derivatives by the cucurbit[7]uril

host: A comparative study of the cucurbituril and cyclodextrin host families. J Am

Chem Soc 127(37):12984–12989.34. Rekharsky MV, et al. (2007) A synthetic host-guest system achieves avidin-biotin af-

finity by overcoming enthalpy-entropy compensation. Proc Natl Acad Sci USA 104(52):

20737–20742.35. Akutagawa T, et al. (2009) Ferroelectricity and polarity control in solid-state flip-flop

supramolecular rotators. Nat Mater 8(4):342–347.36. Setaka W, Yamaguchi K (2013) Order-disorder transition of dipolar rotor in a crys-

talline molecular gyrotop and its optical change. J Am Chem Soc 135(39):14560–14563.37. Ganin AY, et al. (2008) Bulk superconductivity at 38 K in a molecular system. Nat

Mater 7(5):367–371.38. Otwinowski Z, Minor W (1997) Methods in Enzymology, Part A, Macromolecular

Crystallography (Academic, New York), Vol 276, pp 307–326.39. Palatinus L, Chapuis G (2007) Superflip—a computer program for the solution of

crystal structures by charge flipping in arbitrary dimensions. J Appl Cryst 40(4):

786–790.40. Sheldrick GM, Schneider TR (1997) SHELXL: High-resolution refinement. Methods

Enzymol 277:319–343.41. Kabuto C, Akine S, Nemoto T, Kwon E (2009) Release of software (Yadokari-XG 2009)

for crystal structure analyses. J Cryst Soc Jpn 51(3):218–224.42. Burla MC, et al. (2007) IL MILIONE: A suite of computer programs for crystal structure

solution of proteins. J Appl Cryst 40(3):609–613.43. Spek AL (2003) Single-crystal structure validation with the program PLATON. J Appl

Cryst 36(1):7–13.44. van der Sluis P, Spek AL (1990) BYPASS: An effective method for the refinement

of crystal structures containing disordered solvent regions. Acta Crystallogr A 46(3):

194–201.

Sato et al. PNAS | June 10, 2014 | vol. 111 | no. 23 | 8379

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

June

25,

202

0