No abstract

Novel carbon nanotube−metal cluster structures are proposed as prototype systems for molecular recognition at the nanoscale. Ab initio calculations show that already the bare nanotube cluster system displays some specificity because the adsorption of ammonia on a carbon nanotube−Al cluster system is easily detected electrically, while diborane adsorption does not provide an electrical signature. Since there are well-established procedures for attaching molecular receptors to metal clusters, these results provide a "proof-of-principle" for the development of novel, high-specificity molecular sensors.Since their discovery in 1991, carbon nanotubes (CNTs) have become one of the most exciting nanomaterials, combining a range of extraordinary physical properties, such as extremely small size and high aspect ratio, high stiffness and excellent flexibility under different mechanical stimuli, high structural and chemical stability, and a rich spectrum of electrical properties. Many potential applications have been proposed: superstrong composites, energy storage and energy conversion devices, field emission displays, mechanical, chemical and biological probes and sensors, radiation sources, nanoelectronic devices, and more. 1 In this letter we concentrate on the possibility of using carbon nanotubes as chemical sensors and probes. Pioneering experiments from Dai's group 2,3 have shown that at room temperature the resistance of an individual single wall carbon nanotube is very responsive to the adsorption of molecules, such as NO 2 and NH 3 . These measurements demonstrate the excellent sensing capabilities of CNTs, in particular their fast response time, high selectivity, and reversibility. It has also been shown that CNT-based single-molecule biosensors 4-6 can compete with other nanowire nanosensors 7 in detecting biological and chemical molecules. If appropriately functionalized, carbon nanotubes even have the ability to recognize proteins and DNA. [8][9][10] Moreover, their intrinsic strength and resilience makes them ideally suited for ultrasmall sensors capable of exploring complicated geometries at the nanoscale. 5,6 Real-world applications, such as CNTbased resonant-circuit wireless ammonia sensors, have been developed by different groups. 11,12 Due to their strong sp 2 bonding and near perfect hexagonal network, carbon nanotubes are chemically stable and do not form strong chemical bonds with most molecules. However, since experiments have shown that the properties of a CNT can change when it is immersed in a specific chemical or biological environment, there has been a substantial effort toward the development of techniques to enhance the sensing capability of CNTs. The most common route to improvement in reactivity and sensitivity is through functionalization of CNT sidewalls with specific bio/chemical molecules. [4][5][6]8 In fact, chemical functionalization can both ensure better chemical bonding between the nanotube and a specific chemical species as well as improve the selectivity of the adsorptio...

Lu, Krüger, and Pollmann Reply: The preceding Comment [1] criticizes the missing-row asymmetric-dimer (MRAD) model [2] which we have recently suggested in order to rationalize the experimentally observed c͑4 3 2͒ reconstruction of SiC(001). This structural model is based on ab initio energy and grand canonical potential calculations and on the analysis of a whole body of experimental data. We strongly disagree with the notion [1] that the MRAD model is not consistent with most of the existing experimental data, such as photoemission [3] and scanning tunneling microscopy (STM) image [4]. Our calculations have clearly shown that the adsorption of 0.5 monolayers of Si on the Si-terminated SiC(001) surface drastically lowers the surface energy by the formation of asymmetric dimers leading to the observed c͑4 3 2͒ reconstruction. No inconsistency with previous ab initio calculations [5,6] can be observed, since they did not address this adatom induced reconstruction.As to the STM images, we would like to point out that the comparison of our calculated image with the experimental one is very reasonable. First, although the Tersoff-Hamann approach depends in some cases on the altitude selected to represent the surface charge density, our calculations have shown that this dependence plays a negligible role in the case of the MRAD model. For example, basically the same images are obtained when we choose the altitude to be 1.5, 2.5, and 3.5 Å above the upper adatoms. Second, the charge density does not give exactly a z-scale image in the constant current mode, but qualitatively it gives a correct image. The height profiles (see Fig. 4d in our Letter [2]) give a curve in z scale along which the tip goes in the constant current mode. Third, STM elastic-scattering quantum chemistry (ESQC) simulations by the authors of the preceeding Comment [1,4] are based on extended Hückel theory, while our calculations are based on local density approximation. There is not doubt that the LDA is more reliable than the empirical method. Fourth, if the STM-ESQC simulations are applied to our MRAD model, we believe that results similar to ours will be found.A detailed investigation of the temperature induced c͑4 3 2͒ $ ͑2 3 1͒ phase transition [7] is beyond the scope of our Letter [2]. However, one cannot rule out that the addimer breaking at 700 K is impossible just from the experience of investigations carried out for the Si(001) surface. Our calculations have shown that the energy required to break a dimer bond at the Si(001) surface is about 1.6 eV, while it is only 0.5 eV in the case of the SiC(001) surface.In the preceding Comment, Soukiassian et al. have questioned the validity of the LDA, because different LDA methods give different results and because the electronic structure of the c͑4 3 2͒-AUDD and of ͑2 3 1͒ are very similar in LDA calculations but the former is semiconducting with a large gap while the latter is metallic in experiment [7]. In the cluster calculation [6], the

Electrical contact to low-dimensional (low-D) materials is a key to their electronic applications. Traditional metal contacts to low-D semiconductors typically create gap states that can pin the Fermi level (E). However, low-D metals possessing a limited density of states at E can enable gate-tunable work functions and contact barriers. Moreover, a seamless contact with native bonds at the interface, without localized interfacial states, can serve as an optimal electrode. To realize such a seamless contact, one needs to develop atomically precise heterojunctions from the atom up. Here, we demonstrate an all-carbon staircase contact to ultranarrow armchair graphene nanoribbons (aGNRs). The coherent heterostructures of width-variable aGNRs, consisting of 7, 14, 21, and up to 56 carbon atoms across the width, are synthesized by a surface-assisted self-assembly process with a single molecular precursor. The aGNRs exhibit characteristic vibrational modes in Raman spectroscopy. A combined scanning tunneling microscopy and density functional theory study reveals the native covalent-bond nature and quasi-metallic contact characteristics of the interfaces. Our electronic measurements of such seamless GNR staircase constitute a promising first step toward making low resistance contacts.

No abstract

In the bottom-up synthesis of graphene nanoribbons (GNRs) from self-assembled linear polymer intermediates, surface-assisted cyclodehydrogenations usually take place on catalytic metal surfaces. Here we demonstrate the formation of GNRs from quasi-freestanding polymers assisted by hole injections from a scanning tunnelling microscope (STM) tip. While catalytic cyclodehydrogenations typically occur in a domino-like conversion process during the thermal annealing, the hole-injection-assisted reactions happen at selective molecular sites controlled by the STM tip. The charge injections lower the cyclodehydrogenation barrier in the catalyst-free formation of graphitic lattices, and the orbital symmetry conservation rules favour hole rather than electron injections for the GNR formation. The created polymer–GNR intraribbon heterostructures have a type-I energy level alignment and strongly localized interfacial states. This finding points to a new route towards controllable synthesis of freestanding graphitic layers, facilitating the design of on-surface reactions for GNR-based structures.

No abstract

Supramolecular self-assembly on well-defined surfaces provides access to a multitude of nanoscale architectures, including clusters of distinct symmetry and size. The driving forces underlying supramolecular structures generally involve both graphoepitaxy and weak directional nonconvalent interactions. Here we show that functionalizing a benzene molecule with an ethyne group introduces attractive interactions in a 2D geometry, which would otherwise be dominated by intermolecular repulsion. Furthermore, the attractive interactions enable supramolecular self-assembly, wherein a subtle balance between very weak CH/π bonding and molecule-surface interactions produces a well-defined "magic" dimension and chirality of supramolecular clusters. The nature of the process is corroborated by extensive scanning tunneling microscopy/spectroscopy (STM/S) measurements and ab initio calculations, which emphasize the cooperative, multicenter characters of the CH/π interaction. This work points out new possibilities for chemical functionalization of π-conjugated hydrocarbon molecules that may allow for the rational design of supramolecular clusters with a desired shape and size.