⇦Chinese Physics Letters, 2020, Vol. 37, No. 2, Article code 028101 Near-Field Optical Identification of Metallic and Semiconducting Single-Walled Carbon Nanotubes * Lele Wang (王乐乐)1,2, Bosai Lyu (吕博赛)1,2, Qiang Gao (高强)1,2, Jiajun Chen (陈佳俊)1,2, Zhe Ying (应哲)1,2, Aolin Deng (邓奥林)1,2, Zhiwen Shi (史志文)1,2** Affiliations 1Key Laboratory of Artificial Structures and Quantum Control (Ministry of Education), Shenyang National Laboratory for Materials Science, School of Physics and Astronomy, Shanghai Jiao Tong University, Shanghai 200240 2Collaborative Innovation Center of Advanced Microstructures, Nanjing 210093 Received 24 October 2019, online 18 January 2020 *Supported by the National Natural Science Foundation of China under Grant Nos. 11574204 and 11774224, and the National Key Research and Development Program of China (2016YFA0302001). Z.S. acknowledges the Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning and additional support from a Shanghai talent program.
**Corresponding author. Email: zwshi@sjtu.edu.cn Citation Text: Wang L L, Lv B S, Gao Q, Chen J J and Ying Z et al 2020 Chin. Phys. Lett. 37 028101    Abstract Single-walled carbon nanotubes (SWCNTs), due to their outstanding electrical and optical properties, are expected to have extensive applications, such as in transparent conductive films and ultra-small field-effect transistors (FETs). However, those applications can only be best realized with pure metallic or pure semiconducting SWCNTs. Hence, identifying and separating metallic from semiconducting SWCNTs in as-grown samples are crucial. In addition, knowledge of the type of an SWCNT is also important for further exploring its new properties in fundamental science. Here we report employing scanning near-field optical microscopy (SNOM) as a direct and simple method to identify metallic and semiconducting SWCNTs on SiO$_2$/Si substrates. Metallic and semiconducting SWCNTs show distinct near-field optical responses because the metallic tubes support plasmons whereas the semiconducting tubes do not. The reliability of this method is verified using FET testing and Rayleigh scattering spectroscopy. Our result demonstrates that the SNOM technique provides a reliable, simple, noninvasive and in situ method to distinguish between metallic and semiconducting SWCNTs. DOI:10.1088/0256-307X/37/2/028101 PACS:81.05.U-, 78.67.-n, 68.37.Uv © 2020 Chinese Physics Society Article Text Single-walled carbon nanotubes (SWCNTs) possess many outstanding electrical and optical properties, making this material recognized as one of the most promising replacements for silicon transistors in advanced electronic technologies. Recently, scientists from MIT created the largest computer chip made from carbon nanotubes, proving the feasibility of experimentally preparing chips with carbon nanotubes.[1] Whether SWCNTs display semiconducting or metallic depends on their chiral index.[2,3] Semiconducting SWCNTs can serve as the channel material of field-effect transistors (FETs) for integrated circuits (ICs),[4–8] while metallic SWCNTs have extensive applications in transparent conducting films, effective quantum wires,[9] nanometer-sized electrodes,[10–12] interconnect wires for integrated circuits,[13] etc. However, these applications can only be realized with pure metallic or pure semiconducting SWCNTs, whereas as-synthesized SWCNTs are always a mixture of them. In the past decade, great efforts have been made to obtaining pure semiconducting or pure metallic SWCNTs. On the one hand, a variety of physical and chemical ways have been employed to separate them from each other, such as selective chemical reactions.[14,15] On the other hand, some researchers have focused on methods to synthesize metallicity-specific SWCNTs, for example, using ceria as a catalyst support to obtain semiconducting SWCNTs.[16,17] To evaluate the validity of these separation and selective growth methods, reliable and accurate identification of metallic/semiconducting (M/S) nanotubes is necessary. Although some methods can be used to identify the metallicity of SWCNTs, they all exhibit limits or drawbacks to varying degrees. Electrical methods such as field-effect transistor (FET) testing[9] and electrostatic force microscopy (EFM)[18–20] typically require a complex process to fabricate microelectrodes on nanotubes. The scanning electron microscopy (SEM) method, developed in recent years, also requires prefabricated electrodes attached to nanotubes.[21] The Raman spectroscopy method is simpler, yet it presents large uncertainties due to broad and overlapping spectral peaks.[22–24] Here we show that scanning near-field optical microscopy (SNOM)[25–29] can be used to reliably and accurately identify metallic and semiconducting SWCNTs on SiO$_{2}$/Si substrates. The reported near-field optical method is based on the different near-field optical responses of metallic and semiconducting nanotubes, metallic nanotubes support plasmon modes that can be excited and observed through their interference, whereas semiconducting nanotubes do not support plasmons due to the lack of free electrons. Previously, Nemeth et al. reported using SNOM technique to distinguish metallic and semiconducting nanotubes based on their different near-field optical phases.[30,31] However, their method has large uncertainty because the phase contrast is rather small and the phase can even change continuously between metallic and semiconducting nanotubes. We identify metallic and semiconducting carbon nanotubes based on whether they support plasmon modes or not, which only yield two discrete results. The reliability of our method has been verified by using electrical tests and Rayleigh scattering spectroscopy. Our result demonstrates that the SNOM technique provides a reliable, simple, noninvasive and in situ method to distinguish between metallic and semiconducting SWCNTs.
cpl-37-2-028101-fig1.png
Fig. 1. Illustration of the near-field optical identification. (a) A schematic of scanning near-field optical microscopy (SNOM). Infrared (IR) light at 10.6 µm was focused onto a metal-coated AFM tip, and the tip-scattered infrared radiation light was measured by an MCT detector in the far field. (b) The AFM topography of several carbon nanotubes. (c) A near-field infrared microscopy image of the same area as in (b). The contrast between these carbon nanotubes is evident. In some nanotubes, a strong infrared response and periodic structure can be observed, indicating those tubes are metallic, while in others nearly no infrared response is apparent, revealing that the nanotubes are semiconducting.
The SWCNT samples used in this experiment were synthesized by chemical vapor deposition (CVD) on SiO$_{2}$/Si substrates.[16,32–38] The as-grown samples exhibited high quality as determined by scanning electron microscopy (SEM) images and their diameters were ranging from 1.3 to 2.0 nm as determined by the atomic force microscopy (AFM) characterization. Near-field optical identification was carried out on a homemade SNOM system (Fig. 1(a)), which is based on a tapping mode AFM. Infrared light at a wavelength of 10.6 µm is focused onto the apex of an AFM tip with a metal coating. The tip, with a small radius of $r \approx 10$ nm, carries a large near-field momentum that enables the excitation of plasmon wave in nanotubes with free electrons. The backscattered light from the tip apex is collected and measured by a HgCdTe detector in the far-field area. The tip-excited plasmon wave propagates along this one-dimensional tube and is reflected at the nanotube ends. The reflected plasmon wave interferes with the tip-stimulated plasmon wave, which modulates the intensity of the infrared radiation, generating a periodic oscillation. Scanning along a metallic carbon nanotube, one would observe a periodic interference pattern with a strong infrared response from the free electrons. Figure 1(b) displays an AFM topography image of nanotube networks containing both metallic and semiconducting nanotubes, which are indistinguishable from their topography. Figure 1(c) shows the corresponding near-field infrared image of the same area in which only some of the nanotubes appear as a bright color, while the others display a rather weak contrast. We attribute the bright features to metallic nanotubes and the hardly visible features to semiconducting SWCNTs. To confirm this hypothesis, we carried out a set of experiments, as discussed in the following in detail.
cpl-37-2-028101-fig2.png
Fig. 2. Near-field optical identification of carbon nanotubes. (a)–(d) AFM topography images of representative nanotube samples. (e)–(h) Corresponding near-field infrared images obtained with a 10.6 µm excitation. The nanotubes in (e) and (f) exhibit a prominent optical response and periodic oscillations. In contrast, the nanotubes in (g) and (h) show little near-field infrared response. (i) Statistics for the counts of identified carbon nanotubes. Among 41 samples, there are 27 semiconducting and 14 metallic nanotubes.
First, we carefully scanned several individual SWCNTs at high resolution. Figure 2 demonstrates the detailed infrared microscopy of several typical SWCNTs exhibiting two different infrared responses. AFM topography images of these representative nanotube samples are shown in Figs. 2(a)–2(d). Accordingly, Figs. 2(e)–2(f) provide the infrared SNOM images, in which the excitation wavelength is 10.6 µm, for the nanotubes shown in Figs. 2(a)–2(d), respectively. We found that the nanotubes exhibiting a bright color in the near-field images (Figs. 2(e) and 2(f)) always show a period structure near the end of the nanotubes, whereas the ones with weak contrast or a dark color do not show any periodic features (Figs. 2(g) and 2(h)). This result agrees well with the existence of plasmons in metallic SWCNTs that decay exponentially from the ends of the nanotubes. As reported previously, Luttinger-liquid plasmons in metallic SWCNTs could be directly probed using the SNOM system.[27] Luttinger-liquid plasmons, collective oscillations of conduction electrons, would only occur in metallic carbon nanotubes with a zero bandgap, while in semiconducting carbon nanotubes with a small bandgap, there are no conduction electrons. Therefore, a natural conjecture would be that the periodic oscillations correspond to the Luttinger-liquid plasmons in metallic SWCNTs. The intensity peaks and valleys in Figs. 2(e) and 2(f) correspond to the constructive and destructive interference patterns generated via the plasmons excited by the tip and the waves reflected by the endpoint of a tube. Note that although the wavelength of excitation is 10.6 µm, the surface plasmon wavelength is only approximately 100 nm, a hundred times smaller than the photon wavelength in free space.[27] This tightly confined surface plasmon could be useful for future nanophotonic devices and circuits. Our observation also indicates that the ratio between metallic and semiconducting nanotubes is approximately $1\!:\!2$, which is consistent with the theoretical prediction for naturally grown samples. Carbon nanotubes with different chiral indices $(m, n)$ possess different band structures and bandgaps.[2] When the value of $(m-n)$/3 is an integer, the SWCNT is metallic, otherwise the nanotube is semiconducting. Thus, the result implies that approximately one-third of the carbon nanotubes are metallic and the other two-thirds are semiconducting. For additional supporting data, we characterized many SWCNTs and analyzed them by SNOM, providing more evidence (see Supporting Materials, Section 3). The statistical results for the number of each different type of SWCNT are as expected. Among the tested 41 samples, there are 14 metallic SWCNTs and 27 semiconducting SWCNTs, close to the theoretical ratio of $1\!:\!2$ (Fig. 2(i)). It should be noted that the bundled SWCNTs with a diameter over 2 nm were not included in the analysis. The agreement between our observations and the theoretical prediction further confirms the validity of this identification method. It should be noted that this method is also applicable for identifying SWCNTs on other substrates. We have measured carbon nanotube samples on hBN substrates and observed the same phenomenon. To further verify the developed method, we compared the near-field optical identification with electrical transport measurements. Since the SWCNT samples were synthesized directly on a doping Si substrate with a dielectric layer of 285-nm-thick SiO$_{2}$, we were able to directly fabricate field-effect transistor devices from semiconducting and metallic carbon nanotubes. The SEM image of a representative device is shown in the inset of Fig. 3(a). We chose one device with a metallic nanotube and one device with a semiconducting nanotube and display their transport data in Figs. 3(a) and 3(d), respectively. In Fig. 3(a), for different gate voltages, the ratio of maximum to minimum source-drain current is less than 5, clearly revealing that this is a metallic nanotube. Next, we severed the metallic nanotube to achieve two terminals (Fig. 3(b)) for the near-field optical identification. The cutting was achieved by using an electrode-free local anodic oxidation (EFLAO) technique,[39] which enables in situ etching of low-dimensional materials with high precision and great flexibility. Figure 3(c) reveals the infrared near-field optical image of this carbon nanotube, in which the Luttinger-liquid plasmon oscillations are apparently observed at the nanotube endpoints. Meanwhile, the transport measurement of another device (Fig. 3(d)) features an ON/OFF ratio over 10$^{3}$, indicating that it is a semiconducting nanotube. The following near-field measurement result in Fig. 3(f) shows barely any near-field infrared response and the absence of plasmon signatures. The consistency between the presented approach and transport measurements confirms unambiguously the reliability of the near-field optical identification method.
cpl-37-2-028101-fig3.png
Fig. 3. Verification of the near-field optical identification method with transport measurements. (a) The $I_{\rm d}$–$V_{\rm g}$ curve of a carbon nanotube device. The small ON/OFF ratio reveals that this is a metallic tube. Inset: an SEM image of a representative FET device. (b) and (c) AFM and infrared SNOM images of the nanotube in (a) after being cut using local anodic oxidation. The periodic oscillations at the ends of the SWCNT show that this is a metallic tube, which is consistent with the FET result. (d) The $I_{\rm d}$–$V_{\rm g}$ curve of a semiconducting carbon nanotube, and the ON/OFF ratio is over 10$^{3}$. (e) and (f) AFM and infrared SNOM images of the nanotube in (d) after cutting. The weak infrared response indicates that this is a semiconducting tube, which corroborates the FET conclusion. Source-drain voltage, 0.1 V.
We further verified the developed method by comparing its results with those of Rayleigh scattering spectroscopy.[40–42] Rayleigh scattering spectroscopy can directly probe the optical transition resonance peaks of carbon nanotubes, which correspond to electronic transitions between the top of the conduction band and the bottom of the valence band. Carbon nanotubes of different chiral indices have different band structures and therefore different transition energies. By performing Rayleigh scattering spectroscopy, one can extract the chiral index of a nanotube. From the Rayleigh spectra of two representative nanotubes plotted in Figs. 4(a) and 4(d), we deduce that the chiral indices $(m, n)$ are (29, 8) and (15, 10), and therefore, they are metallic (Fig. 4(a)) and semiconducting (Fig. 4(d)), respectively. As before, we cut the nanotubes using the EFLAO technique and characterized them via SNOM. The first nanotube shows a strong near-field infrared signal with a periodic structure, while the second produces almost no near-field response. The near-field optical characterization clearly indicates that the first and second tubes are metallic and semiconducting, respectively. Therefore, we have confirmed the result through another independent method that the near-field optical identification is correct and reliable.
cpl-37-2-028101-fig4.png
Fig. 4. Rayleigh scattering verification of the developed method. (a) The Rayleigh scattering spectrum of an SWCNT. From the spectrum, we could extract that the chiral index of this carbon nanotube is (29, 8). Therefore, this is a metallic nanotube. (b) and (c) The AFM topography image (b) and near-field infrared image (c) of this nanotube. The infrared image in (c) also reveals the nanotube to be metallic. (d) The Rayleigh scattering spectrum of another SWCNT. From the spectrum, we could extract a chiral index of (15, 10). Therefore, this is a semiconducting nanotube. (e) and (f) The AFM topography image (e) and near-field infrared image (f) of this nanotube.
In conclusion, we have reported a near-field optical method that can identify metallic and semiconducting carbon nanotubes, providing a novel solution to the metallicity-distinguishing challenge. The reliability of this method has been verified by using transport measurements and Rayleigh scattering spectroscopy. The reported identification method is simple, noninvasive, in situ, and of high spatial resolution. In addition, the near-field optical identification results in no modification to the intrinsic structure and properties of the tested nanotubes. The developed method is applicable to a wide range of circumstances that require knowledge of the nanotube metallicity, such as evaluating the metallicity-specific separation or growth of SWCNTs and diagnosing subsequent SWCNT-based electronic circuits. Acknowledgements: We thank Feng Wang at UC Berkeley, Kaihui Liu and Yan Li at Peking University for helpful discussions. References Modern microprocessor built from complementary carbon nanotube transistorsRoom-temperature transistor based on a single carbon nanotubeLogic Circuits with Carbon Nanotube TransistorsAn Integrated Logic Circuit Assembled on a Single Carbon NanotubeHigh Performance n-Type Carbon Nanotube Field-Effect Transistors with Chemically Doped ContactsImpact of Temperature Variation on Performance of Carbon Nanotube Field-Effect Transistor—Based on Chaotic Oscillator: A Quantum Simulation StudyIndividual single-wall carbon nanotubes as quantum wiresCarbon nanotubes as long ballistic conductorsQuantum Interference and Ballistic Transmission in Nanotube Electron WaveguidesMagnetic Transport Properties of Fe-Phthalocyanine Dimer with Carbon Nanotube ElectrodesA 1 GHz Integrated Circuit with Carbon Nanotube Interconnects and Silicon TransistorsA Route for Bulk Separation of Semiconducting from Metallic Single-Wall Carbon NanotubesBulk Separative Enrichment in Metallic or Semiconducting Single-Walled Carbon NanotubesChirality-specific growth of single-walled carbon nanotubes on solid alloy catalystsGrowth of Semiconducting Single-Walled Carbon Nanotubes by Using Ceria as Catalyst SupportsScanned Probe Microscopy of Electronic Transport in Carbon NanotubesA Scanning Probe Microscopy Based Assay for Single-Walled Carbon Nanotube MetallicityLocal Electronic Structure of Single-Walled Carbon Nanotubes from Electrostatic Force MicroscopyDirect Identification of Metallic and Semiconducting Single-Walled Carbon Nanotubes in Scanning Electron MicroscopyRaman spectroscopy of carbon nanotubesStructural ( n , m ) Determination of Isolated Single-Wall Carbon Nanotubes by Resonant Raman ScatteringSystematic Comparison of the Raman Spectra of Metallic and Semiconducting SWNTsUltrabroadband infrared nanospectroscopic imagingPhase-Resolved Surface Plasmon Interferometry of GrapheneObservation of a Luttinger-liquid plasmon in metallic single-walled carbon nanotubesCorrelation of Electron Tunneling and Plasmon Propagation in a Luttinger LiquidGate-tuning of graphene plasmons revealed by infrared nano-imagingScattering near-field optical microscopy on metallic and semiconducting carbon nanotube bundles in the infraredNanoscale Characterization of Individual Horizontally Aligned Single-Walled Carbon NanotubesGrowth of Millimeter-Long and Horizontally Aligned Single-Walled Carbon Nanotubes on Flat SubstratesGrowth Mechanism of Oriented Long Single Walled Carbon Nanotubes Using “Fast-Heating” Chemical Vapor Deposition ProcessQuasi-Continuous Growth of Ultralong Carbon Nanotube ArraysUltra-high-yield growth of vertical single-walled carbon nanotubes: Hidden roles of hydrogen and oxygenUltralow Feeding Gas Flow Guiding Growth of Large-Scale Horizontally Aligned Single-Walled Carbon Nanotube ArraysThe reason for the low density of horizontally aligned ultralong carbon nanotube arraysChirality-Controlled Synthesis and Applications of Single-Wall Carbon NanotubesElectrode-Free Anodic Oxidation Nanolithography of Low-Dimensional MaterialsAn atlas of carbon nanotube optical transitionsHigh-throughput optical imaging and spectroscopy of individual carbon nanotubes in devicesProbing Electronic Transitions in Individual Carbon Nanotubes by Rayleigh Scattering
[1] Hills G et al 2019 Nature 572 595
[2]Dresselhaus G and Riichiro S 1998 Physical Properties of Carbon Nanotubes (London: Imperial College Press)
[3]Dresselhaus M S et al 1996 Science of Fullerenes and Carbon Nanotubes: Their Properties and Applications (Amsterdam: Elsevier)
[4] Tans S J et al 1998 Nature 393 49
[5] Bachtold A et al 2001 Science 294 1317
[6] Chen Z et al 2006 Science 311 1735
[7] Javey A et al 2005 Nano Lett. 5 345
[8] Ha N V and Song H J 2015 Chin. Phys. Lett. 32 038201
[9] Tans S J et al 1997 Nature 386 474
[10] White C T and Todorov T N 1998 Nature 393 240
[11] Kong J et al 2001 Phys. Rev. Lett. 87 106801
[12] Lv Y Z et al 2017 Chin. Phys. Lett. 34 047302
[13] Close G F et al 2008 Nano Lett. 8 706
[14] Debjit C et al 2003 J. Am. Chem. Soc. 125 3370
[15] Chen Z, Du X, Du M H, Rancken C D, Cheng H P and Rinzler A G 2003 Nano Lett. 3 1245
[16] Yang F, Wang X, Zhang D, Yang J, Luo D, Xu Z, Wei J, Wang J Q , Xu Z and Peng F 2014 Nature 510 522
[17] Qin X, Peng F, Yang F, He X, Huang H, Luo D, Yang J, Wang S, Liu H and Peng L 2014 Nano Lett. 14 512
[18] Bachtold A, Fuhrer M, Plyasunov S, Forero M, Anderson E H, Zettl A and McEuen P L 2000 Phys. Rev. Lett. 84 6082
[19] Lu W, Xiong Y, Hassanien A, Zhao W, Zheng M and Chen L 2009 Nano Lett. 9 1668
[20] Heo J and Bockrath M 2005 Nano Lett. 5 853
[21] Li J, He Y, Han Y, Liu K, Wang J, Li Q, Fan S and Jiang K 2012 Nano Lett. 12 4095
[22] Dresselhaus M S, Dresselhaus G, Saito R and Jorio A 2005 Phys. Rep. 409 47
[23] Jorio A, Saito R, Hafner J, Lieber C, Hunter D, McClure T, Dresselhaus G and Dresselhaus M 2001 Phys. Rev. Lett. 86 1118
[24] Gao B, Zhang Y, Zhang J, Kong J and Liu Z 2008 J. Phys. Chem. C 112 8319
[25] Bechtel H A, Muller E A, Olmon R L, Martin M C and Raschke M B 2014 Proc. Natl. Acad. Sci. USA 111 7191
[26] Gerber J A, Berweger S, O'Callahan B T and Raschke M B 2014 Phys. Rev. Lett. 113 055502
[27] Shi Z, Hong X, Bechtel H A, Zeng B, Martin M C, Watanabe K, Taniguchi T, Shen Y R and Wang F 2015 Nat. Photon. 9 515
[28] Zhao S, Sheng W, Fanqi W, Wu S, Bakti U I, Tairu L, Lili J, Yudan S, Siqi W and Kenji W 2018 Phys. Rev. Lett. 121 047702
[29] Fei Z, Rodin A S, Andreev G O, Bao W, McLeod A S, Wagner M, Zhang L M, Zhao Z, Thiemens M and Dominguez G 2012 Nature 487 82
[30] Németh G, Datz D, Tóháti H M, Pekker Á and Kamarás K 2016 Phys. Status Solidi 253 2413
[31] Németh G, Datz D, Tóháti H M, Pekker á Otsuka K, Inoue T, Maruyama S and Kamarás K 2017 Phys. Status Solidi B 254 1700433
[32] Huang S, Cai X and Liu J 2003 J. Am. Chem. Soc. 125 5636
[33] Huang S, Woodson M, Smalley R and Liu J 2004 Nano Lett. 4 1025
[34] Hong B H, Lee J Y, Beetz T, Zhu Y, Kim P and Kim K S 2005 J. Am. Chem. Soc. 127 15336
[35] Zhang G, Mann D, Zhang L, Javey A, Li Y, Yenilmez E, Wang Q, McVittie J P, Nishi Y and Gibbons J 2005 Proc. Natl. Acad. Sci. USA 102 16141
[36] Jin Z, Chu H, Wang J, Hong J, Tan W and Li Y 2007 Nano Lett. 7 2073
[37] Zhang R, Xie H, Zhang Y, Zhang Q, Jin Y, Li P, Qian W and Wei F 2013 Carbon 52 232
[38] Liu B, Wu F, Gui H, Zheng M and Zhou C 2017 ACS Nano 11 31
[39] Li H Y, Ying Z, Lyu B S, Deng A L, Wang L L, Taniguchi T, Watanabe K and Shi Z W 2018 Nano Lett. 18 8011
[40] Liu K, Deslippe J, Xiao F, Capaz R B, Hong X, Aloni S, Zettl A, Wang W, Bai X and Louie S G 2012 Nat. Nanotechnol. 7 325
[41] Liu K, Hong X, Zhou Q, Jin C, Li J, Zhou W, Liu J, Wang E, Zettl A and Wang F 2013 Nat. Nanotechnol. 8 917
[42] Sfeir M Y, Feng W, Limin H, Chia-Chin C, Hone J, O'Brien S P, Heinz T F and Brus L E 2004 Science 306 1540