Nanomaterial Engineering and Property Studies ... - Wiley Online Library

6 downloads 0 Views 14MB Size Report
Oct 14, 2011 - Modern methods of in situ transmission electron microscopy (TEM) allow .... Photographic views of (a) a scanning tunneling microscope.
www.advmat.de

www.MaterialsViews.com

PROGRESS REPORT

Nanomaterial Engineering and Property Studies in a Transmission Electron Microscope Dmitri Golberg,* Pedro M.F.J. Costa, Ming-Sheng Wang, Xianlong Wei, Dai-Ming Tang, Zhi Xu, Yang Huang, Ujjal K. Gautam, Baodan Liu, Haibo Zeng, Naoyki Kawamoto, Chunyi Zhi, Masanori Mitome, and Yoshio Bando level (that gives the clearest picture free of artifacts) is of prime importance as far as its real, rather than speculative integrations into modern technologies are concerned. Such properties should precisely be measured in order to clearly demonstrate that such integration and, thus, substitution of existing well-tested and historically reliable materials, e.g., Si in electronics, is essential, realistic and profitable. However, in most cases, the property measurements are performed using instruments with no direct access to the nanomaterial internal structure, namely, scanning electron microscope (SEM) and atomic force microscope (AFM).[1,2] This has significantly limited the relevance of acquired data since any particular structural features of an object prior/during/ after its testing have entirely been hidden. Therefore, the measured properties have not been directly linked to a particular nano-morphology, crystallography, spatially-resolved chemistry and existing and/or appearing defects. This explains a common and wide scatter of the mechanical and electrical data reported by various scientific

Modern methods of in situ transmission electron microscopy (TEM) allow one to not only manipulate with a nanoscale object at the nanometer-range precision but also to get deep insights into its physical and chemical statuses. Dedicated TEM holders combining the capabilities of a conventional highresolution TEM instrument and atomic force -, and/or scanning tunneling microscopy probes become the powerful tools in nanomaterials analysis. This progress report highlights the past, present and future of these exciting methods based on the extensive authors endeavors over the last five years. The objects of interest are diverse. They include carbon, boron nitride and other inorganic one- and two-dimensional nanoscale materials, e.g., nanotubes, nanowires and nanosheets. The key point of all experiments discussed is that the mechanical and electrical transport data are acquired on an individual nanostructure level under ultimately high spatial, temporal and energy resolution achievable in TEM, and thus can directly be linked to morphological, structural and chemical peculiarities of a given nanomaterial.

1. Introduction The exact knowledge of physical and chemical properties of a given nanomaterial, in particular on the individual structural Prof. D. Golberg, Dr. X. L. Wei, Dr. D.-M. Tang, Dr. N. Kawamoto, Dr. M. Mitome Nanotube Unit, International Center for Materials Nanoarchitectonics (MANA) National Institute for Materials Science (NIMS) Namiki 1–1, Tsukuba, Ibaraki, 3050044, Japan E-mail: [email protected] Dr. P. M. F. J. Costa CICECO, Department of Ceramics and Glass Engineering University of Aveiro 3810-193 Aveiro (Portugal) and IFW Dresden Helmholtzstrasse 20 01069 Dresden, Germany Dr. M.-S. Wang Laboratory for Nanophotonics and Electronics Department of Materials Science and Engineering Massachusetts Institute of Technology 77 Massachusetts Avenue Cambridge, MA 02139, USA Dr. D.-M. Tang Shenyang National Laboratory for Materials Science Institute of Metal Research Chinese Academy of Sciences 72 Wenhua Road Shenyang, 110016, P.R. China

Dr. Z. Xu Beijing National Laboratory for Condensed Matter Physics Institute of Physics Chinese Academy of Sciences, Beijing, 100190, P.R. China Dr. Y. Huang School of Mechanical and Mining Engineering University of Queensland, St. Lucia, QLD 4072, Australia Dr. H. Zeng, Dr. C. Y. Zhi, Prof. Y. Bando Inorganic Nanostructures Unit International Center for Materials Nanoarchitectonics (MANA) National Institute for Materials Science (NIMS) Namiki 1–1 Tsukuba, Ibaraki, 3050044, Japan Dr. U. K. Gautam New Chemistry Unit Jawaharlal Nehru Centre for Advanced Scientific Research (JNCASR) Jakkur, Bangalore, 560064, India Prof. B. Liu School of Materials Science and Engineering Dalian University of Technology Dalian, 116024, P.R. China

DOI: 10.1002/adma.201102579

Adv. Mater. 2012, 24, 177–194

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

177

PROGRESS REPORT

www.advmat.de

www.MaterialsViews.com

groups which has confused the practical engineers and led to many uncertainties with respect to the real nanomaterials’ potentials. Only rather recently the efforts with respect to true nanomaterial property measurements under the high-spatial resolutions peculiar to transmission electron microscopes (TEM) have become popular worldwide and started to attract full attention.[3–14] In order to perform such studies, special types of dedicated TEM holders, with either scanning tunneling microscope (STM) or AFM capabilities, have been designed. Such holders have been commercialized, for instance, by ‘‘Nanofactory Instruments AB’’, Goteborg (Sweden).[15] In this Progress Report we demonstrate the full usefulness of these advanced in situ TEM techniques during mechanical property analysis, e.g., elasticity, plasticity and strength data while employing direct bent or tensile tests, detailed electrical transport tracing and on-demand nanoengineering (thinning, filling/emptying, soldering, doping etc.) of inorganic nanostructures. The objects highlighted in this contribution are empty or metal-, or halide-filled multi-walled carbon nanotubes, multiwalled boron nitride (BN) nanotubes of various diameters, semiconducting galium nitride (GaN) nanowires and nanotubes, and standard C “black,” and novel BN “white” graphene-like nanosheets. In addition, several important functions which the nanostructures are believed to be good for, like field emission, charge and mass transports are modeled inside TEM giving the exact process clues which have never been uncovered before. It is noted that, to date, apart from the case of standard multi-walled carbon nanotubes, most of the electrical, mechanical and/or thermal properties of many individual inorganic nanostructures, and their hybrides, have remained basically unknown. Fabrication of the numerous pristine, doped and filled carbon nanotubes,[16–24] boron nitride nanotubes,[25–33] inorganic nanowires[34–36] and C and BN nanosheets[37–40] for the below described in situ probing is a challenging work, in which our group has historically had significant achievements. In order to gain clear insights into the properties of these nanoinorganics both STM-TEM and AFM-TEM ‘‘Nanofactory Instruments’’ holders were adopted. The general setups of the holders employed are shown in Figure 1. Both holder types were assembled within a JEOL JEM-3100FEF (Omega Filter) field-emission high-resolution TEM (HRTEM) with a spatial resolution of 0.17 nm. The microscope is operated at 300 kV and has energy dispersion X-ray (EDX) and electron energy loss spectroscopy (EELS) capabilities for spatially resolved chemical analyses and elemental mapping of the tested nanomaterials before/during and after probing. The procedure for accurate mounting of a nanosample within a holder is the key step in obtaining the reliable and reproducible property data set afterwards. In our case it was typically designed as follows. Firstly, a freshly cut gold (or tungsten) wires (250 μm in diameter) were delicately immersed into a given nanomaterial powder and, then, placed on either fixed or movable terminal of the STM holder, or on the movable part of the AFM holder. The copper hats with the sample-holding wires or sharp metal tips were mounted on a sapphire ball of the piezo-driven tubes. The hats embraced the sapphire balls of a piezo-motor rod using six legs, as shown in Figure 1a and Figure 1b. It is noted that during the sample deposition the 178

wileyonlinelibrary.com

Dmitri Golberg earned his PhD in Solid State Physics from Moscow Institute for Ferrous Metallurgy in 1990. He worked as a Fellow of the Japan Society for Promotion of Science, University of Tsukuba, Japan, and MaxPlanck-Society, Duesseldorf, Germany, before joining National Institute for Materials Science (NIMS), Tsukuba, Japan, in 2001 as a Senior Researcher. Currently, he is a Nanotube Unit Director and a Principal Investigator of the International Center for Materials Nanoarchitectonics (MANA) of NIMS, and a Professor of the University of Tsukuba.

nanoobjects were attracted to the metal wires due to simple physical adhesion forces. No organic pastes were used for sample assembling in order to avoid circuit contaminations. The counterparts of the two-terminal holders consisted of a sharply etched fixed gold (or tungsten) tips (STM-TEM) or silicon micro-cantilevers (AFM-TEM), either non-conducting (Si) or conducting (Si coated with a thin layer, ∼15 nm, of

Figure 1.  Photographic views of (a) a scanning tunneling microscope (STM)–transmission electron microscope (TEM); and (b) atomic force microscope (AFM)–TEM “Nanofactory” setups used in the present work. The structural details of both holders are arrowed and highlighted.

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2012, 24, 177–194

www.advmat.de

www.MaterialsViews.com

2. Nanoengineering of Carbon Nanotubes Carbon nanotubes (CNTs) are the first nanotube system discovered and explored. To date, thousands of research papers

Adv. Mater. 2012, 24, 177–194

PROGRESS REPORT

Pt). With the help of an optical microscope and through delicate manipulations with tweezers, the minimum possible gap between the sample wires and the probes was achieved. Then during the piezo-motor-driven manipulations inside TEM the positions of the wires with the loaded nanosamples (or metal STM tips) were accurately adjusted in three dimensions, X, Y and Z, inside the pole piece of the microscope with a precision better than 1 nm. Finally, the relative heights of a STM tip (or a Si cantilever) and the nanoobjects were accurately set using the TEM wobbler function. After this, the tight physical contact between the nanoobject of interest and the STM or AFM probes was achieved. For the STM-TEM system the sharply etched gold or tungsten STM tips (or sample wires, depending on the circuit polarity) could be biased up to ±140 V. This was highly desirable for the analysis of electrical transport in inorganic nanostructures, e.g., BN, GaN, which, as a rule, are not good electrical conductors (except the case of metallic carbon nanotubes). The force measurements by the AFM-TEM holder were accomplished by a Microelectromechanical Systems (MEMS) sensor located at the bottom of a Si cantilever, Figure 1b. Prior to measurements the spring constants of cantilevers were calculated and the (MEMS) sensor mV-nN ratios were calibrated using soft pre-indentation of a blank metal wire. In contrast to many pre-existing methods of nanostructure property evaluations mentioned above (e.g., SEM, AFM), for the regarded in situ TEM systems precise structural and chemical analyses of a nanotube sample before, during and after manipulation/deformation/electrical probing were performed using parallel highresolution imaging, electron diffraction and spatially resolved spectroscopic methods. This is the most relevant to the growing area of nanodevice/nanosystem/nanoarthitectonics research for uncovering not false information as regards the nanomaterial mechanical, electrical and thermal performances in various interconnectors, transducers, field emitters, sensors, actuators and mass conveyers. This Progress Report is organized as follows. After this introduction the 2nd section gives the electrically-driven experiments on standard multi-walled carbon nanotubes and their engineering, soldering and testing; 3rd section presents the in situ TEM studies on nanotubes filled with metals or inorganics; 4th part illustrates direct mechanical bending and tensile tests on multi-walled boron nitride and single-walled carbon nanotubes; 5th section describes STM-TEM experiments on GaN nanotubes and nanowires; section 6 illustrates the electrical probing and manipulation with “hot” two-dimensional nanostructures, i.e. “black” C and “white” BN graphenes; 7th section shows several practical schemes for the in situ analysis of functional nanomaterials properties, e.g., field-emission, electrical switching, conductivity modulations under on-demand doping; and, finally, section 8 provides the authors outlook on the future developments of in situ TEM by taking as an example the stateof-the-art local temperature measurements on nanotubes and graphenes inside TEM.

have been emerged in diverse areas and fields, from CNT mechanics and electronics to its biological and medical applications. However, the potentials of nanotubes for industries have yet been realized and, notably, since recently, the interest of researchers in nanocarbon system has shifted towards a sister material - graphene, a flat sheet of graphite, or, in some sense, the unwrapped CNT. Historically, the electronic applications of nanotubes have grabbed most of the attention due to possibilities of a ballistic transport, huge current densities and transistor applications. The usage of nanotubes in field-effect transistors and interconnects has indeed been nicely demonstrated on the laboratory level,[41] but yet been pursued to the real markets. Clearly, each nanotube electrical contact within a transistor or other electrical circuit becomes a demanding issue - its quality, e.g., Ohmic or Shottky type, determines the overall performance. Typically, low-resistant metallic contacts, e.g., Au, Pt, Pd, Ti, Cr, W are applied to CNTs. However, the detailed picture of the contact alternations and its kinetics during a current flow has been unknown. We model a nanotube/metal contact system inside the STM-TEM holder and traced its performance in real time and at the ultimately high spatial resolution of 0.17 nm.[42,43] Figure 2 illustrates this difficult experiment. The setup is depicted in Figure 2a, where the tungsten electrode has been chosen. During a marginal current flow through the circuit, i.e. 1–2 μA, and affiliated Joule heating, we noticed a drastic change in the contact morphology, which has never been mentioned before. The metal electrode starts to swell (Figure 2c and Figure 2d) and its TEM contrast dramatically changes. In addition, the tube morphology itself also alters; the former hardly visible tube channel gradually becomes wider. Finally, new carbon nanophases, which have not existed prior to biasing, become noticeable on the electrode, e.g., new carbon tubular layers, shielding the electrode, and nested fullerene-like particles, Figure 2e and Figure 2f. This sequence of events was further analyzed using energy filtering TEM, electron diffraction and high-resolution imaging. Closer to a contact point, i.e. ball-like tip electrode end, the carbon map becomes bright, while it is fully dark within the electrode body. Combined with the electron diffraction analysis it was concluded that the bright part represents not a pure metal anymore, but a tungsten carbide. Under continuous biasing, the interface between the carbide phase and the pure metal was driven from the contact point to the electrode body. The whole kinetics was understood as follows. Under Joule heating and a directed current flow, C species from the tube diffuse into the electrode body. Once the C concentration exceeds the solubility limit, a phase transformation occurs and a tungsten carbide (WC) phase forms. Under a continuous feedstock of C atoms the carbide phase expands into the metal body. Once the solubility limit exceeds that natural for WC, C atoms precipitate from it in the form of new carbon nanophases, either tubular layers or nested fullerenes. Interestingly, such newly formed C-WC-W circuit is quite different from the original C-W one. It becomes much less electrically resistant, compared to a tube which has a simple physical contact between its layers and the electrode, and ultimately strong. It now requires a stress of ∼6 GPa to break the formed C-WC-C heterostructure, as revealed by direct tensile tests employing the AFM-TEM holder.[44] Such strength is already comparable to that of a spider silk and pure

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

179

www.advmat.de

PROGRESS REPORT

www.MaterialsViews.com

post-synthesis treatments. Such on-demand channel widening is useful for nano-transport and nano-delivery CNT functions discussed later. Modification of C nanotubes morphologies, their interconnections and branching is also nicely shown in the next in situ TEM experiment. In that case, we need a Co particle filled in the C nanotube to start with.[45,46] The experimental layout is marked in Figure 3a. Under in-tandem convergent electron beam irradiation and a current flow through the whole Au-Co@ CNT-W circuit (where @ stands for encapsulated in) a series of complex intradiffusion processes takes place. And the formerly filled CNT transforms to the resultant Co-clamped CNT with independent two CNT branches welded to the Co particle. The electrical resistance of the joined state becomes lower than that of the filled state, Figure 3b. Using the designed process branched tubes may be created in various shapes. Through physical overlapping of two tubes using in situ TEM nanomanipulations, biasing and irradiating, a 4-branched tube with a central Co node was created, Figure 3c and Figure 3d. Such Cojoined states were found to be both mechanically robust and electrically conductive. The two developed methods of making electrically and mechanically advantageous metal-nanotube contacts should be of high value in the nanotube electronics which, by all means, presumes contacting external metal leads to a nanotube. Finally, a word of caution should be mentioned in regards of treating the electrical data from such metal-CNT circuits. They cannot be envisioned as a solid and electrically unchangeable system. The effects of C bulk and surface diffusion, Joule overheating and morphology changes should be kept in mind, and all necessary measures should be undertaken in order to minimize the negative effects of these changes on a given electrical nanodevice performance.

3. Phase Transformations, Mechanics and Electronics of Metals and Inorganics Confined in Carbon Nanotubes

Figure 2.  (a) Schematic setup illustrating an individual multi-walled carbon nanotube placed inside the STM-TEM holder; (b-f) TEM images showing consecutive stages of the carbon nanotube/tungsten electrode interfacial changes under a current flow through the CNT-W contact. The initial tube (a) with the ill-defined channel transforms to that with a wide open channel while C diffuses through the interface and finally precipitates on the originally clean metal electrode in the form on new tubular C shells (f). White arrows in (c) and (d) depict the structural boundary between tungsten carbide and pure elemental tungsten; whereas black arrowhead in (e) shows the onset of the first C layer. Figure adapted from Ref. [42].

tungsten, and dramatically exceeds all values reported in the literature for metal/nanotube interconnects’ strengths. It worth noting that the regarded process can also be treated in another practical way: the nanotube channel (due to a loss of diffusing C atoms towards the electrode) can easily be re-shaped and made wider (Figure 2f), that is difficult to achieve under any other 180

wileyonlinelibrary.com

Previously, in situ TEM has been used for the investigation of high temperature structural transformations, such as superplastic deformation of CNTs,[47] electromigration of metals filled in CNTs[4] and formation of metal-CNT junctions.[42–46] These experiments have been conducted under gradual Joule heating and, therefore, nearly equilibrium processes are usually observed. In the newest experiments of us, a fast current pulse was applied to a metal-filled CNT (Co or Fe fillings) for the investigation of possible metal structural transformations at the extreme and non-equilibrium conditions.[48] Figure 4a and Figure 4b show the designed experimental configuration and its TEM view, respectively. A metal-filled CNT is connected to gold and tungsten electrodes in a TEM-STM holder. Figure 4c and Figure 4e display high-resolution TEM images (HRTEM) of Fe and Co single crystals encapsulated into multi-walled CNTs. After a flash current pulse, the structures have immediately and completely transformed into disordered, amorphous-like phases, as marked on the corresponding HRTEM images in Figure 4d and Figure 4f. The corresponding fast Fourie transform (FFT) patterns are shown on the insets. This further documents the

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2012, 24, 177–194

www.advmat.de

www.MaterialsViews.com

PROGRESS REPORT

the boxed area. Initially, an Au-contacted tube has its channel fully filled, Figure 5c. However, after applying a few μA current pulse a wide section of the channel becomes empty (filling level marked with an arrow). Figure 5e demonstrates in parallel measured current-voltage characteristics of the structure configurations in Figure 5c and Figure 5d. The electrical performance of the tube drastically changes. The system resistance drops. The electrical conductance is related to the extension of the empty channel by a negative exponential, as summarized in Figure 5f. The filling levels are arrowed in correspondence to points 3 and 5 in the curve. This in situ recorded tube emptying process gives a rare possibility to: 1) modify the CNT electrical performance within a desired range which can hardly be altered by all other means, like control of chirality or doping; 2) to deliver a sublimable material to a desired place. The spatial and mass accuracy of such material delivery is extremely high (1 nm and 10−18 g, respectively) which implies Figure 3.  (a) Schematic setup showing a Co particle filled multi-walled C nanotube placed possibly smart usages of the phenomenon in inside the STM-TEM holder and simultaneously experiencing electron irradiation under a the prototype drug delivery systems. 300 kV electron beam, and a passing current through the circuit; (b) Comparative I–V plots showing the decreased electrical resistance under a structural change of the original Co-filled C This sort of experiments was also extended nanotube to resultant Co-particle-branched tube after in-tandem electron irradiation and a curto the more complex nanosystems, i.e. rent flow leading to a complex intra-diffusion process; the inset depicts the low-magnification multi-walled carbon nanotube filled with a TEM image of the starting setup; (c) and (d) Higher magnification TEM images illustrating Zn0.92Ga0.08S compound placed between two the formation of a 4-branched Co-clamped nanoarchitecture (d) from an overlay of the two conductive clamps.[20–23] After applying a current separated and physically crossed tubes (c) under the regarded intradiffusion-induced structural transformation. Co clamping is rather robust and withstands a ∼5 GPa tensile stress, as was pulse followed by a compressive load, the origirevealed by a direct AFM-TEM test. Figure adapted from Refs. [43] and [45]. nally fully filled matter transforms to discrete incorporated domains. Interestingly, during in regarded crystalline-to-amorphous phase transitions. This set of situ bending, the kinks only appear on the empty sections of the experiments opens up an interesting and rich field of the metal carbon shell. Comparative in situ measured stress-strain curves of research. A CNT may be imagined as a nano-container or nanothe structures show that the emptied tube becomes highly flexible crucible for diverse metallurgical operations at the nanoscale. and soft, whereas the filled tube has had a meaningful force under Unusual P-T conditions may exist in such nanovessels due to bending. This means that the intuitively expecting (from a macextremely small sizes and corresponding confinement effects. roworld experience) gain in the hollow material’s rigidness, while The regarded metal amorphization processes may be underit is filled with a solid, works for the small system too. This result stood in terms of the starting single crystals flash melting under is important in regards of CNT mechanical usage in composite the immediate Joule heating and extremely high cooling rates research: the tubes which are filled should deliver more strength achieved due to small system size and superb thermal conducto a composite rather than those which are empty and which have tivity of a nanotube container. The metals undergo a sort of commonly been utilised to date. melt quenching being first rapidly heated and confined in tube Using the regarded emptying process the nanodelivery run channels, and then suddenly cooled at a cooling rate exceeding mentioned above for the CuI system was also attempted on several million K s−1 once the electrical pulse is seized up. This short Zn0.92Ga0.08S-filled CNT contacted to two metal electrodes new phenomenon is related to the exciting area of so-called inside TEM.[21] After the current pulse, the nanotube becomes “nanometallurgy” which may bring to the front some rare, or nearly empty with the majority of the core filling delivered to an even unknown, metal phases with peculiar properties. allocated position. Another type of fillings studied by us in TEM were the inorganics having much lower meting and sublimation points com4. Mechanical Properties of Boron Nitride pared to standard metals.[19] Figure 5a depicts a curled copper and Carbon Nanotubes iodide (CuI)-filled CNT before its contacting to the gold electrode. The inset shows an EDX spectrum recorded from the filling and 4.1. Bending Tests showing the presence of Cu and I species. Figure 5b presents the structural details of a filled nanotube where the well-ordered lattices of carbon shells and the halide core are visible. This is addiUtilization of in situ TEM techniques for mechanical testing tionally confirmed using a fast-Fourier transform (FFT) pattern of of individual nanostructures is a relatively new field. This

Adv. Mater. 2012, 24, 177–194

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

181

PROGRESS REPORT

www.advmat.de

www.MaterialsViews.com

bent until it kinks (drastic force drop) and then starts to be superplastically bent without any force required. Second, once a thick tube approaches the cantilever, it starts to deform elastically, revealing the different curve slop. After force-to-stress recalibration using tube cross-section values accurately measured in HRTEM, the two structures give exactly the same stress values pointing at the full validity of data. Note, that even a very fine detail during the deformation, i.e. a sudden slippage of the thick tube over the cantilever at its displacement of 413 nm was nicely recorded as a small force drop. The value of such drop was only 9 nN, which may be considered to be above the sensor sensitivity limit. The serrations on the F-d plot of the order of 1–2 nN give the sensitivity range and the associated uncertainty of the force value measurements. Another important characteristic of the tube deformation and the sensor/piezo-tube performances is a complete reversibility of both force and displacement values during loading/reloading cycles, Figure 6b. The details of kinking deformation and a superplastic flow mentioned above are repeated herein, and, in addition, the F-d plot on a backward move comes exactly to the same point where it started from. Thus the regarded AFM-TEM holder has an excellent performance as regards evaluation of the deformation properties of nanoobjects with the ultimately high precision. We then tested very thin BN multi-walled nanotubes, 5 shells only, whose properties have remained unknown in a bending test.[33] The F-d curve, featuring both elastic and nearly superplastic regions, once Figure 4.  (a) Schematics illustrating a metal filled multi-walled C nanotube placed inside the the force reached ∼25 nN, was recorded. No STM-TEM holder and experiencing a short voltage/current pulse; (b) TEM image of the setup; significant deformation strengthening was (c) and (d) transformation from a crystalline (c) to amorphous (d) iron wire under an electrical pulse; (e) and (f) analogous transformation in the filled cobalt wire. The insets in (d) and seen. The tube deformed easily and showed (f) show halo-like electron diffraction patterns manifesting the amorphous states of the metal no any traces of the residual plastic deforfillings after electrical pulsing.[48] mation on unloading after multiple cycles. This type of elastic kinking is a natural feapowerful technique of gaining deep insights into the nanomature of BN nanotubes highlighting their excellent potentials in terial bending, compression and tensile properties has only mechanical applications. It is noted that the F-d curves for thin recently received the deserved attention of researchers. AFM tubes were much more serrated compared to the case of relaholders integrated within a standard TEM instrument make tively thick nanotubes in Figure 6 due to the smaller system size use of MEMS sensors attached to the bottom of a Si cantilever, and high sensitivity of the MEMS sensor to the tube clamping Figure 1b. Deflection of the cantilever during nanostructure quality. The thinner insulating BN tubes are not that stable loading gives an electrical signal to the sensor which is recalibetween clamps due to charging effects in TEM. brated to get the true values of the force applied. Figure 6 testifies the accuracy and validity of such measurements. Two sets 4.2. Tensile Tests of bending tests on a thin bundle made of two BN nanotubes, one relatively thin and another - relatively thick, are shown in panels 6a and 6b. Various force needed to bend such tubes Although bending experiments of nanoobjects have been percould be recorded due to a different protruding length of the formed in many laboratories possessing the “Nanofactory tubes from the bundle, Figure 6a. A raw force-displacement Instruments” holders and become a sort of routine, the direct plot illustrating all stages of the deformation is shown below tensile tests, which give the true values of the elastic modulus the consecutive TEM images. First, the thin tube is elastically and the ultimate tensile strength have been unavailable. This is 182

wileyonlinelibrary.com

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2012, 24, 177–194

www.advmat.de

www.MaterialsViews.com

PROGRESS REPORT

clamps, Figure 7b. The nanotube was tensed until it broke and the true stress/true strain curve was recalculated, Figure 7c. For doing so we need to accurately determine the crosssectional morphology of a given BN nanotube (stress calculation) and to measure the sole cantilever displacements in regard of the whole system displacement (plus elastic and plastic tube deformations). Thus such deformation cannot be performed continuously since during the loading the electron beam should be blanked to prevent its effect on the electrical circuit and the MEMS sensor. Therefore, a tensile test was carried out in steps, after each stage, the beam was switched on, and the displacements were accurately measured using a high-resolution TEM mode. Then it was switched off again, and the next step of the deformation proceeded. All these experimental tricks allowed us to precisely measure the elastic modulus and the true strength of 14 individual multi-walled BN nanotubes; the representative values obtained are marked on the plot in Figure 7c. It is noted that this data sheds a new light on extreme elasticity and strength of a BN nanotube which a-priori could hardly been expected keeping in mind the BN compound iconicity compared to a covalently bond graphitic structures. Given the ultimately high chemical and thermal inertness of BN nanotubes compared to their C counterparts, the former can be envisaged to be excellent structural materials as far as Figure 5.  (a) A curled CuI@CNT before being contacted to an Au electrode. Inset: EDX spechigh temperature and aggressive environtrum showing the presence of Cu and I. (b) Detail of a CuI@CNT where the well-ordered latments during the nanomaterials’ employtices of the carbon shell and the halide core are visible. Inset: fast-Fourier transform (FFT) of the ment are considered. boxed area. (c) Initial configuration of an Au-contacted CuI@CNT with the nanotube channel As a typical morphology of BN nanotubes, fully filled. (d) After the application of a few μA current pulses a wide section of the channel became empty (filling level marked with an arrow). (e) Current-voltage characteristics of the so-called bamboo-like nanostructures (BNNB), structure configurations (c) and (d). (f) The electrical conductance is related to the extension are quite widely spread and are much more of the empty channel by a negative exponential. Inset: filling levels (arrowed) corresponding to easily fabricated opposed to well-structured points 3 and 5 in the curve. Scale bar: 50 nm. Figure adapted from Ref. [19]. BN tubes. For example, many commercially available tubes named as “BN nanotubes”, because significant difficulties involved in clamping of a nanoe.g., those produced by a famous ball-milling process, are actuscale object inside TEM which is required for a reliable tenally made of nanobamboos. Thus, once one thinks about the sile test. In case of conducting nanotubes the sort of electron structural application of BN nanotubes, the nanobamboo morbeam soldering can be utilised, however this technique does phology should definitely be tested and the right tensile data not work for insulators, like BN nanotubes. The tensile tests should be acquired. We have recently performed such tests.[49] on such nanotubes are thus rather challenging. We developed The analogous clamping technique, as mentioned above, was an original technique of an individual BN nanotube clamping again adopted here. Initially, we noticed that the nanobamboo using a sort of electron-beam induced carbon deposition inside joints may be arranged in two fashions: (1) the neighbouring HRTEM.[30] For this purpose, deliberately, tiny wax droplets bamboo-like sections are tightly connected through the entire were placed on a Si cantilever and a moving electrode, Figure 7a. embracement of the preceding segment by the following one Following-up electron beam irradiation of the spots in the in a way resembling a human bone’s joint; (2) the joints are vicinity of these droplets led to hydrocarbon decompositions loosen while being overlapped only partially. Then, we investiand C precipitation on the tube-lead joints. The nanotube ends gated both joint types to study the influence of interface pecubecame buried under thick debris of an amorphous-like carbon. liarities on the mechanical properties of nanobamboo archiSignificantly large contact area between the nanotube external tectures. The toughest BNNB broke when the force reached surface and an amorphous C deposit allowed the tube to withabout 2400 nN. It was found that the strength of BNNBs had stand decently high level of loadings without slippage over the varied over three multiples depending on the regarded interface

Adv. Mater. 2012, 24, 177–194

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

183

www.advmat.de

PROGRESS REPORT

www.MaterialsViews.com

Figure 6.  (a) A bending experiment using the AFM-TEM holder on an individual thin bundle made of the two tubes: one–relatively thin; the other one– relatively thick; the experimental force-displacement curve with all testing stages linked to the corresponding TEM images on the lower panel. The curve highlights the ultimate sensitivity of the MEMS sensor used: all peculiarities of the deformation, e.g., elastic deformation of the thin tube, its sudden kinking with a force drop followed by a sort of the “superplastic” flow, and elastic deformation of the thick tube are all clearly visible on the curve. Note that even a flash sliding of the thick tube over the cantilever within only a ∼9 nN force drop at 413 nm displacement gives a clear serration on the curve. (b) A full bending cycle on the same bundle showing all the regarded features plus an additional complete tube displacement and force recovery after reloading, e.g. neither plastic flow nor tube damage are seen in spite of the severe tube bending; this manifests the BN tube superelasticity.

184

wileyonlinelibrary.com

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2012, 24, 177–194

www.advmat.de

www.MaterialsViews.com

structure. Due to a geometry-related strengthening, so called interlocking effect, BNNBs with the type-(1) interface, with an acute angle of the merged part section to the axis show the fracture strength and Young modulus of up to ∼8.0 GPa and ∼225 GPa, respectively. Such values may be considered as the decent ones for practical applications. In fact, even non-perfect bamboo-like BN morphologies are still about 20 times stronger than the common steels. This fact opens up a bright prospect for BN tubular materials applications for reinforcement of various polymeric, ceramic and metallic composites.

Adv. Mater. 2012, 24, 177–194

PROGRESS REPORT

Figure 7.  (a) Schematics showing the ideology of the direct tensile test on a multi-walled BN tube. Before test the tube was glued to the cantilever and the retracting tungsten node using a deliberately placed wax droplets within the AFM-TEM setup; (b) Consecutive TEM images showing the tube tension and its final breakage in the end of the experiment; (c) Corresponding stress-strain curve recorded using the MEMS sensor data and detailed HRTEM imaging of all deformation stages. The curve yields the BN tube Young’s modulus and the ultimate tensile strength of 924 GPa and 18.8 GPa, respectively. Figure adapted from Ref. [30].

Another system of interest with respect to tensile tests is single-walled C nanotubes. The mechanical properties of multi-walled C nanotubes have been studied and appeared to be good.[47] However, the measured true strength values were still behind the brightest theoretical estimates pointing at ∼100 GPa strength levels. At present, single-walled C nanotube can routinely be produced, however their mechanical performance has never been evaluated due to a minute size, diameter of ∼0.7 nm, and many experimental difficulties involved in tube visualization, placing and clamping. Thus we decided to start with multi-walled C nanotubes which might easily be seen and manipulated inside TEM.[50] By using the above-mentioned Joule heating under a current flow in the conducting AFM-TEM holder, such tubes were firstly “undressed”: the C layers were sublimated one by one after moderate current pulses, Figure 8a. The process was stopped once a single-layered nanotube fragment was fixed in the tube center, as verified by HRTEM, Figure 8b and Figure 8c. Then a direct tensile test under the tungsten tip retraction was performed, Figure 8d. In such manner twelve individual single-walled C nanotubes were tested in tension. The highest value of the ultimate tensile strength was ∼100 GPa on a defect-free single C shell, thus being exactly as the theory predicts. It is worth mentioning that the pre-existing structural defects, like pentagon-heptagon pairs in a honeycomb graphite lattice led to a drastic decrease in tube strength, down to 25 GPa or less, being the stress-concentrators. The single shell fracture always took place at these points, as was confirmed by high-resolution TEM imaging. These results point at the importance of structural control of the C nanotube shells which are envisaged for mechanical applications as individual structural elements or as reinforcing components in composite matrixes. Avoiding the defects appearance, for example through correctly designed postsynthesis high temperature heat treatments would solve the problems of unexpected failures of the structural components.

5. Studies of Inorganic Nanowire and Nanotube Electrical and Mechanical Properties Over the years using the in situ TEM-STM and AFM instruments we have studied numerous nanoscale inorganics shaped as nanowires, nanobelts, nanosheets and nanocones.[14,34–36] Among those III-V nanoscale semiconductors have received a prime attention. For example, GaN has a wide band gap (3.4 eV),

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

185

www.advmat.de

PROGRESS REPORT

www.MaterialsViews.com

However, all previous evaluations of GaN nanowires and/or nanotubes electronic properties have relied on GaN field effect transistor (FET) geometries on Si substrates.[51,52] A deposition process for such devices included chemical solvents and led to surface contamination. Also the initial surface states were affected by a photoresistor used for fabrication of metal electrodes under electron beam deposition. Such surface contamination plays the key role in the transport when a GaN nanostructure size falls into the nanoscale. Therefore, we have recently attempted the GaN nanowires and nanotubes electrical property analysis in a high vacuum of the high-resolution TEM in a free-standing setup.[34,36] Figure 9a and Figure 9b depict the experimental STM-TEM layout composed of an Au wire, a GaN nanowire and a sharp W tip, and a TEM image showing its real view inside the microscope. Figure 9c shows a representative I–V curve of an individual GaN nanowire under a bias voltage sweeping from -3 to 3 V. The symmetrical I–V plot exhibits a typical nonlinear characteristic of the Schottky contact on both Au-GaN and GaN-W interfaces, rather than linear Ohmic

Figure 8.  (a) Schematics explaining the way of a multi-walled C nanotube shell “undressing” using an electrical current inside the conducting AFM-TEM holder; (b) a starting multi-walled C nanotube TEM view; (c) that after consecutive stepwise shells sublimation under Joule heating yielding only a single-layer tubular structure in the tube center (the singlelayer framed area is enlarged in the inset); (d) TEM view of a resultantly broken tube under a direct tensile test; (e) corresponding force-displacement curve yielding a huge tensile strength of a single tubular C shell of ∼100 GPa. The inset displays a HRTEM image of the broken tube ends. Figure adapted from Ref. [50].

high melting point, high carrier mobility, chemical inertness, abundant optical and electrical properties, and decent integration with the well-established Si-based nanodevice technology. No doubt, GaN nanostructures should be regarded as prospective building blocks for modern electronics. 186

wileyonlinelibrary.com

Figure 9.  (a) Schematics illustrating an individual GaN nanowire inside the STM-TEM holder; (b) corresponding TEM image of the setup; and (c) experimental (black) and theoretical (red) I–V curves of the GaN nanowire exhibiting the perfect match. Figure adapted from Ref. [36].

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2012, 24, 177–194

www.advmat.de

www.MaterialsViews.com

PROGRESS REPORT

The present electrical and mechanical data acquired on the behavior. The currents through the Au-GaN and GaN-W interrepresentative GaN nanowires and nanotubes using in situ faces are rather low or negligible under a small bias voltage TEM setups show that the individual nanostructure perform1 GΩ to 10 kΩ (corresponding to a conductivity of 1.5 × 105 S m−1). An important finding of this dedicated experiment was a clue that under the transformation, first, the epoxy groups had been released from the sheet, and then the hydroxyl ones started to emerge. The first process gave rise to a drastic increase in conductance, whereas the later one only marginally altered the nanosheet transport. Since to date no solid evidences have been demonstrated that during the preparation graphene can selectively be oxidized by one of the regarded oxygen containing groups, the present findings of the effect of different oxygen groups on graphene oxide conductivity and its changes during transition to pure graphene are envisaged to be highly useful for graphene and/or its oxide integrations into future nanotechnologies. 6.3. Electrical Measurements on “White” Graphenes: Boron Nitride Nanoribbons and Nanosheets While graphene synthesis has made significant progress,[59] large-scale and reliable production of inorganic monolayers and studies on their properties remain a major research bottleneck.[61,62] BN two-dimensional structures, as one of the most important representative of inorganic nanosheets so called “white” graphenes, are thought to have mechanical strength and thermal conductivity comparable to C graphenes.[40,63] They are valuable as substrates for graphene electronics, dielectric coatings with high heat dissipation, non-wetting films, supports for catalysts, thermally conductive insulating fillers for polymer or ceramic composites, thermal radiators, deep UV light emitters and lasers, nanoelectronic and spintronic devices, and radiation-hard aerospace applications.[63,64] A particularly appealing feature of BN nanosheets is their superior chemical, thermal and oxidation stability at temperatures up to 1100 °C, compared to about half that temperature for nanostructured carbons. This

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2012, 24, 177–194

www.advmat.de

www.MaterialsViews.com

PROGRESS REPORT

However, there have been a few reports pointing at the marginal conductance appearance under downsizing their dimensions or applying deformations.[26] We synthesized BN nanosheets and nanoribbons via plasma-etching-induced multiwalled nanotube unzipping and analyzed the I-V curves in the two-terminal STM-TEM.[38] Figures 12a–c consecutively show a HRTEM image of a BN nanosheet with the characteristic zig-zag edge, the “white” graphene structural model, and an AFM topography image verifying the sheet thickness of only 0.5 nm. Figure 12c depicts the comparative electrical measurements on various BN nanostructures: e.g., non-stretched nanotubes, nanoribbons and nanosheets. Surprisingly, the latter two structures started to show the I–V curves peculiar to semiconductors rather than to an insulator. For example, high currents up to of ∼15 μA under several dozen volts, high conductance of 104 S m−1 and carrier mobility of 58.8 cm2 V−1 s−1 were measured in such 2D-BN crystals without a gate bias at room temperature. First-principle calculations verified that the edge states and dangling bond states located in the peculiar zig-zag edges, and surface vacancies (due to plasma treatment) were behind this distinct carrier transport alternations. Such way of the 2D-BN electrical conductance tuning is important as far their future electronic and spintronic applications are concerned. Also this results open a very rich area of physics, if one thinks about ensembles of mixed two-dimensional heterostructures composed of pure zeroband-gap C (“black”) and pure wide-band-gap BN (“white”) domains in lateral or transverse directions. Such, so called “meta-materials” should bring to the materials front novel properties and functionalities.

7. Practical Nanomaterial and Nanotechnology Issues Solved 7.1. Electron Field Emitters Figure 11.  (a) Schematics illustrating an individual graphene oxide film placed on the fixed terminal of the multi-probe STM-TEM holder; (b) corresponding TEM image of the setup; (c) and (d) alternations of the HRTEM images under reduction of graphene oxide to graphene under in situ TEM Joule heating. The insets show corresponding fast-Fourie-transform patterns and an enlarged image of the graphene sheet edge; (e) Changes in the conductance and conductivity under gradual transformation of the graphene oxide to graphene during in situ TEM heating; the two different models of bonded oxygen-containing groups (e.g., epoxy and hydroxyl) corresponding to a definite two-stage oxygen release process are shown for clarity. Figures reprinted with permission from Ref. [60]. Copyright (2011) American Chemical Society.

reflects the high promise in nanoelectronics where high currents and considerable resistive heating make a problem for a reliable device use. BN nanostructures are typically insulating. Adv. Mater. 2012, 24, 177–194

Carbon nanotube electron sources have a high promise for commercial field emission (FE) displays and high-resolution electron-beam instruments. Compared with the commonly used emitters such as tungsten, most advantages of a nanotube emitter originate from its closed cap. The CNT caps usually have a small diameter of a few to tens nanometers and C atoms are joined by strong covalent sp2 bonds. Such CNT emitters have low operation voltage and good FE stability. That is because the activation energy for surface migration of

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

189

PROGRESS REPORT

www.advmat.de

www.MaterialsViews.com

FE failure) is shown in Figure 13b. The onset voltage for field evaporation is 54 V, and the corresponding local field on the CNT cap Fevap is calculated to be 9.1 V nm−1 (by Fevap = βVevap; where β is the field-enhancement factor). This value can be considered as the maximum value applied on this CNT cap, and is consistent with the previous results for a critical evaporation field of ∼10 V nm−1. The high robustness of the present CNT-WCW heterostructure (tensile strength of ∼6 GPa or more) and the specific all-shells contact between CNT and intermediate tungsten carbide (that gives extremely low resistance of only hundred Ω) prevent the CNT emitter dispatch from the substrate, and therefore such structure is indeed attractive for the real CNT-based point-electron emitters. However, once an individual CNT is employed as a point electron source its thermal vibration significantly deteriorates the FE performance. The well emitting nanotubes tend to be shorter. The extreme case of a shortened capped multi-walled CNT is a nested C onion. Such emitter would be able to greatly lower FE onset voltage of the tungsten emitter, and to display a fairly stable emission at a maximum FE current comparable to shortest nanotubes. We thus fabricated the regarded onion-like emitter inside TEM using delicate nanomanipulations and current flows in STM-TEM.[65] First, a balllike W tip was moved to face a gold wire. A voltage sweep up to 140 V was applied onto the gold anode. At such conditions Figure 12.  (a) HRTEM image of an individual boron nitride nanosheet taken using aberration no obvious FE current from the W tip was corrected TEM operating at 80 kV; the inset presents the zig-zag edge considered within the detected. Then the W tip touched a C onion boxed area; (b) and (c) a structural model and a standard topography AFM image of the sheet, that was pre-attached to the edge of the gold respectively, demonstrating its atomic thickness; (d) Comparative I-V curves obtained using wire. A bias was applied to the Au wire and a in situ STM-TEM probing on various nano-BNs. BN ribbons and sheets, opposed to standard multi-walled BN nanotubes, reveal a meaningful current. The insets depict an electrical circuit half of the C onion suddenly adhered to the utilized and TEM image of the two-terminal testing scheme for a BN sheet. Figure adapted W end. The top half of the onion was exposed from Ref. [38]. upon the W tip which next was retracted to its original position. The repeated FE measthe C atoms is much larger than for a tungsten electron source, urement showed a low onset voltage (for a current of 1 nA) of which makes the CNT cap withstand the extremely strong only ∼70 V. The newly formed emitter sustained a FE current fields up to ∼10 V nm−1. However, at such strong fields there is of ∼17 μA and its Fowler-Nordheim plot still remained linear a problem of a CNT dispatch from a metal substrate. Through documenting a cold emission. By contrast, previous studies the C/W intradiffusion process under a current flow and associhave reported that a FE current of only ∼1 μA already heats a ated Joule heating (discussed in Section 2), we created a CNT 40 μm CNT to 2000 K at the apex. This resulted in a consideremitter which was bound to a tungsten substrate through an able deviation of the emission law from the electron tunneling intermediate layer of tungsten carbide within the STM-TEM to thermal emission. Therefore, it is concluded that a shortsetup.[46] Such emitter is shown in Figure 13a at different magened emitter length effectively suppressed the negative Joule nifications. Positive bias scans with a short sweeping time of heating effect. In addition, in situ TEM elemental mapping of 200 ms were applied to the counter gold wire, and a FE current the whole emitter showed that a lower half of the C onion had was recorded. The sweeping ranges increased in steps of 5 V already been dissolved into the W head to form a layer with to push the CNT to its current limit. In the end, the tube failed faint contrast. Such layer represents a tungsten carbide (WC) at its middle part (Figure 13b, inset), but not at the contact, whose formation indicates a strong coupling between C onion leading to the removal of the top CNT half and an abrupt curand W tip and the superb emitter robustness analogous to that rent drop. The I–V curve of the last voltage scan (that led to the of CNT emitters discussed above. 190

wileyonlinelibrary.com

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2012, 24, 177–194

www.advmat.de

www.MaterialsViews.com

7.2. Ionic Conductor Atomic Nanoswitch Development of microelectronics makes a call for low-power, low-cost, and high-density memory devices used in portable digital systems. Many efforts have been put in finding socalled atomic or molecular switches that may substitute for those based on Si-based technologies. The possible candidates are ionic and electronic mixed conductor-based solid electrolyte nonvolatile memories, such as silver sulfide (Ag2S). Their switching behavior was attributed to repetitive formation and breakage of the conductive filament inside a solid electrolyte. Some experimental and theoretical studies have been performed to uncover the switching mechanism, but still no solid evidences for the regarded electrical pathways have been found. Furthermore, it has not been clear how a conducting filament grows microscopically given a rigid sulfide lattice response. We entirely reproduced the switching behavior of an Ag/Ag2S/W sandwich structure using the STM-TEM unit, Figure 14.[66] And the chemical composition and crystal structure of the initial state, i.e. the off-state, and the conducting, i.e. the on-state,

Adv. Mater. 2012, 24, 177–194

PROGRESS REPORT

Figure 13.  (a) TEM image of an individual thin-walled C nanotube tested with respect to its field-emission ability inside the STM-TEM setup. The nanotube was attached to the tungsten tip using the above described process of the contact interface modification through the formation of a tungsten carbide under a current flow and Joule heating; thus the contact becomes ultimately robust as evidenced in (b); the tube bottom end remains on the electrode even after the tube tip failure (inset of the boxed area) under an ultimately high emission current. The corresponding I–V FE plot in which the moment of tube evaporation at a bias of 54 V is arrowed. Figure adapted from Ref. [46].

were traced using both spatially-resolved energy dispersive X-ray spectroscopy (EDS) and HRTEM lattice imaging paired with in situ electrical measurements. The conductive pathway was indeed directly imaged under the atomic resolution (∼1.7 Å), Figure 14d. Based on the observations, a refined switching model of an Ag2S-based nonvolatile memory device was proposed. Initially, Ag2S is in the non-conducting acanthite monoclinic phase. When a positive bias is applied, Ag cations start to migrate towards the cathode and are reduced to Ag atoms during their transport; the acanthite phase transforms into the argentite cubic phase (low left inset in Figure 14d). Thus a continuous conducting channel made of argentite and metallic Ag is built, as evidenced by a rising current. Because of the total switch volume expands, the conducting mixture can grow out of the solid Ag2S electrolyte, as seen on the HRTEM image in Figure 14b. Next, when a negative bias is applied, the just-grown Ag crystal dissolves back into the argentite phase, Ag cations migrate towards the anode, and the conducting argentite phase back-transforms into the non-conducting acanthite phase. The former conducting channel is now trespassed by the non-conductive acanthite phase (black contrast on the upper-left phase map in Figure 14d); the switch is turned off. The conducting mixture shrinks back, Figure 14c. If a positive bias is applied again, the broken conductive channel is rebuilt due to the argentite phase and Ag formation. The regarded switch works an indefinite number of cycles without failure, as evidenced by multiple cycling biasing in TEM. It is noted that although the switching behavior has previously been attributed to the growth and dissolution of an Ag crystal out of the Ag2S matrix,[67] the present dedicated TEM experiment further enlightens its physics, chemistry and the working principle. It is worth noting that the Ag2S system is not the only one which exhibits an ionic-based conductivity. Other materials, like Cu2S, Ag2Se, (Pr, Ca)MnO3, SrTiO3, Ta2O5, VO2, TiO2 have already been shown or may have similar functions.[68] The detailed mechanisms of their functioning have not yet been properly elucidated due to the absence of reliable in situ, real time and atomic resolution observations. The TEM experiment described here provides a general route towards answering all related fundamental and practical questions as regards atomic switching operations. 7.3. Doping of Boron Nitride Nanophases with Carbon: Targeting Controllable Conductivity At room temperature, pristine BN nanophases, without defects and/or deformations, do not allow any current to pass through due to their wide and rigid energy band gaps of 5–6 eV. This fact significantly hinders nano-BNs applications for interconnects. By contrast, C nanomaterials may be metals, semimetals or semiconductors. Therefore, the ternary B-C-N system presents an exciting opportunity to prepare diverse structures with various compositions ranging from pure C to pure BN. Film-like bulk B-C-N materials have been known for decades,[69] but large-scale synthesis of ternary B-C-N nanotubes and nanosheets has lingered far behind. Only a few experimental successes have been reported till now,[70] but even in these works the atomic distributions of B, C and N species within the nanophases have basically remained unknown due to spatial resolution limitations and a

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

191

www.advmat.de

PROGRESS REPORT

www.MaterialsViews.com

283.8 eV after the electron beam irradiation and doping. And the final goal of such doping, i.e. turning an insulator into a conductor was indeed accomplished, as verified by the real time in situ electrical measurements in TEM, Figure 15c. Spatially resolved elemental mapping using ELLS showed that C dopants are rather evenly distributed within the doped sheets with a predominant appearance at the sheet edges and/or structural defects, e.g., voids or holes, which are subjected to more pronounced destruction under irradiation due to loosen B and N atoms at the edges. HRTEM imaging revealed that the sheets remained in the well-structured state, no considerable sheet destruction or amorphization took place and C atoms positioned themselves within the honeycomb graphenelike lattice islands rather than being formed amorphous-like coatings on the pristine BN sheets. Such doping strategy may be scaled up using larger beams of electrons, or ions for the preparation of macroamounts of B-C-N nano-semiconductors. Clearly, the beam energies and irradiation times are variables which should determine the doping levels and thus the levels of the on-demand conductivity increase.

8. Future Challenges and Outlook 8.1. Local Temperature Measurements on Individual Nanomaterials Figure 14.  (a-c) TEM images of a silver sulphide ionic conductor nanocrystal in its initial, and ON, and OFF states inside the STM-TEM holder, respectively; (d) a current loop under changing bias with the phase maps for OFF and ON states shown in the insets. A continuous conducting channel made of mixed argentite and pure silver domains in the ON state becomes discrete and discontinuous in the OFF state, this leads to the loss of conductance. Figures reprinted with permission from Ref. [66]. Copyright (2010) American Chemical Society.

lack of appropriate characterizations. We were able to reproduce a single C doping event inside TEM, while taking pure BN nanotubes or graphene-like sheets and/or ribbons (Figure 15a) as mother phases.[31] In order to do so, we intentionally placed a C feedstock (candle wax droplets) inside a pole piece of the highresolution TEM in the vicinity of the moving W electrode of the STM-TEM holder. The TEM image in Figure 15a shows such electrode contacting an individual BN multi-layered (“white”) graphene attached to the gold counterpart electrode. The convergent electron beam was used as a splitter for hydrocarbons which dissociated onto pure carbon and hydrogen species under high dose electron beam irradiation. Surface diffusion of C atoms towards the pristine BN nanophase then took place. Along with hydrocarbon dissociations, the 300 kV electron beam knocked out B and N atoms from a honeycomb BN lattice creating the structural vacancies which were immediately filled by the incoming C atoms. This type of substitution was confirmed by EEL spectroscopy, Figure 15b. The C K-edge appeared at 192

wileyonlinelibrary.com

Although there has been a significant progress in electrical and mechanical property studies on diverse nanomaterials, as evidenced by the preceding chapters, these days it becomes of vital importance to develop novel nanoscale temperature measurements in order to reliably integrate such materials into practice. In fact, overheating and thermal destructions of nanocomponents in modern electronic devices still remain a problem for engineers. Thus nanomaterial thermophysical properties, e.g., Joule heating effects, specific heats, thermal conductivities should be understood prior to the marketing stage. However, such studies have been very much restricted due to minimal dimensionality problem and the lack of the proper experimental techniques. In order to target this goal we introduced a newly-developed twin-probe STMTEM holder (“Nanofactory Instruments”) and for the first time realized local temperature measurements in every arbitrary area of a nanosample.[71] The twin-probe holder employs two independent piezoelectric rods attached with Cu hats to the two independent piezo-tube drivers through the sapphire balls, as shown in Figure 16a and Figure 16b. Two metallic tips are fixed at the grounded parts of the double-probe holder and may be physically and electrically connected to the nanosamples. In this study, Cu and Cu-Ni tips were used for constructing a nano-thermocouple

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2012, 24, 177–194

www.advmat.de

www.MaterialsViews.com

PROGRESS REPORT Figure 16.  (a) The top part of a twin-probe STM-TEM holder and (b) its electrical circuit; (c) and (d) bright-field TEM images of constructed thermocouples attached to Joule heated C nanotube and graphene samples, respectively. A thermoelectromotive force generated between Cu and Cu-Ni tips is measured by a voltmeter.[71]

Figure 15.  (a) Schematics displaying the setup for BN nanosheet electrical probing and chemical engineering inside the STM-TEM holder and the cartoons of the B and N atoms electron irradiation-induced substitutions with C atoms. The latter originated from the decomposition of deliberately placed hydrocarbons in the form of wax droplets on the tungsten electrode inside TEM. Atomic models of the tested objects of substitution, e.g., BN nanosheets, nanoribbons and nanotubes are also drawn; (b) Comparative EEL spectra showing the K-edges of interest before the electron irradiation-induced doping and after that. The appearance of C edge after doping is apparent; (c) Drastic rise in electrical conductivity of originally insulating BN nanosheet after C-doping. Figure adapted from Ref. [39].

which was capable of a local temperature measurement in TEM under Joule heating of a nanostructure. The ultimately thin Cu and Cu-Ni tips were prepared by a chemical etching method. Other thermocouples (depending on the temperature range of interest, e.g., containing Ir, Pt, Rd metals may also be fabricated using FIB techniques). The minimum diameter of the tips-ends was down to 50 nm, implying a high precision of the contacting spot and an intimate contact to the nanostructure. Three-dimensional independent piezoelectric mechanics of the twin-probe holder precisely controlled the desired position of a thermocouple on the object of interest. For example, the thermocouple was successfully placed on individual nanotubes or graphene-like samples, as displayed in Figure 16c and Figure 16d, respectively. Then the electromotive forces, which depend on a temperature difference between a measuring point on a given specimen and a standard point of the thermocouple at room T were measured.

Adv. Mater. 2012, 24, 177–194

The presently developed novel nano-methodology is envisaged to be highly useful for diverse thermophysical studies at the nanoscale. For instance, in situ observations of phenomena involved in phonon excitation due to the electron irradiation and/or resistive heating[72] become possible in real time and at a high spatial, temporal and energy resolution provided by a high-resolution TEM instrument. In addition to the above-mentioned challenging temperature measurements other future directions of in situ TEM should include: (i) increasing complexity of the measuring schemes moving from simple two-terminal to three-terminal and multiprobe measurements which would significantly increase the richness of physics data acquired; (ii) integration of STM- and AFM-holders with the in situ optical capabilities, e.g., real time in situ collection of various wavelength light absorption and transmission spectra under current flows and/or deformations. These would give a direct access to the electronic structure alternations of a nanomaterial. Completion of such challenging tasks would lift up the in situ TEM techniques to the highest quality level in regards of getting deepest insights into the nanomaterial physical and chemical properties, and would make it the prime technique among other rivaling methods which fell short of spatial resolution. The latter may reach 80 pm in the modern aberration-corrected TEM instruments.

Acknowledgements The authors are grateful to the International Center for Materials Nanoarchitectonics (MANA) of the National Institute for Materials Science (NIMS), Tsukuba, Japan, for a continuous support of the in situ TEM Project. PMFJC acknowledges support from the Portuguese Foundation for Science and Technology (Ciencia 2007 Fellowship), the Alexander von Humboldt Foundation (Experienced Researcher

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

193

PROGRESS REPORT

www.advmat.de

www.MaterialsViews.com

Fellowship), and Short Term Visiting Research Fellowship tenable at MANA. DT acknowledges MOST (Grant 2011CB932601) and NSFC (Grants 50921004 and 50872137) projects, Peoples Republic of China. Received: July 6, 2011 Published online: October 14, 2011

[1] E. W. Wong, P. E. Sheehan, C.M. Lieber, Science 1997, 277, 1971. [2] J. P. Salvetat, A. D. Briggs, J. M. Bonard, R. R. Bacsa, A. J. Kulik, T. Stockli, N. A. Burnham, L. Forro, Phys. Rev. Lett. 1999, 82, 944. [3] Y. Nakayama, S. Akita, New J. Phys. 2003, 5, 128. [4] K. Svensson, H. Olin, E. Olsson, Phys. Rev. Lett. 2004, 93, 145901. [5] B. C. Regan, S. Aloni, R. O. Ritchie, U. Dahmen, A. Zettl, Nature 2004, 428, 924. [6] M. S. Wang, L. M. Peng, J. Y. Wang, Q. Chen, Adv. Funct. Mater. 2006, 16, 1462. [7] L. X. Dong, X. Y. Tao, L. Zhang, X. B. Zhang, B. J. Nelson, Nano Lett. 2006, 7, 58. [8] J. Y. Huang, S. Chen, Z. Q. Wang, L. Kempa, Y. M. Wang, S. H. Jo, G. Chen, M. S. Dresselhaus, Z. F. Ren, Nature 2006, 439, 281. [9] D. Golberg, M. Mitome, K. Kurashima, C. Y. Zhi, C. C. Tang, Y. Bando, O. Lourie, Appl. Phys. Lett. 2006, 88, 123101. [10] K. Jensen, W. Mickelson, A. Kis, A. Zettl, Phys. Rev. B. 2007, 76, 195436. [11] P. M. F. J. Costa, D. Golberg, M. Mitome, Y. Bando, J. Mater. Sci. 2007, 43, 1460. [12] P. M. F. J. Costa, X. S. Fan, S. Wang, Y. H. He, Y. Bando, M. Mitome, J. Zou, H. Huang, D. Golberg, J. Micr. Res. Technol. 2009, 72, 93. [13] D. Golberg, P. M. F. J. Costa, M. Mitome, Y. Bando, Nano Res. 2008, 1, 166. [14] D. Golberg, P. M. F. J Costa, M. Mitome, Y. Bando, J. Mater. Chem. 2009, 19, 909. [15] http://www.nanofactory.com [16] D. Golberg, P. M. F. J. Costa, M. Mitome, Ch. Mueller, S. Hampel, A. Leonhardt, Y. Bando, Adv. Mater. 2007, 19, 1937. [17] P. M. F. J. Costa, D. Golberg, M. Mitome, Y. Bando, Appl. Phys. Lett. 2007, 91, 2231081. [18] P. M. F. J. Costa, D. Golberg, M. Mitome, Y. Bando, Appl. Phys. A. Mater. Sci. Proc. 2008, 90, 225. [19] P. M. F. J. Costa, D. Golberg, M. Mitome, S. Hampel, A. Leonhardt, B. Buchner, Y. Bando, Nano Lett. 2008, 8, 3120. [20] P. M. F. J. Costa, U. Gautam, M. S. Wang, Y. Bando, D. Golberg, Carbon 2009, 47, 541. [21] P. M. F. J. Costa, U. K. Gautam, Y. Bando, D. Golberg, Carbon 2011, 49, 342. [22] P. M. F. J. Costa, P. B. Cachim, U. K. Gautam, Y. Bando, D. Golberg, Nanotechnology 2009, 20, 405706. [23] P. M. F. J. Costa, P. B. Cachim, U. K. Gautam, Y. Bando, D. Golberg, Nanotechnology 2009, 20, 405707. [24] X. L. Wei, M. S. Wang, Y. Bando, D. Golberg, Sci. Tech. Adv. Mater. 2011, 12, 044605. [25] D. Golberg, X. D. Bai, M. Mitome, C. C. Tang, C. Y. Zhi, Y. Bando, Acta Mater. 2007, 55, 1293. [26] X. D. Bai, D. Golberg, Y. Bando, C. Y. Zhi, C. C. Tang, M. Mitome, K. Kurashima, Nano Lett. 2007, 7, 632. [27] D. Golberg, P. M. F. J. Costa, O. Lourie, M. Mitome, C. C. Tang, C. Y. Zhi, K. Kurashima, Y. Bando, Nano Lett. 2007, 7, 2146. [28] Z. Xu, D. Golberg, Y. Bando, Nano Lett. 2009, 9, 2251. [29] Z. Xu, D. Golberg, Y. Bando, Chem. Phys. Lett. 2009, 480, 110. [30] X. L. Wei, M. S. Wang, Y. Bando, D. Golberg, Adv. Mater. 2010, 22, 4895. [31] X. L. Wei, M. S. Wang, Y. Bando, D. Golberg, J. Am. Chem. Soc. 2010, 132, 13592. [32] R. Arenal, M. S. Wang, Z. Xu, A. Loiseau, D. Golberg, Nanotechnology 2011, 22, 265704. [33] Y. Huang, Y. Lin, C. C. Tang, Y. Bando, C. Y. Zhi, D. Golberg, 2011, unpublished.

194

wileyonlinelibrary.com

[34] B. D. Liu, Y. Bando, M. S. Wang, C. C. Tang, M. Mitome, D. Golberg, Nanotechnology 2009, 20, 185705. [35] B. D. Liu, Y. Bando, M. S. Wang, C. Y. Zhi, X. S. Fang, C. C. Tang, M. Mitome, D. Golberg, J. Appl. Phys. 2009, 106, 034302. [36] B. D. Liu, Y. Bando, M. S. Wang, D. Golberg, J. Nanosci. Nano­ technol. 2010, 10, 3945. [37] A. K. Schaper, H. Q. Hou, M. S. Wang, Y. Bando, D. Golberg, Carbon 2011, 49, 1821. [38] H. Zeng, C. Y. Zhi, Y. Bando, Z. H. Zhang, X. L. Wei, X. B. Wang, W. L. Guo, Y. Bando, D. Golberg, Nano Lett. 2010, 10, 5049. [39] X. L. Wei, M. S. Wang, Y. Bando, D. Golberg, ACS Nano 2011, 5, 2916. [40] D. Golberg, Nat. Nanotechnol. 2011, 6, 200. [41] P. Avouris, R. Martel, MRS Bull. 2010, 35, 306. [42] M. S. Wang, D. Golberg, Y. Bando, Adv. Mater. 2010, 22, 93. [43] J. A. Rodriguez-Manzo, F. Banhart, M. Terrones, H. Terrones, N. Grobert, P. M. Ajayan, B. G. Sumpter, V. Meunier, M. S. Wang, Y. Bando, D. Golberg, Proc. Natl. Acad. Sci. USA 2009, 106, 4591. [44] M. S. Wang, D. Golberg, Y. Bando, Adv. Mater. 2010, 22, 5350. [45] J. A. Rodriguez-Manzo, M. S. Wang, F. Banhart, Y. Bando, D. Golberg, Adv. Mater. 2009, 21, 4477. [46] M. S. Wang, Y. Bando, J. A. Rodriguez-Manzo, F. Banhart, D. Golberg, ACS Nano 2009, 3, 2632. [47] B. Peng, M. Locascio, P. Zapol, P. S. Li, S. L. Mielke, G. C. Schatz, H. D. Espinosa, Nat. Nanotechnol. 2008, 3, 626. [48] D. M. Tang, M. S. Wang, Y. Bando, D. Golberg, 2011, unpublished. [49] D.M. Tang, C. L. Ren, X. L. Wei, M. S. Wang, C. Liu, Y. Bando, D. Golberg, ACS Nano 2011; doi: 10.1021//nn202283a. [50] M. S. Wang, D. Golberg, Y. Bando, Adv. Mater. 2010, 22, 4071. [51] Y. Huang, X. Duan, Y. Cui, C. M. Lieber, Nano. Lett. 2002, 2, 101. [52] T. Kuykendall, P. Pauzauskie, S. Lee, Y. Zhang, J. Goldberger, P. Yang, Nano Lett. 2003, 3, 1063. [53] R. Calarco, M. Marso, T. Richter, A. I. Aykanat, R. Meijers, A. V. D. Hart, T. Soica, H. Luth, Nano Lett. 2005, 5, 981. [54] C. Y. Nam, P. Jaroenapibal, D. Tham, D. E. Luzzi, S. Evoy, J. E. Fischer, Nano Lett. 2006, 6153. [55] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva, A. A. Firsov, Science 2004, 306, 666. [56] A. K. Geim, Science 2009, 324, 1530 [57] D. Rangappa, K. Sone, M. S. Wang, U. K. Gautam, D. Golberg, H. Itoh, M. Ichihara, I. Honma, Chem. Eur. J. 2010, 16, 6488. [58] M. S. Wang, D. Golberg, Y. Bando, 2011, unpublished. [59] S. Park, R. S. Ruoff, Nature Nanotechnol. 2009, 4, 217. [60] Z. Xu, Y. Bando, L. Liu, W. L. Wang, X. D. Bai, D. Golberg, ACS Nano 2011, 5, 4401. [61] D. Pacilé, J. C. Meyer, C. O. Girit, A. Zettl, Appl. Phys. Lett. 2008, 92, 133107. [62] W. Q. Han, L. Wu, Y. Zhu, K. Watanabe, T. Taniguchi, Appl. Phys. Lett. 2008, 93, 223103. [63] D. Golberg, Y. Bando, Y. Huang, T. Terao, M. Mitome, C. C. Tang, C. Y. Zhi, ACS Nano 2010, 4, 2979. [64] D. Golberg, Y. Bando, Y. Huang, Z. Xu, X. L. Wei, L. Bourgeois, M. S. Wang, H. Zeng, J. Lin, C. Y. Zhi, Isr. J. Chem. 2010, 50, 405. [65] M. S. Wang, D. Golberg, Y. Bando, ACS Nano 2010, 4, 4396. [66] Z. Xu, Y. Bando, W. L. Wang, X. D. Bai, D. Golberg, ACS Nano 2010, 4, 2515. [67] K. Terabe, T. Hasegawa, T. Nakayama, M. Aono, Nature 2005, 433, 47. [68] R. Waser, M. Aono, Nat. Mater. 2007, 6, 833. [69] M. Kawaguchi, Adv. Mater. 1997, 9, 615. [70] D. Golberg, Y. Bando, P. S. Dorozhkin, Z. C. Dong, MRS Bull. 2004, 1, 38. [71] N. Kawamoto, M. S. Wang, M. Mitome, D. Golberg, Y. Bando, 2011, unpublished. [72] P. M. F. J. Costa, U. K. Gautam, Y. Bando, D. Golberg, Nat. Commun. 2011, 2, 421.

© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Mater. 2012, 24, 177–194