Raman and XPS analyses of pristine and annealed N-doped double-walled carbon nanotubes
Lei Shi, Markus Sauer, Oleg Domanov, Philip Rohringer, Paola Ayala, Thomas Pichler
pphysica status solidi, 8 September 2015
Raman and XPS analyses ofpristine and annealed N-dopeddouble-walled carbon nanotubes
Lei Shi *,1 , Markus Sauer , Oleg Domanov , Philip Rohringer , Paola Ayala , Thomas Pichler University of Vienna, Faculty of Physics, Strudlhofgasse 4, 1090 Vienna, AustriaReceived XXXX, revised XXXX, accepted XXXXPublished online XXXX
Key words:
Nitrogen dopant, double-walled carbon nanotubes, annealing, Raman scattering, X-ray photoelectron spectroscopy. ∗ Corresponding author: e-mail [email protected] , Phone: +43-0676-4456949, Fax:+43-1427751375
N-doped single/multi-walled carbon nanotubes (CNTs) were studied for long time from synthesis to properties.However, the stability of N in the CNT lattice still needs further developments. In this work, to obtain more stableN-doped CNTs, concentric double-walled (DW) CNTs with more N were synthesized using benzylamine as C and Nsource. In order to test the stability of N-doped DWCNTs, high-temperature annealing in vacuum was performed. ByXPS and Raman spectroscopic measurements, we found that the N-doped DWCNTs are still stable under 1500 ◦ C :the graphitic N does not change at all, the molecular N is partly removed, and the pyridinic N ratio greatly increases bymore than two times. The reason could be that the N atoms from the surrounded N-contained materials combine intothe CNT lattice during the annealing. Compared with the undoped DWCNTs, no Raman frequency shift was observedfor the RBM, the G-band, and the G’-band of the N-doped DWCNTs. Copyright line will be provided by the publisher
In order to tailor the band structure,the Fermi level, and the density of states of the carbonnanotubes (CNTs), nitrogen as electron donor was used todope the pristine CNTs [1,2]. The N-doping changes theelectronic properties of the CNTs, including the densityof mobile charge carriers, hopping mobility, and workfunction[3]. Hence, compared with the pristine CNTs,the so–called N-doped CNTs are with better N–typeconductivity [3], ideal gas sensor [4], enhanced fieldemission properties [5], and high performance transistor[6]. Although the advantage of the N-doped CNTs is veryclear, the effective and controllable synthesis of N-dopedCNTs with specific type of doping is still not solved.Synthesis of N-doped CNTs was studied for longtime, but most of the researches focus on N-dopedmulti-walled CNTs (MWCNTs), partly on single-wallCNTs (SWCNTs) [7,8,9,10], and few on double-walledCNTs (DWCNTs) [5,11,12] because of the challengedifferences for the synthesis. DWCNTs have alreadyshow their advantage on both properties of SWCNTsand MWCNTs, for example, combined the excellent mechanical, electrical, and chemical properties [13].Hence, DWCNTs were attracted more attention recentlyby physicists, chemists, materials scientists, and engineers.Even more, the N-doped DWCNTs would be even betterfor the future fundamental studies as well as applicationsdue to their improved electrical and chemical propertiescompared to pristine DWCNTs. Hence, synthesis ofhigh-purity N-doped DWCNTs is one of the key pointsfor future researches and applications.In-situ doping of nitrogen atoms in the structure ofthe CNTs during the growth gives much better nitrogenstabilities than the post-treatment method, for example,under thermochemical treatment [14] or by hydrogennitrate [15]. The nitrogen source in the in-situ synthesisnormally can be both ammonia gas [3,5,11,12] or liquidcontained nitrogen (N-liquid) including ethylenediamine[10], acetonitrile [8,9], pyridine [8,12], and benzylamine[7,16,17,18]. The big advantage using N-liquid is thatit not only can be used as N supplier, but also as Csource. Till now, high-purity N-doped DWCNTs can beonly synthesized using ammonia gas, but not N-liquid.
Copyright line will be provided by the publisher a r X i v : . [ phy s i c s . a t m - c l u s ] S e p L. Shi et al.: Raman and XPS analyses of N-doped DWCNTs
Here, we report a detailed study on the N-dopedDWCNTs with small diameter by HVCVD using purebenzylamine as C source as well as N supplier. Higherdoping ratio, smaller diameter, and high-purity N-dopedDWCNT buckypaper make them good candidate forapplications, for example, in the fields of flexible transparentconductive film [15] and oxygen reduction reactions [19].In addition, we also testing the stability of the N-dopedDWCNTs under high temperature. More interestingly, wefound out that after annealing the doping ratio increasesdramatically compared with the pristine N-doped DWCNTs.Under the high temperature, the graphitic N is very stable,the ratio for the pyridinic N increases a lot due to theconversion from the N-contained materials, and the ratiofor the molecular N decreases, but it is not completelyremoved.
N-doped DWCNTs weresynthesized by high vacuum chemical vapor deposition(HVCVD). The HVCVD system was successfully used forthe growth of pristine DWCNTs with small diameter usingethanol as the C source in our group [20]. Here, in orderto synthesize N-doped thin DWCNTs, pure benzylamine(Sigma-Aldrich, 99.9 wt. % ) was used as C source as wellas N supplier. Ammonium iron citrate (Sigma-Aldrich, 3wt. % ) was mixed with MgO (Sigma-Aldrich, 97 wt. % )in ethanol, sonicated in bath for 1 hour, and dried inbeaker at 70 ◦ C . The growth was carried out at differenttemperatures between 825 and 900 ◦ C with a pressure at ∼ ∼
37 wt. % ) for 2 hoursto dissolve and remove most of the MgO and part of ironparticles. Secondly, the sample was heated in dry air forhalf hour at 400 ◦ C to remove the amorphous carbon andthe graphitic carbon surrounded the iron particles, andthen immersed in HCl again for 24 hours to remove therest of the MgO and exposed iron particles. Thirdly, thedried powder was heated in air at 500 ◦ C for 2 hours toremove the thin or destroyed SWCNTs. Finally, a thinbuckypaper was formed after sonicating the obtainedDWCNT powder in ethanol and filtered by membrane(MF-Millipore, 0.22 µ m). Typically, 0.5 g catalyst wasused for synthesis, and 30 mg N-doped DWCNTs can beobtained after purification.The purified N-doped DWCNTs were annealed invacuum ( ∼ − mbar) at 1500 ◦ C to test the nitrogenstability and to improve the graphitization of the N-dopedDWCNTs.The samples were measured by a Raman spectrometer(Horiba Jobin Yvon, LabRAM HR800) in ambient conditions Figure 1 µ m. Standard-resolution mode was used for themeasurement by the help of a x50 objective, so the spectralresolution was ∼ − . For ease of comparison, thespectra were all normalized by the intensity of the G-band.The nitrogen contents of the N-doped DWCNT sampleswere firstly annealed in vacuum (10 − mbar), and thenprobed with monochromatic Al K α radiation (1486.6eV) by a hemispherical SCIENTA RS4000 photoelectronanalyzer. We firstly analyze theRaman spectra of N-doped DWCNTs before and afterpurification. As seen in Figure 1a, the G/D intensityratio slightly increases after purification due to theremoval of the catalysts, which also indicates that thepurification process does not damage the structure ofthe N-doped CNTs. Some N-doped SWCNTs werealso synthesized simultaneously together with N-doped
Copyright line will be provided by the publisher ss header will be provided by the publisher 3
Figure 2 % of N-doped CNTs in the purified sampleare N-doped DWCNTs, which is very similar as the value Figure 3 ◦ C excitedby 568.188 nm laser. Inset: A close view for the D-band.for the undoped DWCNTs purified by a similar purificationprocess [20,21,22]. Compared to the radial breathingmode (RBM) peaks of the as-grown N-doped DWCNTs,some of the RBM peaks of the purified sample disappearafter purification as indicated by the arrows in Fig. 1b,suggesting that these RBM peaks belong to the SWCNTs.From the Kataura plot, we found that the disappearedRBM peaks corresponding to the metallic SWCNTs [23].In the G-band region, the Breit-Wigner-Fano lineshapefor the metallic SWCNTs is normally very wide andintense [24,25]. So the diminished G − after purificationwas observed (Fig. 1a) because of the population of themetallic SWCNTs decreases.We didn’t observe any change for the RBM peaksdue to the doping, as shown in Figure 2a. From the wellestablished inverse proportionality of the radial breathingmode (RBM) to the tube diameter [26], we calculated theaverage diameter of the purified N-doped DWCNTs isaround 0.9 and 1.5 nm for the inner and the outer tubes,respectively. This value of the diameter is the same asthe undoped DWCNTs [20], suggesting that the dopingdoes not change the diameter distribution of DWCNTssynthesized using the same catalysts. This is differentfrom that the thinner N-doped SWCNTs were usuallysynthesized by introducing the N-dopant [7,27].The G-band can be fitted by four peaks correspondingto the G − and G + of the inner and outer tubes. No Ramanshifts of the G + for both the inner and outer tubes wereobserved between the doped and undoped DWCNTs (Fig.2b), because the low-ratio dopants can not change thefrequency of the G + mode [18]. In addition, the G − forthe doped and undoped tubes are at the same frequency,indicating that the diameter distribution of the doped andthe undoped tubes is very similar, because the G − is highlysensitive to the diameter of the tubes [28]. Copyright line will be provided by the publisher
L. Shi et al.: Raman and XPS analyses of N-doped DWCNTs
Similar as the G-band behaviour, the G’-band shiftwas also not observed due to the N-doping. Typically, theG’-band of SWCNTs can be fitted to one Lorentzian [29].For the DWCNT system, the G’-band can be fitted by twoLorentzians corresponding to the inner and the outer tubes.The overall G’-bands of the undoped and doped samplesare at almost the same frequency. However, both fittingpeaks for the N-doped DWCNTs were maybe shifted tolower frequencies a little bit (Fig. 2c), which confirms thatthe doping is N-type [18,30]. Also, the frequency shiftis a little bit different for the inner and the outer tubes,suggesting that the doping ratio for the inner and the outertubes depends on their diameters. This relates with thediameter change of N-doped SWCNTs when doped withdifferent ratio of nitrogen in previous studies [7,27].In order to test the stability of the N-doped DWCNTs,the sample was annealed at 1500 ◦ C in vacuum. Asshown in Figure 3, compared with the pristine N-dopedDWCNTs, the RBM peaks for the annealed ones keepat the same frequencies, which means that the N-dopedDWCNTs are very stable, and not destroyed or meltedinto larger tubes. Theoretically the RBM of high-dopingSWCNTs should shift to lower frequency by a fewwavenumbers [1]. However, in our case no frequencyshift was observed, indicating that the doping ratio of theN-doped DWCNTs is not high (further discussion in nextsection). Furthermore, the G/D intensity ratio increasesa little bit after annealing, confirming the stability ofN-doped DWCNTs again, and also revealing a bettercrystallinity by annealing. With the same reason, theintensity of RBM peaks of annealed sample is higher thanthat of the pristine sample as seen in Fig. 3. Table 1
Different types of N in pristine and annealedN-doped DWCNTs. The absolute value (relative ratio) fordifferent types of N are shown in at. % ( % ). Sample Pyridinic quaternary N-related gas totalPristine 0.18(27.5) 0.36(55.3) 0.11(17.2) 0.65(100)Annealed 0.66(60.2) 0.36(32.5) 0.08(7.3) 1.10(100)
Interestingly, the high-temperature annealing not onlyincreases the CNT crystallinity, but also changes the Nratio from 0.65 to 1.10 at. % . As shown in Figure 4b and 4c,there are three types of nitrogen in the N-doped DWCNTsfor both of the pristine and the annealed samples: pyridinic,graphitic (or called quaternary, or sp ), and molecular. Theresults are summarized in Table 1. Firstly, after annealingthe molecular nitrogen ratio decreases due to the removaleffect by high temperature, but they are not completelyremoved. It is also confirmed by the largely decreased Oratio after annealing (Fig. 4a) [2,16,17]. Secondly, thegraphitic nitrogen ratio does not change at all, indicatingthat the very good stability of graphitic nitrogen, which cansurvive under the extremely high temperature in vacuumwithout problem. Thirdly, the relative content of pyridinicnitrogen is greatly enhanced by annealing from 0.18 to Figure 4 % . The nitrogen could be from the molecular Copyright line will be provided by the publisher ss header will be provided by the publisher 5 nitrogen, because it is possible to convert the molecularnitrogen to the pyridinic nitrogen during the annealingby deoxygenation process [31]. However, apparently themolecular nitrogen is not enough for this large change.Further possibilities will be discussed in next section.Another question is how does the N-dopants affectthe G’-band excited by multi-lasers. Previous studiesshowed that the G’-band of N-doped CNTs almost doesnot shift after doping. Also, the frequency of G’-bandhighly depends on the excitation laser energy [29], and thediameter of the CNTs as well [32]. We exactly observedthese behaviours as shown in Figure 5a. In addition, withall of these knowledge, we would like to analyze theannealing effect on the G’-band of the N-doped DWCNTs.As shown in Fig. 5a, the left and right peaks in the G’-bandwere assigned to the G’-bands for the inner tubes andthe outer tubes, respectively [32]. A linear fitting wasnormally used in the studies of the D and G’ dispersions[29]. The linear relation can be wrote as ω G (cid:48) = AE laser + B, where A and B are fitting parameters, and E laser is the excitation laser energy. A higher doping ratioshould give out a lower slope [30]. As seen in Fig. 5band 5c, we obtained very well linear fittings for all thesamples, although the slopes are slightly different due tothe different doping ratios. Before the annealing, the slopesfor the inner and outer tubes ( ω G (cid:48) − inner = (94 ± laser + (2424 ± ω G (cid:48) − outer = (96 ± laser + (2448 ± ω G (cid:48) − pristine = (97 ± laser + (2442 ± ω G (cid:48) − doped = (71 ± laser + (2465 ± ω G (cid:48) − inner = (86 ± laser +(2446 ± ω G (cid:48) − outer = (73 ± laser + (2510 ± Figure 5
High-purity N-doped DWCNTs werefirstly synthesized from liquid pure benzylamine. Thefollowing purification process does not destroy the N-dopedDWCNT structure, and the high-temperature annealing in
Copyright line will be provided by the publisher
L. Shi et al.: Raman and XPS analyses of N-doped DWCNTs vacuum can increase their crystallinity. The N ratio forthe purified N-doped DWCNTs is 0.65 at. % , which ishigher than that in previous studies [16,17]. This dopingratio still can not shift the frequencies of the RBM, theG-band and the G’-band. Three types of N exist in theN-doped DWCNTs: pyridinic, graphitic, and molecular.After annealing, the ratio for the molecular N decreases,no change for the graphitic N, and increasing a lot for thepyridinic N. This great increasing is mostly contributedfrom the higher-doped outer tubes due to the combinationof the surrounded N-contained materials into the structureof the DWCNTs. Acknowledgements
This work was supported by theAustrian Science Funds (FWF, NanoBlends I 943–N19). L.S. thanks the scholarship supported by the China ScholarshipCouncil (CSC). P. A. was supported by a Marie Curie IntraEuropean Fellowship within the 7th European CommunityFramework Program.
References [1] I. C. Gerber, P. Puech, A. Gannouni, and W. Bacsa,Physical Review B (7), 075423 (2009).[2] P. Ayala, R. Arenal, A. Loiseau, A. Rubio, and T. Pichler,Reviews of Modern Physics (2), 1843 (2010).[3] J. Wiggins-Camacho and K. Stevenson, The Journal ofPhysical Chemistry C (44), 19082–19090 (2009).[4] F. Villalpando-Paez, A. Romero, E. Munoz-Sandoval,L. Martinez, H. Terrones, and M. Terrones, ChemicalPhysics Letters (1), 137–143 (2004).[5] K. Y. Chun, H. S. Lee, and C. J. Lee, Carbon (1),169–177 (2009).[6] K. Xiao, Y. Liu, P. Hu, G. Yu, Y. Sun, and D. Zhu, Journalof the American Chemical Society (24), 8614–8617(2005).[7] F. Villalpando-Paez, A. Zamudio, A. Elias, H. Son,E. Barros, S. Chou, Y. Kim, H. Muramatsu, T. Hayashi,J. Kong, H. Terrones, G. Dresselhaus, M. Endo,M. Terrones, and M. S. Dresselhaus, Chemical PhysicsLetters (4), 345–352 (2006).[8] C. L. Pint, Z. Sun, S. Moghazy, Y. Q. Xu, J. M. Tour, andR. H. Hauge, ACS Nano (9), 6925–6934 (2011).[9] T. Thurakitseree, C. Kramberger, P. Zhao, S. Aikawa,S. Harish, S. Chiashi, E. Einarsson, and S. Maruyama,Carbon (7), 2635–2640 (2012).[10] J. C. Li, P. X. Hou, L. Zhang, C. Liu, and H. M. Cheng,Nanoscale (20), 12065–12070 (2014).[11] S. Y. Kim, J. Lee, C. W. Na, J. Park, K. Seo, and B. Kim,Chemical Physics Letters (4), 300–305 (2005).[12] L. Panchakarla, A. Govindaraj, and C. Rao, ACS Nano (5), 494–500 (2007).[13] C. Shen, A. H. Brozena, and Y. Wang, Nanoscale (2),503–518 (2011).[14] D. Golberg, Y. Bando, L. Bourgeois, K. Kurashima, andT. Sato, Carbon (14), 2017–2027 (2000).[15] H. Z. Geng, K. K. Kim, K. P. So, Y. S. Lee, Y. Chang,and Y. H. Lee, Journal of the American Chemical Society (25), 7758–7759 (2007). [16] P. Ayala, A. Gr¨uneis, T. Gemming, D. Grimm,C. Kramberger, M. H. R¨ummeli, F. L. Freire, H. Kuzmany,R. Pfeiffer, A. Barreiro, B. B¨uchner, and T. Pichler, TheJournal of Physical Chemistry C (7), 2879–2884(2007).[17] P. Ayala, F. L. Freire, M. H. R¨ummeli, A. Gr¨uneis, andT. Pichler, Physica Status Solidi (b) (11), 4051–4055(2007).[18] I. Maciel, J. Campos-Delgado, M. Pimenta, M. Terrones,H. Terrones, A. Rao, and A. Jorio, Physica Status Solidi(b) (11-12), 2432–2435 (2009).[19] K. Gong, F. Du, Z. Xia, M. Durstock, and L. Dai, Science (5915), 760–764 (2009).[20] L. Shi, P. Rohringer, K. Suenaga, Y. Niimi, J. Kotakoski,J. C. Meyer, H. Peterlik, P. Ayala, and T. Pichler, arXivpreprint p. arXiv:1507.04896 (2015).[21] P. Rohringer, L. Shi, X. Liu, K. Yanagi, and T. Pichler,Carbon , 282–290 (2014).[22] L. Shi, L. Sheng, L. Yu, K. An, Y. Ando, and X. Zhao,Nano Research (8), 759–766 (2011).[23] H. Kataura, Y. Kumazawa, Y. Maniwa, I. Umezu,S. Suzuki, Y. Ohtsuka, and Y. Achiba, Synthetic Metals (1), 2555–2558 (1999).[24] M. Pimenta, A. Marucci, S. Empedocles, M. Bawendi,E. Hanlon, A. Rao, P. Eklund, R. Smalley, G. Dresselhaus,and M. S. Dresselhaus, Physical Review B (24), R16016(1998).[25] S. D. M. Brown, A. Jorio, P. Corio, M. S. Dresselhaus,G. Dresselhaus, R. Saito, and K. Kneipp, Physical ReviewB (15), 155414 (2001).[26] H. Kuzmany, W. Plank, M. Hulman, C. Kramberger,A. Gr¨uneis, T. Pichler, H. Peterlik, H. Kataura, andY. Achiba, The European Physical Journal B-CondensedMatter and Complex Systems (3), 307–320 (2001).[27] H. R. Barzegar, E. Gracia-Espino, T. Sharifi, F. Nitze, andT. W˚agberg, The Journal of Physical Chemistry C (48),25805–25816 (2013).[28] A. Jorio, A. Souza Filho, G. Dresselhaus, M. S.Dresselhaus, A. Swan, M. ¨Unl¨u, B. Goldberg, M. Pimenta,J. Hafner, C. Lieber, and R. Saito, Physical Review B (15), 155412 (2002).[29] M. S. Dresselhaus, G. Dresselhaus, R. Saito, and A. Jorio,Physics Reports (2), 47–99 (2005).[30] I. O. Maciel, N. Anderson, M. A. Pimenta, A. Hartschuh,H. Qian, M. Terrones, H. Terrones, J. Campos-Delgado,A. M. Rao, L. Novotny, and A. Jorio, Nature Materials (11), 878–883 (2008).[31] J. Pels, F. Kapteijn, J. Moulijn, Q. Zhu, and K. Thomas,Carbon (11), 1641–1653 (1995).[32] A. Souza Filho, A. Jorio, G. G. Samsonidze,G. Dresselhaus, M. Pimenta, M. S. Dresselhaus, A. K.Swan, M. ¨Unl¨u, B. Goldberg, and R. Saito, PhysicalReview B (3), 035427 (2003).[33] R. Podila, J. C. Chac´on-Torres, J. Spear, T. Pichler,P. Ayala, and A. Rao, Applied Physics Letters (12),123108 (2012).(12),123108 (2012).