Linearly controlled arrangement of ^{13}C isotopes in single-wall carbon nanotubes
LLinearly controlled arrangement of C isotopes in single-wall carbon nanotubes
J. Koltai, H. Kuzmany, T. Pichler, and F. Simon
2, 3 Department of Biological Physics, E¨otv¨os University,P´azm´any P´eter s´et´any 1/A, H-1117 Budapest, Hungary Universit¨at Wien, Fakult¨at f¨ur Physik, Strudlhofgasse 4, 1090 Wien, Austria Department of Physics, Budapest University of Technology and Economics and MTA-BMELend¨ulet Spintronics Research Group (PROSPIN), P.O. Box 91, H-1521 Budapest, Hungary
The growth of single wall carbon nanotubes (SWCNT) inside host SWCNTs remains a compellingalternative to the conventional catalyst induced growth processes. It not only provides a catalystfree process but the ability to control the constituents of the inner tube if appropriate startingmolecules are used. We report herein the growth of inner SWCNTs from C labeled toluene andnatural carbon C . The latter molecule is essentially a stopper which acts to retain the smallertoluene. The Raman spectrum of the inner nanotubes is anomalous as it contains a highly isotopeshifted ”tail”, which cannot be explained by assuming a homogeneous distribution of the isotopes.Semi-empirical calculations of the Raman modes indicate that this unsual effect is explicable ifsmall clusters of C are assumed. This indicates the absence of carbon diffusion during the innertube growth. When combined with appropriate molecular recognition, this may enable a molecularengineering of the atomic and isotope composition of the inner tubes.
INTRODUCTION
The growth of carbon nanotubes from carbonaceousmaterials, which are encapsulated inside host SWCNTs,remains a compelling catalyst free synthesis method ofSWCNTs. While the growth was originally discoveredfrom encapsulated fullerenes (peapods) [1] under inten-sive electron beam irradiation and heating [2], it was latershown that the inner tube can proceed from virtually anycarbon containing materials [3] including small solventmolecules such as e.g. benzene or toluene, azafullerenes[4], or coronene [5]. This synthesis method of SWCNTshave the advantage of allowing a diameter control de-pending on the diameter of the outer tube and that acatalyst free synthesis is performed, which leads to ultraclean inner tubes [6]. Natural disadvantages of the innertube growth are the hindered ability to remove the innertubes from the inside in a non-invasive manner, and thelack of control over the possible inner-outer tube chiral-ity pairs, whose presence complicates the Raman analysis[7].A possible next step to explore the inner tube syn-thesis from various carbon sources is the combinationof several starting components, e.g. the combination ofco-encapsulated fullerenes and small organic molecules,which could be used e.g. for the growth of heteroatomcontaining inner tubes or for their isotope labeling. Inprinciple, the various carbon source molecules would en-capsulate in a random fashion. However if some sort of amolecular recognition was present, it would be a possibil-ity to control the arrangement of the various components.In addition, it is also required that little carbon atom dif-fusion takes place during the inner tube growth in orderto fully exploit the molecular recognition. It was reportedpreviously [8] that the carbon diffusion is limited alongthe inner tube axis: fullerenes of natural carbon and C were co-encapsulated and a Raman analysis of the innertube modes showed a larger than expected inhomogene-ity of the C isotopes on the resulting inner tubes. Alogical continuation of this effort is to co-encapsulate a C isotope labeled small organic molecule (benzene ortoluene) with fullerenes inside SWCNTs and to study thevibrational modes of the resulting inner tubes.We report the synthesis and Raman characterizationof single-wall carbon nanotubes which are grown in-side host nanotubes from fullerenes and other small or-ganic molecules, benzene and toluene. When the lattermolecules are C isotope labeled, unexpected changes inthe Raman spectra are observed: rather than downshift-ing in a uniform manner (which is expected for homoge-neous doping), a tail develops on the small Raman shiftside of the Raman modes. This indicates a clustering ofthe C isotopes. This is supported by first principles cal-culations, where similar features can only be reproducedwhen a significant clustering of the isotopes is present.This effect is probably related to the clustered natureof the C isotopes on the organic rings and it suggeststhat no carbon diffusion takes place during the inner tubegrowth.
EXPERIMENTAL
The starting SWCNT sample was obtained by the arc-discharge method and it was identical to samples as inprevious studies [3] with a mean diameter of 1 . . ◦ C opensthe nanotubes. Commercial fullerenes (Hoechst, SuperGold Grade C , purity 99 . a r X i v : . [ c ond - m a t . m e s - h a ll ] J a n natural (Sigma) and C enriched benzene and toluene(Eurisotop, France). We note that for toluene, only thebenzyl ring was enriched while the methyl group was ofnatural carbon. Co-encapsulation of the fullerenes andthe benzene or toluene proceeds by sonicating the SWC-NTs for 2 hours in the corresponding solvent: C solu-tion of 1 mg / ml. This method is known to result in aclathrate structure where the smaller benzene or toluenemolecules occupy half-half of the available inner volumein an alternating fashion [3, 9]. The resulting materialwas filtered to obtain nanotube bucky-papers and it wasrinsed with the corresponding non-enriched solvent (ben-zene or toluene) to remove any non-encapsulated excessfullerenes from the outside of the nanotubes. This stepwas followed by the final filtering and drying of the bucky-paper samples under a fume hood. The samples were an-nealed in dynamic vacuum at 1250 ◦ C for 1 hour. Thisprocess is known to yield high quality double-wall carbonnanotubes [3, 10]. In the following, we denote double-wall carbon nanotubes grown from C and toluene bothcontaining natural carbon as DWCNT. / DWCNTdenotes double wall carbon nanotubes where the innerwall is made of co-encapsulated C and C benzeneand ring-enriched toluene.Raman spectroscopy was performed with a Dilor xytriple monochromator spectrometer with an excitationlines of an Ar-Kr gas discharge laser. We report datawith the 514 . RAMAN SPECTROSCOPY RESULTS
We show the Raman spectra of reference SWCNT, DWCNT and / DWCNT samples around the 2DRaman line spectral range. The latter sample was basedon a mixture of ring-labelled C H -CH3 and naturalcarbon toluene with a ratio of 84:16. This means that thenominal C carbon content in the toluene was 72 %. TheDWCNT samples are characterized by the emergence ofthe lower frequency 2D Raman line which corresponds tothe inner tubes [11]. In fact, the inner tube 2D Ramanline is twice as strong as the present one for C basedinner tubes. This is explained by the fact that tolueneand benzene are relatively large compared to their nom-inal carbon content as compared to the fullerenes [3]. Itmeans that they use a significant amount of the availablevolume while contributing to less carbon atoms to theinner tube growth.The Raman spectra of DWCNT and / DWCNTshows a striking difference which is clearest from the bot-tommost comparison in Fig. 1 (indicated by a blue ar-row in the figure): a sizeable amount of spectral weight isshifted from around the peak of the inner tube 2D Ramanline toward lower Raman shifts which forms a ”tail”. This * SWCNT DWCNT
DWCNT DWCNT&
DWCNTRaman shift (cm -1 ) R a m a n s i gn a l ( a r b . u . ) FIG. 1. Raman spectra of the toluene+C based DWCNTsamples around the 2D Raman line. A corresponding Ramanspectrum on the starting SWCNT sample is shown for com-parison (dashed curve). The DWCNT and DWCNTdata are shown with an offset first in the middle. The twospectra are shown on one another at the bottom in order tohighlight the spectral weight which is downshifted (blue filledarea, indicated by an arrow). Asterisk indicates a small Ra-man line which is present in the pristine sample already. is a surprising observation as a homogeneous downshiftof the Raman line is expected from a naive considerationof isotope labeling, rather than the formation of a lowRaman frequency ”tail”. Previously, the growth of innertubes was investigated from C enriched fullerenes andtherein a uniform downshift was observed with a meandownshift corresponding to the formula [8, 12]: f f = (cid:114) .
011 + c · .
011 (1)where f and f are the Raman shifts with and without C doping, 12 .
011 g / mol is the molar mass of naturalcarbon and it reflects the 1 .
1% abundance of C in nat-ural carbon and c is the C concentration. We notethat previously a similar, anomalous development of alow Raman frequency tail was observed in isotope labeledbenzene based inner tubes in Ref. [13]. -1
33 % CI=27 % S ub t r ac t e d R a m a n s i gn a l ( a r b . u . ) DWCNT- DWCNT toluene based -1
67 % CI=5 % 2624 cm -1 I=18 % -1 I=11 %2590 cm -1
33 % CI=18 %2536 cm -1
85 % CI=7 % benzene based
Raman shift (cm -1 ) FIG. 2. The spectrum obtained after subtracting the 2D Ra-man line of the DWCNT sample from the / DWCNT.A deconvolution into several components is also shown. TheRaman shifts, the nominal C enrichment level and the in-tensity of the particular component with respect to the innertube 2D Raman line are shown. For comparison we show thesame kind of data for benzene based / DWCNT sampleafter Ref. [13].
We show in Fig. 2. the spectrum obtained after sub-tracting the 2D Raman line of the DWCNT samplefrom the / DWCNT. We also show for comparison thesame kind of data on C labeled benzene based DWC-NTs from Ref. [13]. The deconvolution of the subtractedspectrum into Lorentzian components reveal three com-ponents: a significantly downshifted, low intensity line,a moderately downshifted stronger signal, and a line atthe position of the unenriched inner tubes with a negativeintensity. The latter corresponds to the spectral weightwhich is missing in the subtracted spectrum, i.e. it is the spectral weight which is downshifted for the isotope la-beled sample in agreement with Fig. 1. The significantlydownshifted component corresponds to a C isotope en-richment of 67 %. This is smaller than the enrichmentfound for the same line in the benzene based DWCNTin Ref. [13] which may be due to the lower (72 %) Cisotope enrichment of carbon of toluene as compared to99 % in benzene. An interesting observation is that theother downshifted Raman line is found at the same posi-tion for both kinds of samples.It is tempting to associate the significantly downshiftedRaman line to a localized cluster or island of C isotopes,which assumption is studied further below. The analysisof the Raman intensity shows for both kinds of samplesthat no Raman spectral weight conservation applies.
THEORETICAL MODELLING ANDDISCUSSION
We calculated the first order Raman spectrum with thesemi-empirical PM3 method as implemented in the Gaus-sian09 package [14]. We neglected the outer tube and theinner tube was considered as a molecule: a hydrogen-terminated piece of (5,5) armchair type SWCNT consist-ing of 600 carbon and 20 hydrogen atoms. The struc-ture was first relaxed with the opt=tight option, thenwe obtained the force constants and the polarizabilityderivatives. We compared the Raman spectra of vari-ous small molecules (methane, benzene, C fullerene)calculated with the PM3 method to the first principlesbased (DFT/B3LYP) results. In general, the frequen-cies obtained by the semi-empirical method are not veryaccurate, which could be fixed by rescaling the force con-stants to fit the experimental (or first principles) value,but we do not bother the absolute position of the cal-culated Raman peaks, because we always used the sameforce constant matrix and only changed the masses inthe dynamical matrix according to the isotope distribu-tion. However, the Raman intensities calculated with thePM3 method reproduced the results of the more demand-ing DFT/B3LYP method very well. Using the freqchk utility of Gaussian09 [14] for different distribution of iso-tope masses the Raman intensities were then evaluatedfor 2000 random configurations. Two kinds of isotopedistribution were considered i) the homogeneous distri-bution where 60 carbon atoms were selected and replacedindividually, and ii) ring arrangement where 10 completerings of 6 carbon atoms were substituted by C isotopes– both resulting in a nominal 10 % isotope enrichment.In Figs. 3 and 4 the data of each vibrational mode ofevery configurations are presented in the G band regionfor the homogeneous and the clustered distribution, re-spectively. The colors (and sizes) of symbols correspondto the Raman intensity, the x -axis is the usual Ramanshift and the y -axis is the weight of the C substitutedcarbon atoms movements of the vibration mode. Theoverall left-top right-bottom trend confirms the naive ex-pectation, that the more dominant the motion of Catoms in a normal mode are, the more significant its red-shift is. The horizontal solid (black) line marks the po-sition of the isotope-shifted G peak according to Eq. 1with c = 10 %. In the homogeneous case the shift canbe well described with this simple formula, while in theclustered arrangement the shift is clearly larger and couldbe fitted with an effective enrichment of c = 16%. Thevertical solid (black) line indicates the weight of the Csubstituted carbon atoms (1 /
10) in the normal modes re-garding the nominal 10% enrichment. For homogeneousdistribution of the isotopes the center of the peaks liesvery precisely on this line – to no surprise. However, forthe ring configurations there is a convincing deviationfrom the simple 1 /
10 value. This is telling us, that if the substituted carbon atoms are clustered (e.g. in a ring),their presence can give more weight to the normal modeand also to the Raman intensity of the specific normalmode. Therefore they give more weight to the averagedRaman spectrum [13]. This effect might lie behind theanomalous development of a low Raman frequency. It isalso worth to note, that this is specific for the G mode, wedid not find a similar behavior for the Radial BreathingMode.Our theoretical finding supports the experimental ob-servation, that upon clusterization the downshift of theRaman peaks can be higher than the real substitutionratio. However, we did not find as significant downshiftsas measured. This might be due to the specific charac-ter of the 2D Raman line [15]. A similar analysis forthe 2D Raman line would be computationally extremelydemanding, since there is an additional integration overk-space for the 2D Raman line. FIG. 3. Scatterplot showing the calculated Raman intensitiesfor the homogeneous distribution of C isotopes. FIG. 4. Scatterplot showing the calculated Raman intensitiesfor the ring arrangement of C isotopes.
CONCLUSIONS
In conclusion, we presented the synthesis and Ramancharacterization of carbon nanotubes grown from C iso-tope labeled organic solvent (benzene and toluene) insidehost outer tubes. Raman spectroscopy analysis indicatesthat the C isotopes are non uniformly distributed onthe inner tube walls. This indicates that little or no car-bon diffusion takes places during the inner tube growth.The analysis is supported by semi-empirical calculationsof the vibrational modes for clustered C isotope richconfigurations. The material with C isotope rich clus-ters may find application as local nuclear spin labels orin quantum information storage.
ACKNOWLEDGEMENT
The Hungarian National Research, Development andInnovation Office (NKFIH) Grants Nr. K108676,K115608, and K119442 are acknowledged for support. [1] B. W. Smith, M. Monthioux, and D. E. Luzzi, Nature , 323–324 (1998).[2] B. W. Smith and D. Luzzi, Chem. Phys. Lett. , 169–174 (2000).[3] F. Simon and H. Kuzmany, Chem. Phys. Lett. , 85–88 (2006).[4] F. Simon, H. Kuzmany, J. Bernardi, F. Hauke, andA. Hirsch, Carbon , 1958–1962 (2006).[5] B. Botka, M. E. Fuestoes, H. M. Tohati, K. Nemeth,G. Klupp, Z. Szekrenyes, D. Kocsis, M. Utczas, E. Szekely, T. Vaczi, G. Tarczay, R. Hackl, T. W. Cham-berlain, A. N. Khlobystov, and K. Kamaras, SMALL (7), 1369–1378 (2014).[6] R. Pfeiffer, H. Kuzmany, C. Kramberger, C. Schaman,T. Pichler, H. Kataura, Y. Achiba, J. K¨urti, andV. Z´olyomi, Phys. Rev. Lett. , 225501–1–4 (2003).[7] R. Pfeiffer, F. Simon, H. Kuzmany, and V. N. Popov,Phys. Rev. B , 161404 –1–4 (2005).[8] V. Z´olyomi, F. Simon, A. Ruszny´ak, R. Pfeiffer, H. Peter-lik, H. Kuzmany, and J. K¨urti, Phys. Rev. B , 195419–1–8 (2007).[9] V. Z´olyomi, H. Peterlik, J. Bernardi, M. Bokor, I. L´aszl´o,J. Koltai, J. K¨urti, M. Knupfer, H. Kuzmany, T. Pich-ler, and F. Simon, J. Phys. Chem. C , 30260–30268(2014).[10] F. Simon, A. Kukovecz, C. Kramberger, R. Pfeiffer,F. Hasi, H. Kuzmany, and H. Kataura, Phys. Rev. B , 165439–1–5 (2005).[11] R. Pfeiffer, H. Kuzmany, F. Simon, S. N. Bokova, andE. Obraztsova, Phys. Rev. B , 155409 (2005).[12] F. Simon, C. Kramberger, R. Pfeiffer, H. Kuzmany,V. Z´olyomi, J. K¨urti, P. M. Singer, and H. Alloul, Phys.Rev. Lett. , 017401–1–4 (2005).[13] J. Koltai, G. Mezei, V. Z´olyomi, J. K¨urti, H. Kuzmany, T. Pichler, and F. Simon, The Journal of Physical Chem-istry C (51), 29520–29524 (2016).[14] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone,B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Cari-cato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino,G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toy-ota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima,Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Mont-gomery Jr., J. E. Peralta, F. Ogliaro, M. J. Bearpark,J. Heyd, E. N. Brothers, K. N. Kudin, V. N. Staroverov,R. Kobayashi, J. Normand, K. Raghavachari, A. P. Ren-dell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi,N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B.Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma,V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannen-berg, S. Dapprich, A. D. Daniels, d. Farkas, J. B. Fores-man, J. V. Ortiz, J. Cioslowski, and D. J. Fox, Gaus-sian 09, Rev B.01, 2009.[15] C. Thomsen and S. Reich, Phys. Rev. Lett.85