Band Engineering of Carbon Nanotubes for Device Applications

Abstract

Since the discovery of carbon nanotubes (CNTs) in 1991, research has been performed on their synthesis and applications. Among all the applications, electrical devices are the most promising because of their unique band structures. Nearly all breakthroughs related to devices are accompanied by improvements in the synthesis and processing of CNTs. Therefore, for further applications, the development of CNT-based material processing and device applications should be specially focused upon. With regard to electrical devices, the processing of materials is mainly referred to as band engineering of CNTs, which includes the selective enrichment of CNTs with specific electrical properties, the modulation of energy bands, and the construction of heterostructures. Devices based on CNTs and their aggregates are steadily advancing, and future efforts should be focused on material synthesis and device manufacture together with proper band engineering. Carbon nanotubes (CNTs)—especially single-walled CNTs—are promising for device applications. Although CNTs have excellent intrinsic properties, their diverse band structures bring difficulties to improving the performances of CNT-based devices. Therefore, band engineering is necessary. For diverse electrical properties, selective enrichment of CNTs with specific electrical properties is essential for determining their corresponding application fields. For a certain band structure, methods such as doping can be used to slightly tune the energy bands of CNTs and make them more suitable for specific devices. Additionally, for some intrinsic limitations, construction of heterostructures with other functional materials is an effective way to tune the carrier transport at the interface and broaden the application range of CNTs. In this review, we discuss in detail the band engineering of CNTs and corresponding device applications from the respect of both microscopic and macroscopic devices. We present an outlook that controlled synthesis will determine the future applications and proper manufacture will improve the application qualities. Carbon nanotubes (CNTs)—especially single-walled CNTs—are promising for device applications. Although CNTs have excellent intrinsic properties, their diverse band structures bring difficulties to improving the performances of CNT-based devices. Therefore, band engineering is necessary. For diverse electrical properties, selective enrichment of CNTs with specific electrical properties is essential for determining their corresponding application fields. For a certain band structure, methods such as doping can be used to slightly tune the energy bands of CNTs and make them more suitable for specific devices. Additionally, for some intrinsic limitations, construction of heterostructures with other functional materials is an effective way to tune the carrier transport at the interface and broaden the application range of CNTs. In this review, we discuss in detail the band engineering of CNTs and corresponding device applications from the respect of both microscopic and macroscopic devices. We present an outlook that controlled synthesis will determine the future applications and proper manufacture will improve the application qualities. Carbon nanotubes (CNTs) have long been considered as promising materials for next-generation electronic systems because of their excellent electrical, optical, and mechanical properties.1Franklin A.D. Nanomaterials in transistors: from high-performance to thin-film applications.Science. 2015; 349: aab2750Crossref PubMed Scopus (237) Google Scholar,2De Volder M.F.L. Tawfick S.H. Baughman R.H. Hart A.J. Carbon nanotubes: present and future commercial applications.Science. 2013; 339: 535-539Crossref PubMed Scopus (3087) Google Scholar As one-dimensional (1D) nanomaterials, CNTs have unique and diverse band structures, making them suitable for various devices. Single-walled CNTs (SWNTs) have high carrier mobilities, ballistic transport, and excellent stability, which are advantageous for nanoelectronics and integrated circuits (ICs).3Shulaker M.M. Hills G. Patil N. Wei H. Chen H.-Y. Wong H.S.P. Mitra S. Carbon nanotube computer.Nature. 2013; 501: 526-530Crossref PubMed Scopus (606) Google Scholar Owing to their direct-band-gap structure and excellent optical and thermal properties, CNTs are also frequently investigated in optoelectronics and thermoelectricity. Additionally, the processability of CNT products allows them to be applied in different devices with different morphologies, such as individual tubes, CNT fibers, and films. In 2019, RV16X-NANO—a microprocessor composed of more than 14,000 complementary metal-oxide semiconductor (CMOS) transistors—was developed,4Hills G. Lau C. Wright A. Fuller S. Bishop M.D. Srimani T. Kanhaiya P. Ho R. Amer A. Stein Y. et al.Modern microprocessor built from complementary carbon nanotube transistors.Nature. 2019; 572: 595-602Crossref PubMed Scopus (56) Google Scholar representing a new step in the device application of CNTs. However, in practical devices, diverse band structures of CNTs make it difficult to improve their performances. For example, in CNT field-effect transistors (FETs), the presence of metallic CNTs can lead to device failure. Therefore, purifying, modulating, or designing the CNT band structures to satisfy specific application requirements is crucial for promoting the development of CNT-based devices. For the diverse electrical properties of CNTs, the enrichment of specific electrical properties is essential because it determines the corresponding application fields, such as semiconducting tubes for FETs5Wang C. Takei K. Takahashi T. Javey A. Carbon nanotube electronics—moving forward.Chem. Soc. Rev. 2013; 42: 2592-2609Crossref PubMed Google Scholar and metallic tubes for interconnects.2De Volder M.F.L. Tawfick S.H. Baughman R.H. Hart A.J. Carbon nanotubes: present and future commercial applications.Science. 2013; 339: 535-539Crossref PubMed Scopus (3087) Google Scholar For given band structures of CNTs, methods such as doping can be used to tune the energy bands, making them suitable for specific devices such as n-doped CNTs for n-type FETs.6Zhang Z.Y. Wang S. Ding L. Liang X.L. Xu H.L. Shen J. Chen Q. Cui R.L. Li Y. Peng L.-M. High-performance n-type carbon nanotube field-effect transistors with estimated sub-10-ps gate delay.Appl. Phys. Lett. 2008; 92: 133117Crossref Scopus (57) Google Scholar Additionally, addressing some of the intrinsic limitations of CNTs, construction of heterostructures with other functional materials is an effective way to tune the carrier transport at the interface and broaden the application of CNTs, such as CNT-Si heterostructures for solar cells.7Tune D.D. Flavel B.S. Advances in carbon nanotube-silicon heterojunction solar cells.Adv. Energy Mater. 2018; 8: 1703241Crossref Scopus (19) Google Scholar Here, we propose the band engineering of CNTs, which significantly affects the performance of CNT-based devices. We also review the main electrical applications of CNTs and their aggregates in terms of both materials and devices. Applications in microscopic and macroscopic scales are discussed respectively because devices with different sizes have different requirements for materials. Finally, we discuss the development and design of CNT-based synthesis and manufacturing. In the semiconductor industry, the process of optimizing the electrical and optical properties of semiconductor materials via doping, solid solutions, and the construction of heterostructures is called band engineering. We extend the concept of band engineering to CNTs, and all the approaches for optimizing the band structures of CNTs to improve the device performance are referred to as CNT band engineering. In this section, we introduce the basic physics of CNT band structures and then introduce band engineering, including the selective enrichment of CNTs with specific electrical properties, modulation of energy bands, and construction of heterostructures based on CNTs. The understanding of the band structures of CNTs should start from their atomic structures. An individual SWNT can be structurally regarded as a hollow cylinder rolled up from a graphene sheet. As shown in Figure 1A, the structure of an SWNT can be specifically represented by a vector C, which indicates the direction and length of the rolling up and is called the chiral vector. The chiral vector can be obtained from the basis vectors of the graphene sheet a1 and a2 as C = na1 + ma2, and this SWNT can be labeled according to its chiral index as (n, m) (n and m are integers, and n ≥ m). A given SWNT can also be specified by its chiral angle θ and diameter d, where θ is defined by the angle between C and a1. The structural parameters for an SWNT can thus be obtained as follows:8Charlier J.-C. Blase X. Roche S. Electronic and transport properties of nanotubes.Rev. Mod. Phys. 2007; 79: 677-732Crossref Scopus (1013) Google Scholar,9Laird E.A. Kuemmeth F. Steele G.A. Grove-Rasmussen K. Nygård J. Flensberg K. Kouwenhoven L.P. Quantum transport in carbon nanotubes.Rev. Mod. Phys. 2015; 87: 703-764Crossref Scopus (157) Google Scholar|a1|=|a2|=3acc=0.246nmaccistheC−Cbondlength,|C|=0.246n2+nm+m2,d=0.246πn2+nm+m2,θ=tan−13m2n+m. Figure 1B shows three typical SWNT structures: zigzag tubes (m = 0, θ = 0), armchair tubes (n = m, θ = 30°), and chiral tubes (n ≠ m, 0 < |θ| < 30°). The atomic structures of SWNTs have been directly observed via high-resolution transmission electron microscopy (TEM) and scanning tunneling microscopy (STM) (Figures 1C and 1D).10Warner J.H. Young N.P. Kirkland A.I. Briggs G.A.D. Resolving strain in carbon nanotubes at the atomic level.Nat. Mater. 2011; 10: 958-962Crossref PubMed Scopus (46) Google Scholar,11Venema L.C. Wildöer J.W.G. Dekker C. Rinzler G.A. Smalley R.E. STM atomic resolution images of single-wall carbon nanotubes.Appl. Phys. A. 1998; 66: S153-S155Crossref Scopus (29) Google Scholar The chiral indices (n, m) reflect the geometric structures and determine the metallic or semiconducting behaviors of SWNTs. The band structure of SWNTs can be derived from that of graphene. According to the zone-folding approximation, the band structure of graphene is approximately unperturbed after rolling up, except for the introduction of a periodic boundary condition.13Saito R. Fujita M. Dresselhaus G. Dresselhaus M.S. Electronic structure of chiral graphene tubules.Appl. Phys. Lett. 1992; 60: 2204-2206Crossref Scopus (2425) Google Scholar,14Hamada N. Sawada S.-i. Oshiyama A. New one-dimensional conductors: graphitic microtubules.Phys. Rev. Lett. 1992; 68: 1579-1581Crossref PubMed Scopus (3094) Google Scholar This quantized boundary condition is expressed as k · C = 2πp, where p is an integer. The perpendicular component of k to the tube axis is kc = 2p/d.9Laird E.A. Kuemmeth F. Steele G.A. Grove-Rasmussen K. Nygård J. Flensberg K. Kouwenhoven L.P. Quantum transport in carbon nanotubes.Rev. Mod. Phys. 2015; 87: 703-764Crossref Scopus (157) Google Scholar Thus, in the reciprocal space, only a series of parallel quantization lines with a space of 2/d are allowed in the Brillouin zone of graphene. The band structure of a specific SWNT can be obtained along these cutting lines, whose orientation, length, and number depend on the chiral index of the SWNT.8Charlier J.-C. Blase X. Roche S. Electronic and transport properties of nanotubes.Rev. Mod. Phys. 2007; 79: 677-732Crossref Scopus (1013) Google Scholar The band gap of the SWNT depends on the minimum distance of the quantization lines from the Dirac points.9Laird E.A. Kuemmeth F. Steele G.A. Grove-Rasmussen K. Nygård J. Flensberg K. Kouwenhoven L.P. Quantum transport in carbon nanotubes.Rev. Mod. Phys. 2015; 87: 703-764Crossref Scopus (157) Google Scholar According to the aforementioned simple scheme, if one of the quantization lines passes straight through the Dirac points of the graphene (Figure 1E), the SWNT has a zero band gap and is metallic. Otherwise (Figure 1F), the dispersion relation exhibits a pair of hyperbolas with a band gap, and it is a semiconducting tube. A simple rule for determining the electrical property is as follows: if (n − m)/3 is an integer, the tubes are metallic; otherwise, they are semiconducting, with a band gap.13Saito R. Fujita M. Dresselhaus G. Dresselhaus M.S. Electronic structure of chiral graphene tubules.Appl. Phys. Lett. 1992; 60: 2204-2206Crossref Scopus (2425) Google Scholar, 14Hamada N. Sawada S.-i. Oshiyama A. New one-dimensional conductors: graphitic microtubules.Phys. Rev. Lett. 1992; 68: 1579-1581Crossref PubMed Scopus (3094) Google Scholar, 15Saito R. Fujita M. Dresselhaus G. Dresselhaus M.S. Electronic structure of graphene tubules based on C60.Phys. Rev. B. 1992; 46: 1804-1811Crossref PubMed Scopus (0) Google Scholar Experimentally, “narrow-gap” behavior is more common than metallic behavior, and yields a small band gap (Eg ≈ kBT) at room temperature.16Ouyang M. Huang J.-L. Cheung C.L. Lieber C.M. Energy gaps in “metallic” single-walled carbon nanotubes.Science. 2001; 292: 702-705Crossref PubMed Scopus (573) Google Scholar This is due to the perturbations of the symmetry of the carbon bonds outside of the “zone-folding approximation.” The perturbations may be curvature or strain.17Blase X. Benedict L.X. Shirley E.L. Louie S.G. Hybridization effects and metallicity in small radius carbon nanotubes.Phys. Rev. Lett. 1994; 72: 1878-1881Crossref PubMed Scopus (0) Google Scholar The calculated 1D dispersion relations of two typical tubes are presented in Figure 1G, showing a metallic (9, 0) tube with the Fermi surface located at Γ and a semiconducting (10, 0) tube with a band gap at Γ.8Charlier J.-C. Blase X. Roche S. Electronic and transport properties of nanotubes.Rev. Mod. Phys. 2007; 79: 677-732Crossref Scopus (1013) Google Scholar Figures 1H and 1I present typical schematics of the density of states (DOS) of semiconducting and metallic SWNTs, showing Van Hove singularities18Charlier J.C. Lambin P. Electronic structure of carbon nanotubes with chiral symmetry.Phys. Rev. B. 1998; 57: R15037-R15039Crossref Scopus (165) Google Scholar and electronic transitions between them. In semiconducting SWNTs, S11 and S22 correspond to the first and second interband transitions, respectively. For a hole-doped semiconducting SWNT, the introduction of an S11-doped energy level may result in intraband transitions involving free carriers (S1fc) and may contribute to the far-infrared absorption. For a metallic SWNT, M11 is the first interband transition, and M00 is derived from the aforementioned “narrow-gap” behavior or the pseudogap phenomenon,19Delaney P. Choi H.J. Ihm J. Louie S.G. Cohen M.L. Broken symmetry and pseudogaps in ropes of carbon nanotubes.Nature. 1998; 391: 466-468Crossref Scopus (309) Google Scholar,20Stahl H. Appenzeller J. Martel R. Avouris P. Lengeler B. Intertube coupling in ropes of single-wall carbon nanotubes.Phys. Rev. Lett. 2000; 85: 5186-5189Crossref PubMed Scopus (211) Google Scholar which may also contribute to the far-infrared absorption.12Itkis M.E. Niyogi S. Meng M.E. Hamon M.A. Hu H. Haddon R.C. Spectroscopic study of the Fermi level electronic structure of single-walled carbon nanotubes.Nano Lett. 2002; 2: 155-159Crossref Scopus (259) Google Scholar Further modulation of the energy bands will be discussed later in this paper. According to the foregoing, SWNTs have different band structures according to their atomic structures and local environment, resulting in various electrical properties. In practical applications, uniform band structures and specific electrical properties are usually beneficial to the device performance, which points to the importance of CNT band engineering. One of the most intriguing properties of CNTs is that their electronic structure is closely related to their geometric structure. As mentioned above, under normal conditions approximately one-third of CNTs are metallic and two-thirds are semiconducting. However, a uniform electronic type of CNTs is needed to achieve high performance of certain electrical devices. There are two main ways to increase the electronic purity of CNTs: the direct growth methods and post-treatment. Figure 2A presents typical methods for selectively enriching semiconducting tubes and corresponding purities. Direct growth methods include controlled growth and selective etching during growth. Controlled growth, which is mainly based on catalyst design and growth condition design,25Yang F. Wang X. Zhang D.Q. Yang J. Luo D. Xu Z.W. Wei J.K. Wang J.Q. Xu Z. Peng F. et al.Chirality-specific growth of single-walled carbon nanotubes on solid alloy catalysts.Nature. 2014; 510: 522-524Crossref PubMed Scopus (448) Google Scholar,26Zhang S.C. Hu Y. Wu J.X. Liu D. Kang L.X. Zhao Q.C. Zhang J. Selective scission of C-O and C-C bonds in ethanol using bimetal catalysts for the preferential growth of semiconducting SWNT arrays.J. Am. Chem. Soc. 2015; 137: 1012-1015Crossref PubMed Scopus (0) Google Scholar highly depends on the intrinsic property differences of chiral structures or band structures. Zhang et al. realized the enrichment of metallic (12, 6) tubes and semiconducting (8, 4) tubes using solid Mo2C and WC catalysts, respectively, by thermodynamically and kinetically controlled chemical vapor deposition (CVD) growth (Figure 2B).21Zhang S. Kang L. Wang X. Tong L. Yang L. Wang Z. Qi K. Deng S. Li Q. Bai X. et al.Arrays of horizontal carbon nanotubes of controlled chirality grown using designed catalysts.Nature. 2017; 543: 234-238Crossref PubMed Scopus (163) Google Scholar They further designed a near-equilibrium nucleation process and enriched a new family of (n, n − 1) semiconducting SWNTs.27Zhang S. Wang X. Yao F. He M. Lin D. Ma H. Sun Y. Zhao Q. Liu K. Ding F. et al.Controllable growth of (n, n - 1) family of semiconducting carbon nanotubes.Chem. 2019; 5: 1182-1193Abstract Full Text Full Text PDF Scopus (8) Google Scholar Wang et al. claimed that when an electric field is introduced as a perturbation, metallic CNTs (m-CNTs) may be twisted into semiconducting CNTs (s-CNTs) owing to the electronic DOS-difference-caused “renucleation barrier difference” between m-CNTs and s-CNTs. This electro-renucleation (ERN) approach led to nearly defect-free s-CNT arrays with <0.1% residual m-CNTs (Figure 2C).22Wang J. Jin X. Liu Z. Yu G. Ji Q. Wei H. Zhang J. Zhang K. Li D. Yuan Z. et al.Growing highly pure semiconducting carbon nanotubes by electrotwisting the helicity.Nat. Catal. 2018; 1: 326-331Crossref Scopus (20) Google Scholar Moreover, Zhu et al. reported the spontaneous purification of CNTs with 99.9999% s-CNTs when the lengths were over 154 mm.23Zhu Z. Wei N. Cheng W. Shen B. Sun S. Gao J. Wen Q. Zhang R. Xu J. Wang Y. et al.Rate-selected growth of ultrapure semiconducting carbon nanotube arrays.Nat. Commun. 2019; 10: 4467Crossref PubMed Scopus (2) Google Scholar These rate-selected s-CNT arrays were based on strong interlocking between the atomic assembly rate and the band-gap structure of the CNTs, which resulted in nearly 10-fold faster decay rate of the m-CNTs than s-CNTs (Figure 2D). On the other hand, differences in the chemical reactivity of s-CNTs and m-CNTs are always considered for selectively etching specific CNTs that react more actively with the introduced reagents or applied field. m-CNTs tend to be more reactive than s-CNTs with similar diameters owing to their different ionization energies. Oxidation by etchants28Kang L.X. Zhang S.C. Li Q.W. Zhang J. Growth of horizontal semiconducting SWNT arrays with density higher than 100 tubes/μm using ethanol/methane chemical vapor deposition.J. Am. Chem. Soc. 2016; 138: 6727-6730Crossref PubMed Scopus (23) Google Scholar or perturbation via an external field29Zhang G.Y. Qi P.F. Wang X.R. Lu Y.R. Li X.L. Tu R. Bangsaruntip S. Mann D. Zhang L. Dai H.J. Selective etching of metallic carbon nanotubes by gas-phase reaction.Science. 2006; 314: 974-977Crossref PubMed Scopus (441) Google Scholar,30Hong G. Zhang B. Peng B.H. Zhang J. Choi W.M. Choi J.Y. Kim J.M. Liu Z.F. Direct growth of semiconducting single-walled carbon nanotube array.J. Am. Chem. Soc. 2009; 131: 14642-14643Crossref PubMed Scopus (118) Google Scholar with high energy (e.g., plasma, UV light, microwave) have been proposed to enrich s-CNTs in situ. Nevertheless, a tradeoff must be made between the semiconducting selectivity and the array density because one-third of the CNTs are removed. The weak oxidative effect during CNT synthesis yields a relatively low selectivity, as there is little difference between the two types of CNTs during nucleation and growth in CVD. Direct growth methods have considerable application potential because of the much easier preparation process without the introduction of impurities. However, achieving a high purity and large-scale synthesis remain challenging. In addition to etching during the growth process, post-growth separation, such as ex situ removal, is also a desirable way to purify CNTs, e.g., selective reactions or selective wrapping. Owing to their higher conductivity compared with s-CNTs, m-CNTs can produce more Joule heat (induced by large currents). Jin et al. used this mechanism to enrich polymer-coated s-CNTs while exposed m-CNTs were removed by reactive-ion etching.31Jin S.H. Dunham S.N. Song J.Z. Xie X. Kim J.H. Lu C.F. Islam A. Du F. Kim J. Felts J. et al.Using nanoscale thermocapillary flows to create arrays of purely semiconducting single-walled carbon nanotubes.Nat. Nanotechnol. 2013; 8: 347-355Crossref PubMed Scopus (124) Google Scholar Other methods that initiate thermocapillary flows and selectively heat m-CNTs using microwave radiation32Xie X. Jin S.H. Wahab M.A. Islam A.E. Zhang C.X. Du F. Seabron E. Lu T.J. Dunham S.N. Cheong H.I. et al.Microwave purification of large-area horizontally aligned arrays of single-walled carbon nanotubes.Nat. Commun. 2014; 5: 5332Crossref PubMed Scopus (31) Google Scholar or infrared laser irradiation33Du F. Felts J.R. Xie X. Song J.Z. Li Y.H. Rosenberger M.R. Islam A.E. Jin S.H. Dunham S.N. Zhang C.X. et al.Laser-induced nanoscale thermocapillary flow for purification of aligned arrays of single-walled carbon nanotubes.ACS Nano. 2014; 8: 12641-12649Crossref PubMed Scopus (28) Google Scholar can be adopted to achieve similar results. Selective wrapping for purifying CNTs involves noncovalent selective interactions between CNTs and other molecules. For instance, a polydimethylsiloxane-based smart Scotch tape34Hong G. Zhou M. Zhang R.O.X. Hou S.M. Choi W. Woo Y.S. Choi J.Y. Liu Z.F. Zhang J. Separation of metallic and semiconducting single-walled carbon nanotube arrays by "Scotch tape".Angew. Chem. Int. Ed. 2011; 50: 6819-6823Crossref PubMed Scopus (57) Google Scholar can be used to extract s-CNTs or m-CNTs from the array mixtures, and washing off m-CNTs using sodium dodecyl sulfate35Hu Y. Chen Y.B. Li P. Zhang J. Sorting out semiconducting single-walled carbon nanotube arrays by washing off metallic tubes using SDS aqueous solution.Small. 2013; 9: 1306-1311Crossref PubMed Scopus (18) Google Scholar has proved to be a facile method for separation. However, the final array density for ex situ selective etching or wrapping depends on the density of the original arrays, limiting the applicability of these methods. Among all the post-treatment methods, solution-phase separation exhibits the best reproducibility and yield. Density gradient ultracentrifugation (DGU) (Figure 2E)24Arnold M.S. Green A.A. Hulvat J.F. Stupp S.I. Hersam M.C. Sorting carbon nanotubes by electronic structure using density differentiation.Nat. Nanotechnol. 2006; 1: 60-65Crossref PubMed Scopus (1842) Google Scholar and gel chromatography36Liu H. Nishide D. Tanaka T. Kataura H. Large-scale single-chirality separation of single-wall carbon nanotubes by simple gel chromatography.Nat. Commun. 2011; 2: 309Crossref PubMed Scopus (606) Google Scholar are the most widely used methods. The key factors in these methods are the density gradient medium and the adopted surfactants. Both of these techniques can provide a high selectivity, even a single chirality, with appropriate surfactants. However, the high technical threshold and complexity limit their applicability. Aqueous two-phase extraction,37Khripin C.Y. Fagan J.A. Zheng M. Spontaneous partition of carbon nanotubes in polymer-modified aqueous phases.J. Am. Chem. Soc. 2013; 135: 6822-6825Crossref PubMed Scopus (196) Google Scholar which has low cost, high yield, and high concentration, can be realized within a short separation time but is restricted by the difficulty of strict control of the polymerization degree. Regarding selective dispersion,38Nish A. Hwang J.Y. Doig J. Nicholas R.J. Highly selective dispersion of single-walled carbon nanotubes using aromatic polymers.Nat. Nanotechnol. 2007; 2: 640-646Crossref PubMed Scopus (818) Google Scholar although the separation purity is high, the biomolecules and conjugated polymers are expensive. However, CNTs purified via the aforementioned post-treatments have common drawbacks: they are short in length, have intrinsic structural defects, and require subsequent purification, e.g., sonication; this may induce further damage and overlap to the tubes, which is detrimental to electrical devices. So far, neither the direct growth method nor post-treatment have been ideal for achieving the ultimate goals of 99.9999% semiconducting selectivity and a high density of >125 nanotubes/μm in a horizontal array.39Franklin A.D. The road to carbon nanotube transistors.Nature. 2013; 498: 443Crossref PubMed Scopus (175) Google Scholar More efforts are needed to make a breakthrough. Moreover, the enrichment of CNTs with specific electrical properties determines their applicability in different fields, but for a specific function in a certain device, the design of band structures and band alignment with other materials are still necessary to improve the device performance. In practical devices, specific band structures, such as p-type or n-type semiconductors, closing a band gap in semiconducting CNTs or opening a band gap in metallic CNTs, are usually needed to satisfy specific device requirements. Under ambient conditions, CNTs are p-doped as a result of the physisorption of oxygen molecules on their surfaces. In this section, the reported methods (Figures 3A–3D ) for tuning the energy bands of CNTs are discussed from theoretical and experimental viewpoints. There are two main categories of doping for CNTs: substitutional doping of heteroatoms in the CNT lattice (Figure 3A) and charge-transfer doping (Figure 3B).40Maiti U.N. Lee W.J. Lee J.M. Oh Y. Kim J.Y. Kim J.E. Shim J. Han T.H. Kim S.O. 25th anniversary article: chemically modified/doped carbon nanotubes & graphene for optimized nanostructures & nanodevices.Adv. Mater. 2014; 26: 40-67Crossref PubMed Scopus (383) Google Scholar To realize heteroatom substitution, doped CNTs can be obtained by in situ introduction of foreign atoms in the vapor or through solid sources during traditional synthesis methods of CNTs, such as NH3, B2H6, thiophene/sulfur powder, and triphenylphosphine for N-, B-, S-, and P-doping, respectively. Post-synthetic doping tends to be difficult because of the chemical inertness of CNTs. Therefore, more-reactive oxidized CNTs are frequently employed as the starting material and annealed in the presence of proper doping species. Dopant atoms can create impurity states in the band structure of CNTs and ultimately influence the overall electronic band configuration. The most typical types of doping are N- and B-doping. N-dopants have three possible configurations: quaternary, pyridinic, and pyrrolic. For semiconducting SWNTs, N-doping with a quaternary configuration incorporates localized states into the band gap near the bottom of the conduction band (Figure 3E). Because of the complex hybridization between the impurity state and the existing unoccupied bands, the Fermi level is raised, approaching the conduction band; this endows the CNTs with metallic behavior. For example, STM and scanning tunneling spectroscopy (STS) studies revealed that N-doped multiwalled CNTs (MWNTs) are metallic and exhibit prominent donor peaks above the Fermi energy at approximately 0.18 eV.41Czerw R. Terrones M. Charlier J.C. Blase X. Foley B. Kamalakaran R. Grobert N. Terrones H. Tekleab D. Ajayan P.M. et al.Identification of electron donor states in N-doped carbon nanotubes.Nano Lett. 2001; 1: 457-460Crossref Scopus (633) Google Scholar For metallic SWNTs, the impurity state generated by N-doping resides in the conduction band (Figure 3G).42Sumpter B.G. Meunier V. Romo-Herrera J.M. Cruz-Silva E. Cullen D.A. Terrones H. Smith D.J. Terrones M. Nitrogen-mediated carbon nanotube growth: diameter reduction, metallicity, bundle dispersability, and bamboo-like structure formation.ACS Nano. 2007; 1: 369-375Crossref PubMed Scopus (171) Google Scholar Pyridinic N-doping leads to either p- or n-type doping, depending on the doping level (p-type is more common). If N-doping adopts a pyrrolic configuration, the five-membered ring structure may lead to a positive curvature, promoting tube closure. For B-doping, all three valence electrons of the B atom participate in σ bonding with neighboring C atoms.43Gebhardt J. Koch R.J. Zhao W. Höfert O. Gotterbarm K. Mammadov S. Papp C. Görling A. Steinrück H.P. Seyller T. Growth and electronic structure of boron-doped graphene.Phys. Rev. B. 2013; 87: 155437Crossref Scopus (76) Google Scholar The absence of additional electrons for original Π bonding leads to a p-doping effect. Carroll et al. detected new peaks in the valence band of MWNTs after B-doping using STS.44Carroll D.L. Redlich P. Blase X. Charlier J.C. Curran S. Ajayan P.M. Roth S. Ruhle M. Effects of nanodomain formation on the electronic structure of doped carbon nanot