# Fileset

[Prb155420.pdf](https://mdr.nims.go.jp/filesets/f240d0a2-10d7-41f9-bf8e-7900b7a9eb38/download)

## Creator

Kazuhiro Yanagi, Konstantin Iakoubovskii, Said Kazaoui, Nobutsugu Minami, Yutaka Maniwa, Yasumitsu Miyata, Hiromichi Kataura

## Rights



## Other metadata

[Light-harvesting function of b-carotene inside carbon nanotubes](https://mdr.nims.go.jp/datasets/ba1cd55f-00dd-41e4-947a-f27168cde9c9)

## Fulltext

Light-harvesting function of �-carotene inside carbon nanotubesKazuhiro Yanagi,1,* Konstantin Iakoubovskii,1 Said Kazaoui,1 Nobutsugu Minami,1 Yutaka Maniwa,2Yasumitsu Miyata,1 and Hiromichi Kataura11Nanotechnology Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba 305-8562, Japan2Department of Physics, Tokyo Metropolitan University, Tokyo 192-0397, Japan�Received 31 July 2006; revised manuscript received 31 August 2006; published 18 October 2006�Single-wall carbon nanotubes �SWCNTs� are attractive components for nanoscale electronics, however theiroptoelectronic properties have been limited by the optical characteristics of semiconducting SWCNTs. Toenhance the functionalities of SWCNTs, �-carotene was encapsulated in SWCNTs and its light-harvestingfunction was investigated. The detailed structure of encapsulated �-carotene was clarified using x-ray diffrac-tion and the polarization dependence of the optical absorption spectra. The photoluminescence spectra revealedexcited energy transfer from �-carotene to the SWCNTs.DOI: 10.1103/PhysRevB.74.155420 PACS number�s�: 78.67.Ch, 78.55.Kz, 78.66.TrI. INTRODUCTIONThe unique properties of single-wall carbon nanotubes�SWCNTs�, such as their high aspect ratio, strength, resonantoptical absorption, high carrier mobility, and maximum cur-rent densities, make them an attractive component for na-nometer scale optoelectronic devices.1 Recently, the efficientgeneration2,3 and detection �quantum efficiency �10%��Ref. 4� of light by a single SWCNT has been demonstrated.However, the spectral range of these unique devices is lim-ited by the specific density of states of the semiconductingSWCNTs. Controllable modification of the spectral range isrequired to create efficient optoelectronic devices forSWCNTs.This modification can be achieved by combining aSWCNT with an appropriate molecule �the so-called “nano-tube functionalization”�. Such functionalization is commonlyachieved by a chemical reaction attaching a molecule outsidethe SWCNT. Unfortunately, this process has undesirable sideeffects, such as the creation of defects, as well as quenchingthe optical absorption and luminescence in SWCNTs. A gen-tler and more elegant functionalization strategy is to encap-sulate organic molecules inside the SWCNT without break-ing the chemical bonds.5,6 It is not yet well known howeverwhich molecules are capable of changing the optical proper-ties of SWCNTs via encapsulation.The encapsulation of �-carotene �a carotenoid �Car� witha chemical structure as shown in Fig. 1�a�� in SWCNTs hasbeen reported in our recent study.7 We refer to the producedcomplex as Car@SWCNTs. It is well known that carotenoidsexhibit a light-harvesting function in the pigment-proteincomplexes of photosynthetic bacteria.8 Thus one might ex-pect similar light-harvesting by �-carotene in Car@SWCNTscomplexes as well. We investigated this possibility in thispresent study.In pigment-protein complexes, knowledge of the detailedpositions of carotene and other pigments allows to controlthe light-harvesting process.9 Therefore it is crucial to revealthe structure of the encapsulated dye to understand the en-ergy transfer processes in the dye-SWCNT complexes. InRef. 7, encapsulation of �-carotene inside SWCNTs has beenclarified by optical absorption and Raman measurements.7However, detailed structure and functions of encapsulated�-carotene has not been clearly revealed.Encapsulated �-carotene is not a robust molecule like C60.�The Raman and absorption peaks originating from�-carotene inside the SWCNT disappear upon annealing atapproximately 200 °C for 1 h in vacuum.� Therefore, al-though it is known that high-resolution transmission electronmicroscopy �HRTEM� measurements can directly probe theencapsulation inside SWCNTs, it is difficult to determine thedetailed structure of the encapsulated �-carotene throughHRTEM measurements because the electron beam radiationwould be strong enough to damage the encapsulated�-carotene. Therefore, in this paper, we studied the structureFIG. 1. �a� Chemical structure of �-carotene, and dependence ofthe absorption spectra of SWCNT �b� and Car@SWCNT �c� on theangle � between the stretching direction of the film samples and thepolarization of incident linearly polarized light. The absorption fea-ture associated with the encapsulated �-carotene is labeled as thecarotene band. Absorption intensities of the carotene �squares� andS1 �circles� bands are plotted as functions of � in �d� and �e�,respectively, and simulated �solid lines� by Ai+Bi cos �2 functions.PHYSICAL REVIEW B 74, 155420 �2006�1098-0121/2006/74�15�/155420�5� ©2006 The American Physical Society155420-1http://dx.doi.org/10.1103/PhysRevB.74.155420of encapsulated �-carotene using nondestructive techniquesof x-ray diffraction and the polarization dependence of theoptical absorption spectra. In addition, photoluminescence�PL� was employed to reveal the energy transfer from�-carotene to the SWCNTs.II. EXPERIMENTAL SECTIONA. Sample preparationWe used SWCNT manufactured by laser vaporization inan Ar atmosphere of carbon rods doped with Co and Ni. Thetubes were purified with H2O2, HCl, and NaOH reagents asdescribed previously10 and then annealed in air at 420 °C for20 min. The encapsulation of �-carotene into the SWCNTswas performed as follows. SWCNTs �1 mg� and �-carotene�100 mg, Wako� were dissolved in hexane �100 ml�. Themixture was refluxed for 10 h in an N2 atmosphere. Then thesolution was filtered and washed with tetrahydrofuran �THF�several times to remove nonencapsulated �-carotene mol-ecules. In order to prepare proper nonencapsulated referencesamples, the same procedure has been repeated without�-carotene except for x-ray diffraction �XRD� measure-ments.For polarization-resolved absorption measurements, weprepared stretched polymer films where SWCNTs were dis-persed and aligned to the stretched direction. The films wereprepared as follows. SWCNTs were dispersed in 1% watersolution �5 ml� of deoxycholic acid sodium salt �TokyoKasei Co.�. Polyvinyl alcohol polymer �Kanto Chemical Co.,PVA2000� was added to the solution, and then it was dried ina desiccator. The obtained films were mechanically stretchedup to 4 times at 70 °C.PL and unpolarized absorption measurements were per-formed on solutions �1 cm thick silica cells� prepared withthe following procedure. Nanotubes were dispersed for15 min in 1% D2O solutions of sodium dodecyl-benzene sul-fonate using a tip sonifier �20 kHz, power 100 W�. The dis-persions were ultracentrifuged for 5 h at 150 000 g, and theupper 80% supernatant was collected.B. XRD measurementXRD patterns were recorded using synchrotron radiationwith a wavelength of 0.100 nm at the BL1B beam line at thePhoton Factory, KEK, Japan. All the XRD measurementswere performed at 330 K in vacuum on powder samples.C. Optical characterizationOptical absorption spectra were recorded with a commer-cial Shimadzu spectrophotometer. PL mapping was per-formed with a home-built setup utilizing a tunable Ti-sapphire laser �Spectra-Physics 3900S� or Xe lamp andmonochromator for excitation and a single-grating mono-chromator with an InGaAs diode array for detection.III. RESULTS AND DISCUSSIONA. Polarization-resolved optical absorptionOur previous Raman measurements on Car@SWCNT in-dicate the trans-conformation of the encapsulated �-caro-tene.7 Note that �-carotene is a one-dimensional �-conju-gated molecule, thus it is expected to align to the SWCNTaxis after encapsulation. Therefore, the optical absorptionoriginating from the encapsulated �-carotene �carotene band�should exhibit the same polarization dependence as the hostSWCNT. Figures 1�b� and 1�c� present variations in the ab-sorption spectra of the stretched polymer films ofCar@SWCNT and SWCNT with the angle � between thestretching direction and the polarization vector of the inci-dent linearly polarized light. Figure 1 reveals that the caro-tene band exhibits dependence on the polarization angle verysimilar to that of the absorption bands of the SWCNTs.These bands are labeled Si and Mi, i=1–2, and they origi-nate from the ith interband excitations in semiconducting andmetallic SWCNTs, respectively. The absorption intensities ofthe carotene band �squares� and the Si band �circles� areplotted vs � in Figs. 1�d� and 1�e�, respectively. Both inten-sities fit very well with the Ai+Bi cos �2 functions �lines�,where constants Ai account for the incomplete SWCNTalignment. The determined absorbance polarization �= ���−��� / ��� +��� values of the carotene band ��=0.6� are al-most the same as that of the S1 band ��=0.5�. These resultsclearly indicate that the encapsulated �-carotene moleculesare aligned to the SWCNT axis.B. XRD results and their modelingFigure 2 summarizes XRD results. The circles �top curve�present diffraction pattern from the initial SWCNT powder,which has not been subjected to the encapsulation or associ-ated refluxing and/or washing treatment. The profile is domi-nated by the �10� peak around Q�4� sin � /r=4.5 nm−1originating from XRD on SWCNT bundles. The squares�bottom curve� in Fig. 2 show the profile fromFIG. 2. �Color online� X-ray diffraction patterns of SWCNT�upper curve� and Car@SWCNT �bottom curve�. Circles are experi-mental observations, and the solid red �blue� line is the simulationresult for SWCNT �Car@SWCNT�. R, g, and t are the radius ofSWCNT, intertube distance, and thickness of the nanotube bundle,respectively. Dashed black line shows the calculated XRD patternassuming the loss of crystal order.YANAGI et al. PHYSICAL REVIEW B 74, 155420 �2006�155420-2Car@SWCNT. It reveals that the encapsulation procedureresults in strong quenching of the �10� peak. There are twopossible reasons for this decrease: loss of crystal order in theSWCNT bundles or presence of molecules inside theSWCNTs.11 In order to distinguish between those two possi-bilities and to obtain the detailed structural information, weperformed a numerical simulation of these XRD patternswith reference to Refs. 12 and 13. In this simulation, both theencapsulated molecules and the SWCNT walls were approxi-mated by infinitely thin sheets of homogeneous electron den-sity within the carbon covalent networks.First, we analyzed the XRD pattern of empty SWCNTs:The SWNTs of the radius R are separated by the distance gand compose a bundle of thickness t �see inset in Fig. 2�; theradius R is allowed to spread by the value �R. This modelwas successfully fitted into the profile of the nonencapsu-lated SWCNTs �see solid red line in Fig. 2�, yielding thevalues of R=0.69±0.01 nm, �R=0.06 nm, g=0.32±0.01nm, and t=17 nm in agreement with previous results.12,13The determined values of R, �R, and g were assumed to beunaffected by the encapsulation procedure applied here andthus fixed in the further simulations.The loss of crystal order upon the encapsulation proce-dure, as a possible reason for the quenching of the �10� peak,was modeled by decreasing the bundle size t. Dashed blackline in Fig. 2 presents simulation obtained with t=7 nm.Clearly, it accounts for the reduction in the �10� peak, butalso results in the concomitant broadening of all peaks. Thisbroadening is inconsistent with the experiment suggestingthat the loss of crystal order hardly occurred during the en-capsulation procedure. It is noteworthy that disorder wouldenhance the D-band and no enhancement was observed inthe Raman results for Car@SWCNT.7 Therefore, the deter-mined above optimal bundle size t=17 nm was also fixed infurther simulations of Car@SWCNT.Encapsulation of molecules was modeled by putting theminside the SWCNTs and separating them from the SWCNTwalls by the distance   �see Fig. 3�a��. We observed a strongdependence of the simulated XRD pattern on the distance  ,as demonstrated in Fig. 3�b�. As expected, the encapsulationof molecules in SWCNTs also causes the decrease of �10�peak. Another adjustable parameter in this model is theweight percent w of encapsulated molecules inside SWCNT.In Fig. 3, w is set to be 13%. This parameter changes the �10�peak intensity, but it has minor effect on other profile fea-tures. The remarkable feature of the XRD pattern inCar@SWCNT is that the intensity of peak d is larger thanthat of peak c. It is clear that this remarkable feature can onlybe reproduced at around the distance  =0.4. The loss ofcrystal order cannot reproduce this feature. The best fit couldbe obtained with  =0.4 nm and w=12% �see blue solid linein Fig. 2�.The effect of intercalation on the XRD pattern was re-ported in Ref. 14. The intensity of the �10� peak decreaseswith increasing intercalation levels. However, the line shapeand position of the �10� peak are broadened and shifted tolower Q, respectively. Thus the effect of intercalation cannotaccount for the XRD pattern of Car@SWCNT.C. Analysis of the XRD resultsThe results of the preceding section suggest the presenceof encapsulated molecules, with the weight percent of�12%, inside the nanotubes of the Car@SWCNT sample.The weight percent of �-carotene alone has been indepen-dently estimated as �3% from optical absorption measure-ments.7 �If we assume that �-carotene is present in a lineinside a 1.4 nm diameter SWCNT, the filling rate is esti-mated to be approximately 30%, thus �-carotene is effi-ciently encapsulated.� There is a difference in the estimatedweight percent between encapsulated molecules and encap-sulated �-carotene. The difference indicates that the compa-rable amounts of solvent molecules and �-carotene were en-capsulated into the Car@SWCNT nanotubes.Therefore the derived distance   between the encapsu-lated molecules and the tube wall does not directly reflect theposition of �-carotene. However, considering the strong sen-sitivity of the XRD profile to the position of the moleculeinside the SWCNT �see Fig. 3�b��, the derived distance  suggests that both the solvent molecules and �-carotene arepositioned off the SWCNT center at a distance of �0.4 nmto its walls.The off center location of the encapsulated molecules isremarkable and suggests an attractive interaction with theSWCNT walls. Note that the derived distance between theencapsulated molecules and the SWCNT walls �0.4 nm� issimilar to that predicted theoretically �0.33 nm� for thepolyacetylene@SWCNT system.15 It is also similar to theFIG. 3. �a� Simulation model for XRD pattern of Car@SWCNT.�b� The XRD patterns calculated when the distance   from theencapsulated molecule to the nanotube wall is changed from 0 to0.69 nm �center of the tube�.LIGHT-HARVESTING FUNCTION OF �-CAROTENE… PHYSICAL REVIEW B 74, 155420 �2006�155420-3distance between the two-dimensional graphitic layers.16Therefore, it could be �-� interaction which shifts the en-capsulated molecules off center of the SWCNTs.D. Photoluminescence excitation spectraFigures 4�b� and 4�c� present 2D maps of the PL excita-tion and/or emission peak intensities in the SWCNT andCar@SWCNT, respectively. Figure 4�a� shows the corre-sponding absorption spectra measured from the same solu-tions. Here the density of the sample solution ofCar@SWCNT, which was estimated from the peak intensityof S2 band �940 nm�, was adjusted to be the same as that ofSWCNT. No luminescence could be detected in the emissionrange of the isolated �-carotene. Although the luminescenceof the �-carotene was completely quenched by encapsula-tion, PL signals originating from the nanotubes were easilydetected and assigned to the �11,9�, �12,7�, and �13,5�tubes.17 Note that the radii of the �11,9�, �12,7�, and �13,5�tubes �0.69, 0.66, and 0.64 nm, respectively� are similar tothose estimated from the XRD results �0.69 nm�.The peaks in the PL excitation �PLE� range 800 nm to1000 nm in Fig. 4 originate from excitation into the S2 bandfollowed by emission from the S1 band. Two processes couldaccount for the features in the excitation range of 350 nm to600 nm for the Car@SWCNT sample: �i� direct excitationinto the S3 band of SWCNT, and �ii� absorption by�-carotene followed by energy transfer to the SWCNTbands. To reveal the latter contribution, we evaluated thedifference excitation spectra �PLE between the SWCNT andCar@SWCNT for the �11,9�, �12,7�, and �13,5� tubes �seeFigs. 5�b�–5�d�, respectively�. The �PLE signals were de-rived as follows: PLE spectra of the S1 emission were ex-tracted for each tube as indicated by the dotted lines in Figs.4�b� and 4�c�. Those spectra were normalized by intensity inthe 600 nm to 800 nm region where the contribution of�-carotene is negligible. The normalization was necessary tocompensate for slight differences in the dispersion conditionin the SWCNT and Car@SWCNT samples. Then theSWCNT spectra were subtracted from the Car@SWCNTcurves thus revealing the �-carotene contribution to the PLEspectra in Car@SWCNT.Figure 5�a� shows the absorption of the carotene bandobtained as the difference between the absorption ofCar@SWCNT and SWCNT �Fig. 4�a��. Remarkably, the dif-ference PLE spectra, which are similar to the carotene ab-sorption band �dotted red lines�, were clearly observed forFIG. 6. Schematic illustration of energy transfer processes from�-carotene to SWCNT which might occur after photoexcitation ofthe 1 1Bu+ excited state of �-carotene.FIG. 4. �Color online� �a� Absorption spectra of the samplesused in photoluminescence �PL� measurements. Also shown are PLmaps of SWCNT �b� and Car@SWCNT �c� dispersions. Dottedlines indicate excitation profiles for the major SWCNT chiralities,identified as �11,9�, �12,7�, and �13,5�.FIG. 5. �Color online� �a� presents the difference absorptionspectrum �A between Car@SWCNT and SWCNT revealing thecarotene band. �b�, �c�, and �d� show the difference excitation spec-tra �PLE between Car@SWCNT and SWCNT for the �11,9�,�12,7�, and �13,5� tubes, respectively. The carotene band is alsoshown with the dashed red line for comparison.YANAGI et al. PHYSICAL REVIEW B 74, 155420 �2006�155420-4the �11,9� and �12,7� �Figs. 5�b� and 5�c��, but not for the�13,5� tubes �Fig. 5�d��. This observation suggests that pho-toexcitation can be transferred from �-carotene to theSWCNTs, but only to those with relatively large diameters. Itcan be interpreted that the SWCNTs with small diameterscould not incorporate �-carotene.E. Energy transfer from �-carotene to SWCNTIt is commonly known that there are two important singletexcited states in �-carotene for excited energy transfer, 1 1Bu+and 2 1Ag−.8,18 Optical transition between the 1 1Bu+ and theground 1 1Ag− state are allowed �transition energy E�2.6 eV�, but between the 2 1Ag− and 1 1Ag− states �E�1.8 eV� are symmetry forbidden.18 When �-carotene is en-capsulated in SWCNT the energy of 1 1Ag−-1 1Bu+ transitiondecreases by �0.1 eV, as revealed by optical absorption.7Comparison of those energies with the SWCNT transitionsreveals that the energies of the two excited states of�-carotene overlap with the S3 �2.5–3.0 eV� and S2�1.2–1.6 eV� transitions. Figure 6 shows a schematic illus-tration of the possible energy transfer processes from�-carotene to SWCNT. These revealed near-resonant excita-tion conditions might be responsible for the experimentallyobserved energy transfer from �-carotene to SWCNT.IV. CONCLUSION�-carotene can be efficiently encapsulated in SWCNT,and the detailed structure and optical properties of the encap-sulated �-carotene are clarified. The polarization dependenceof absorption spectra indicates that encapsulated �-caroteneis aligned to the tube axis. This experimental observationsuggests that the encapsulation of molecules inside SWCNTcan be used to orient molecules. Analysis of the XRD patternof Car@SWCNT suggests that the encapsulated �-carotene,as well as the solvent, molecules are shifted off the nanotubecenter. PL measurements indicate that encapsulated �-caro-tene can transfer photoexcitation energy to the SWCNT, andthus does exhibit the light-harvesting function. We hope thatthis new and exciting observation will open the door for thefine-tuning, via the encapsulation of various molecules, ofthe optical properties of SWCNTs for numerous optoelec-tronic applications.ACKNOWLEDGMENTSOne of the authors �K.Y.� acknowledges Grant-in-Aidfrom the Foundation of Advanced Technology Institute,the Sumitomo Foundation �Grant No. 050645�, andJSPS.KAKENHI�18740187�. This study was supported inpart by the Industrial Technology Research Grant Program in2003 from the New Energy and Industrial Technology De-velopment Organization �NEDO� of Japan.*Corresponding author. Electronic address: k-yanagi@aist.go.jp1 T. Dürkop, B. M. Kim, and M. S. Fuhrer, J. Phys.: Condens.Matter 18, R553 �2004�.2 J. A. Misewich, R. Martel, P. Avouris, J. C. Tsang, S. Heinze, andJ. Tersoff, Science 300, 783 �2003�.3 J. Chen, V. Perebeinos, M. Freitag, J. Tsang, Q. Fu, J. Liu, and P.Avouris, Science 310, 1171 �2005�.4 M. Freitag, Y. Martin, J. A. Misewich, R. Martel, and P. Avouris,Nano Lett. 3, 1067 �2003�.5 B. W. Smith, M. Monthioux, and D. E. Luzzi, Nature �London�396, 323 �1998�.6 T. Takenobu, T. Takano, M. Shiraishi, Y. Murakami, M. Ata, H.Kataura, Y. Achiba, and Y. Iwasa, Nat. Mater. 2, 683 �2003�.7 K. Yanagi, Y. Miyata, and H. Kataura, Adv. Mater. �Weinheim,Ger.� 18, 437 �2006�.8 N. J. Fraser, H. Hashimoto, and R. J. Cogdell, Photosynth. Res.70, 249 �2001�.9 R. Z. Desamero, V. Chynwat, I. v. d. Hoef, F. J. Jansen, J. Lugten-burg, D. Gosztola, M. R. Wasielewski, A. Cua, D. F. Bocian, andH. A. Frank, J. Phys. Chem. B 102, 8151 �1998�.10 H. Kataura, Y. Maniwa, T. Kodama, K. Kikuchi, K. Hirahara, K.Suenaga, S. Iijima, S. Suzuki, Y. Achiba, and W. Krätschmer,Synth. Met. 121, 1195 �2001�.11 Y. Maniwa, Y. Kumazawa, Y. Saito, H. Tou, H. Kataura, H. Ishii,S. Suzuki, Y. Achiba, A. Fujiwara, and H. Suematsu, Jpn. J.Appl. Phys., Part 2 38, L668 �1999�.12 M. Abe, H. Kataura, H. Kira, T. Kodama, S. Suzuki, Y. Achiba,K. I. Kato, M. Takata, A. Fujiwara, K. Matsuda, and Y. Maniwa,Phys. Rev. B 68, 041405�R� �2003�.13 H. Kadowaki, A. Nishiyama, K. Matsuda, Y. Maniwa, S. Suzuki,Y. Achiba, and H. Kataura, J. Phys. Soc. Jpn. 74, 2990 �2005�.14 X. Liu, T. Pichler, M. Knupfer, and J. Fink, Phys. Rev. B 67,125403 �2003�.15 G. C. McIntosh, D. Tománek, and Y. W. Park, Phys. Rev. B 67,125419 �2003�.16 S. Bandow, F. Kokai, K. Takahashi, M. Yudasaka, L. C. Qin, andS. Iijima, Chem. Phys. Lett. 321, 514 �2000�.17 R. B. Weisman and S. M. Bachilo, Nano Lett. 3, 1235 �2003�.18 K. Onaka, R. Fujii, H. Nagae, M. Kuki, Y. Koyama, and Y. Wa-tanabe, Chem. Phys. Lett. 315, 75 �1999�.LIGHT-HARVESTING FUNCTION OF �-CAROTENE… PHYSICAL REVIEW B 74, 155420 �2006�155420-5