TY - JOUR
AU - Nakata,, Toshihiko
AB - Abstract We evaluated an independent multi-walled carbon nanotube (MWNT) probe by using tomography with a high-resolution transmission electron microscope to verify the three-dimensional structure of the probe tip. The new method of probe evaluation revealed the following features: (i) cutting the end of the MWNT probe caused the wall structure to disintegrate and encapsulated the graphene sheets fragmented by the discharged pulse; (ii) the cap of the MWNT probe was an open cylinder covered by walls similar in shape to a rectangular slit; (iii) the grooves of the inner walls of the MWNT probe, which were created by the discharge cutting method, maintained a cylindrical shape that was different from the peeling-off mechanism. tomography, nanotube, TEM, high resolution, scanning probe microscope Introduction Carbon nanotubes (CNTs) have been studied by many researchers for various engineering applications and are considered to be the next frontier of nanotechnology. One such application may be to use an independent CNT for the tip of a scanning probe microscope. Indeed, the use of an independent CNT tip for metrology of large-scale integration (LSI) was suggested [1]. When the half-pitch of the line-and-space in LSIs becomes as fine as 32 nm and beyond, in order to meet the requirements of higher capacity and faster computation in a small space, the metrology must deliver accurate measurement with high reproducibility in the infinitesimal dimensions of an LSI [2,3]. The novel properties, such as high rigidity and strength, of a CNT make it a candidate material for the tip of the probe-microscope metrology used in LSI device manufacturing. However, the physical and mechanical properties of an independent CNT have not been investigated in detail, because the properties strongly depend on graphene defects in the CNT walls as well as production method such as arc-discharging or chemical vapor deposition. In general, the arc-discharge method has been used to produce multi-walled carbon nanotubes (MWNTs) for preparing an independent MWNT probe because it can produce a satisfactory probe for device manufacturing. Fabrication of the probe requires discharge cutting to adjust the CNT to the proper length as a probe. The peeling-off effect has been proposed for discharge cutting of MWNTs during preparation of the probe, in which the MWNT walls are peeled off layer by layer during discharging of direct current [4]. The aim of this study was to evaluate the damage to the MWNT probe caused by discharge cutting with a high pulse current within a few microseconds. Tomography with high-resolution transmission electron microscopy (HREM) has been studied by many researchers [5–7], but has not been attained due to many technological difficulties. If we provide high-resolution evaluation in tomography, then it will be possible to use three-dimensional diagnostics for precise evaluation of the probes. In this report, we introduce a new sample preparation technique for high-resolution tomography and describe the remarkable results obtained by this method. Methods If we observe a probe with a rotation device attached to the HREM, with the rotation axis precisely fixed along a CNT, then tomography of an independent MWNT at the lattice image of graphite layers could be realized by HREM. As a three-dimensional image can be reconstructed from the projection data contained in a set of micrographs, the rotation axis must be precisely straight along the center of the MWNT [5]. If the MWNT is inclined at a few degrees, deviation of the focal point may cause the magnification of the image to change and even a small difference in magnification may cause lack of continuity in lattice spacing. If this happens, three-dimensional images cannot be obtained by tomography. Installation of the independent MWNT on the needle is the most critical step in sample preparation. We used a high-resolution scanning electron microscope (SEM; S-4300, Hitachi High-Technologies, Japan) with a dual stage that allows individual motion of each stage so that pulsed direct current can be introduced between the two stages through a bridge made up of the MWNT [3,4]. The MWNT was welded on the needle by chemical vapor deposition of gold with electron beam irradiation. The particles are formed by decomposition of gaseous dimethyl gold acetylacetonate ((CH3)2Au(acac)) during irradiation in the SEM in vacuo. The reaction can be written as (CH3)2Au(acac) → Au + hydrocarbon + CO2. The angular deviation of the MWNT from the needle was measured on the SEM monitor by a conventional program of critical dimension metrology. An HREM (HF-2000 equipped with cold-FE gun, Hitachi High-Technologies, Japan) was operated at 100 kV with the needle rotation device [8] as shown in Fig. 1. The lattice spacing of the MWNT (0.34 nm) could be observed, provided Scherzer's condition [9] was satisfied: Defocus yields 76.6 nm for our experiment. Fig. 1. Open in new tabDownload slide Schematic view of the needle rotation device for TEM. Fig. 1. Open in new tabDownload slide Schematic view of the needle rotation device for TEM. Point resolution (ds) is given by Point resolution yields 0.316 nm, where l is the wavelength of the electron beam and the relativistic wavelength operated at 100 kV is 0.00370 nm, and Cs = 1.1 mm is spherical aberration. Thus, HREM is feasible for observing lattice images of the MWNT. HREM images were taken every two degrees of 180° probe rotation in order to reconstruct the MWNT tip by using tomography. The three-dimensional reconstructed structure of the MWNT was calculated from multiple HREM images using the tomography technique. We used the simultaneous iterative reconstruction technique (SIRT) [10] built into commercial software (EMIP, Hitachi High-Technologies, Japan) because the contrast of the reconstructed images was satisfactory and line artifacts were diminished. Results and discussion The MWNT probe after installation on the needle is shown in Fig. 2. Figure 2a shows a typical SEM image of the mounted MWNT. Figure 2b shows a transmission electron microscopic (TEM) image of the MWNT which was evaluated by tomography in this work. The length of the MWNT was about 500 nm and its diameter was about 25 nm. The deviation angle of the probe from the axis was also confirmed to be less than 1.0°. A capped MWNT observed as a flat-headed cylindrical probe is preferable to the metrology at the critical dimensions of LSI. Fig. 2. Open in new tabDownload slide CNT probe weld on the needle by Au deposition: (a) a typical SEM image; (b) a TEM image of the MWNT evaluated by tomography. Fig. 2. Open in new tabDownload slide CNT probe weld on the needle by Au deposition: (a) a typical SEM image; (b) a TEM image of the MWNT evaluated by tomography. Figure 3 shows slice sections of the reconstructed MWNT. Figure 3a shows a vertical slice section of the reconstructed MWNT. The lattice fringe of the graphite layer of the MWNT spacing with 0.34 nm can be clearly observed in the reconstructed image. An amorphous layer coating the surface of the MWNT is also observed. There were some graphene sheets within the MWNT. Figure 3b shows transverse slice sections of the reconstructed MWNT. Each section is separated by 5.3 nm. The number at the top left in each image of Fig. 3b corresponds to the slice section indicated in Fig. 3a. When determining a reconstructed layer structure, the technical limitations must be considered carefully. When a specimen is very thin, the contrast of an HREM image is linear to the projected potential of the specimen. This is called weak phase object approximation. Even though the MWNT is formed by carbon, the thickness limitation of the weak phase object approximation is not more than 20 nm for the dense graphite area. The maximum thickness of the crystalline region of the graphite wall was about 20 nm. Therefore, the contrasts of the HREM images containing the thick crystalline area were not completely linear to the sample thickness due to the effects of multiple scattering. The intense contrasts in HREM images used for reconstruction were mainly generated by the lattice fringes, which were observed only when the Bragg reflection was satisfied, namely when the lattice plane was approximately parallel to the direction of the incident beam within the permitted excitation error. When using these images for tomography, the intense lattice fringe contrast contributes much and weak contrast from nonparallel graphite sheets contributes little to the reconstruction. Even using these tilted HREM images with the extra contrast for tomography, SIRT gave approximately the proper three-dimensional structure. This was because the contrast of the lattice fringes survived at the proper three-dimensional position and the lattice fringe contrast at improper three-dimensional positions were reduced by overlapping with lattice fringe contrasts from other images at other tilt angles through the simultaneous iterative reconstruction process. However, some imperfections remained in this reconstruction, particularly in the vicinity of the bending area of the graphite walls of the MWNT. In addition, we must consider the possibility that the lattice fringes in the HREM images may shift by focusing. Although the focus was carefully controlled in the experiments, there may have been some reconstruction errors with the order of the lattice spacing. Thus, we cannot discuss the atomic arrangements of the reconstructed MWNT. In this report, the reconstructed image was used for understanding the outline of the MWNT. Fig. 3. Open in new tabDownload slide Slice sections of the reconstructed MWNT: (a) vertical slice section; (b) transverse slice sections of the reconstructed MWNT. The positions of the transverse slice sections are indicated in (a). 1, 0 nm; 2, 5.3 nm; 3, 10.6 nm; 4, 15.9 nm; 5, 21.2 nm; 6, 26.5 nm from the point of 1, respectively. Fig. 3. Open in new tabDownload slide Slice sections of the reconstructed MWNT: (a) vertical slice section; (b) transverse slice sections of the reconstructed MWNT. The positions of the transverse slice sections are indicated in (a). 1, 0 nm; 2, 5.3 nm; 3, 10.6 nm; 4, 15.9 nm; 5, 21.2 nm; 6, 26.5 nm from the point of 1, respectively. The interior angle of the heptagonal MWNT was recognized in the transverse section 31.8 nm distant from the tip of the MWNT (see Fig. 3b1). The strong properties of the MWNT having anisotropic transverse structure may affect the vibration direction of the CNT probe. The number of graphene walls at each section is one index for studying the geometrical construction of an MWNT, that is, the number of pentagons necessary to enclose the MWNT tip. We have considered introducing five pentagons on the MWNT, but the results of the end structure due to discharge cutting did not agree with the graphic cones proposed by Krishnan et al. [11]. The longitudinal section of the MWNT shows peeling-off of the interior walls and encapsulation of graphene at the tip. Although the outer walls seem to cover the tip of the MWNT, they may bend over to protect the capped MWNT at the tip. The amount of debris becomes greater closer to the tip. At the center of the tip where the outer walls swelled, a lattice image could be taken, but the graphene sheets were wavy. Moreover, the curvature of the graphene sheets did not coincide with any known atomic arrangements of a pentagonal-carbon-atom structure of a bent graphene sheet. The pulse power of discharge cutting might exclusively affect the disconnection of the MWNT. There was no evidence of rearrangement of the graphene formation. Figure 4 shows deviation of the vertex of the heptagonal MWNT in the longitudinal direction. The origin of the transverse section was defined far (i.e. at 31.8 nm) from the tip. Superposition of the inner and outer walls of the polygons at (a) 5.3 nm and (b) 26.5 nm from the original point are shown in Fig. 4. The crease in the MWNT wall shows a small displacement in Fig. 4a, but no significant change in the shape was observed. On the other hand, the crease in the MWNT wall at 26.5 nm from the original point shows a remarkable difference from that at the origin (see Fig. 4b). It was found that the crease in the MWNT indicates differences in the positions of the vertices of the polygonal pillar. Such differences in the polygonal pillars may correspond to degradation of crystalline graphite due to discharge cutting. In this study, discharge cutting affected a section about 20-nm long from the edge of the MWNT. The length of damage might depend on the power applied during discharge cutting. Fig. 4. Open in new tabDownload slide Superimposition of the crease in the graphite walls of the MWNT transverse sections: (a) 5.3 nm point shown in Fig. 3 as 2 distant from the point shown in Fig. 3 as 1; (b) 26.5 nm point shown in Fig. 3 as 6 distant from the point shown in Fig. 3 as 1. Fig. 4. Open in new tabDownload slide Superimposition of the crease in the graphite walls of the MWNT transverse sections: (a) 5.3 nm point shown in Fig. 3 as 2 distant from the point shown in Fig. 3 as 1; (b) 26.5 nm point shown in Fig. 3 as 6 distant from the point shown in Fig. 3 as 1. The decrease in wall thickness is proportional to the decrease in the area of the polygons along the MWNT probe. The number of outer walls decreases as a function of distance in the longitudinal direction of the MWNT as shown in Fig. 5. Note that the inner and outer layers of the walls are created by generalized marching cubes (GMCs). The benefit of the three-dimensional observation having a resolution of the lattice image is that we can separate the graphene and amorphous parts of a nanotube. The main feature to be resolved by the lattice image of the reconstructed image is the number of walls. The wall also bends at the tip of the nanotube. The grooved area of the MWNT tip is shown in Fig. 5. The strength of the MWNT tip is thought to be different from that of the main body of the MWNT due to the reduced graphene wall thickness. This structural change may affect the electronic properties. In the next step of our research, we will determine the best cutting conditions for the MWNT for producing a proper probe tip for a probe microscope. We found that the effect of discharge cutting could be observed by using our three-dimensional observation technique, which we will use to analyze the relation between pulse energy and the structural change of graphene. Fig. 5. Open in new tabDownload slide The inner and outer surfaces of the MWNT tip shown by GMCs. This figure is available in black and white in print and in colour at JEM online. Fig. 5. Open in new tabDownload slide The inner and outer surfaces of the MWNT tip shown by GMCs. This figure is available in black and white in print and in colour at JEM online. Conclusions By using the SEM sampling technique, the sample was properly set on the needle stub as the ideal orientation for the tomography. The three-dimensional structure was reconstructed without any missing zone. We could not define the lattice fringe observed in the reconstructed image as the real atomic arrangement of graphite due to the effects of the sample thickness and the electron interference, but we could identify the outline of the MWNT. The technique revealed the following: (1) The discharged pulse used for cutting the end of the MWNT probe caused the wall structure to disintegrate and formed the curved graphene sheet fragments. (2) The cap of the MWNT probe was an open cylinder covered by curved walls, which were shaped similar to a rectangular slit. (3) The walls of the MWNT probe tip, which was cut by the discharge cutting method, maintained a cylindrical arrangement that was different from the peeling-off mechanism. This technique is a useful method for diagnostic analysis of the internal structure and surface geometry of an independent MWNT and can also be used for analyzing other types of probes. Acknowledgements We acknowledge Ms Chie Uchiyama for her assistance in the preparation of MWNT probes. References 1 Dai H , Hafner H , Rinzler AG , Colbert DT , Smalley RE . Nanotubes as nanoprobes in scanning probe microscopy , Nature , 1996 , vol. 384 (pg. 147 - 150 ) Google Scholar Crossref Search ADS WorldCat Crossref 2 Edamura E , Kunitomo Y , Morimoto T , Sekino S , Kurenuma T , Kenbo Y , Watanabe M , Baba S , Hidaka K . New inline AFM metrology for LSI manufacturing at the 45-nm node and beyond , Proc. SPIE , 2007 , vol. 6518 (pg. 65183M-1 - 65183M-9 ) WorldCat 3 Sekino S , Morimoto T , Kurenuma T , Hirooka M , Tanaka H . Dimension controlled CNT probe of AFM metrology tool for 45-nm node and beyond , Proc. SPIE , 2008 , vol. 6922 (pg. 6922L-1 - 6922L-12 ) WorldCat 4 Fujieda T , Hidaka K , Hayashibara M , Kamino T , Matsumoto H , Ose Y , Abe H , Shimizu T , Tokumoto H . In situ observation of field emissions from an individual carbon nanotube by Lorenz microscopy , Appl. Phys. Lett. , 2004 , vol. 85 (pg. 5739 - 5741 ) Google Scholar Crossref Search ADS WorldCat Crossref 5 Van Dyck D , Van Aert S , Croitoru M . Atomic resolution electron tomography: a dream? , Int. J. Mater. Res. (formerly Z. Metallkd.) , 2006 , vol. 97 (pg. 872 - 879 ) Google Scholar Crossref Search ADS WorldCat 6 Li ZY , Young NP , Di Vece M , Palmer RE , Bleloch AL , Curley BC , Johnston RL , Jiang J , Yuan J . Three-dimensional atomic structure of size-selected gold nanoclusters , Nature , 2008 , vol. 451 (pg. 46 - 48 ) Google Scholar Crossref Search ADS PubMed WorldCat Crossref 7 Jinschek JR , Batenburg KJ , Calderon HA , Kilaas R , Radmilovic V , Kisielowiski C . 3-D reconstruction of the atomic positions in a simulated gold nanocrystal based on discrete tomography: prospects of atomic resolution electron tomography , Ultramicroscopy , 2007 , vol. 108 (pg. 589 - 604 ) Google Scholar Crossref Search ADS PubMed WorldCat Crossref 8 Yaguchi T , Kamino T , Onishi T , Hashimoto T , Asayama K . Three-dimensional device evaluation system for advanced semiconductor processes , Hitachi Hyoron , 2004 , vol. 86 (pg. 481 - 484 ) WorldCat 9 Scherzer O . The theoretical resolution limit of the electron microscope , J. Appl. Phys. , 1949 , vol. 20 (pg. 20 - 29 ) Google Scholar Crossref Search ADS WorldCat Crossref 10 Gilbert P . Iterative methods for the three-dimensional reconstruction of an object from projections , J. Theor. Biol. , 1972 , vol. 36 (pg. 105 - 117 ) Google Scholar Crossref Search ADS PubMed WorldCat Crossref 11 Krishnan A , Dujardin E , Treacy MMJ , Hugdal J , Lynum S , Ebbesen TW . Graphic cones and the nucleation of curved carbon surfaces , Nature , 1997 , vol. 388 (pg. 451 - 454 ) Google Scholar Crossref Search ADS WorldCat Crossref © The Author 2010. Published by Oxford University Press [on behalf of Japanese Society of Microscopy]. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
TI - Three-dimensional evaluation of an independent multi-walled carbon nanotube probe by tomography with high-resolution transmission electron microscope
JO - Journal of Electron Microscopy
DO - 10.1093/jmicro/dfq072
DA - 2011-02-01
UR - https://www.deepdyve.com/lp/oxford-university-press/three-dimensional-evaluation-of-an-independent-multi-walled-carbon-HAWwQJb1yt
SP - 19
EP - 24
VL - 60
IS - 1
DP - DeepDyve
ER -