Ultramicroscopy 148 (2015) 75–80
Contents lists available at ScienceDirect
Ultramicroscopy journal homepage: www.elsevier.com/locate/ultramic
Imaging of soft material with carbon nanotube tip using near-ﬁeld scanning microwave microscopy Zhe Wu a,c, Wei-qiang Sun d, Tao Feng b, Shawn Wenjie Tang a, Gang Li b, Kai-li Jiang e, Sheng-yong Xu b,n, Chong Kim Ong a,nn a
Centre for Superconducting and Magnetic Materials, Department of Physics, National University of Singapore, Singapore 117542, Singapore Key Laboratory for Physics and Chemistry of Nanodevices, Department of Electronics, Peking University, Beijing 100871, PR China c National Laboratory for Vacuum Electronics, School of Physical Electronics, University of Electronic Science and Technology of China, Chengdu 610054, PR China d Singapore Synchrotron Light Source, National University of Singapore, Singapore 117603, Singapore e Department of Physics & Tsinghua-Foxconn Nanotechnology Research Centre, Tsinghua University, Beijing 100084, PR China b
art ic l e i nf o
a b s t r a c t
Article history: Received 25 April 2014 Received in revised form 2 September 2014 Accepted 22 September 2014 Available online 6 October 2014
In this manuscript, a near-ﬁeld scanning microwave microscope (NSMM) of our own design is introduced while using a multi-walled carbon nanotube (MWCNT) bundle as the tip (referred to as ‘CNT tip’). Clear images of gold-patterned numbers, photoresist stripes and corneal endothelial cells (cell line B4G12) were obtained by mapping the resonant frequency fr and S11 amplitude of a given area while the NSMM is operating in tapping mode. The CNT tip helps to improve image quality and reveals more information about the sample as compared to a traditional metallic tip. The CNT tip is ﬂexible and does not scratch the surface of the sample during the scan, which is useful for imaging soft material in biological science. In the imaging of the B4G12 endothelial cells, the nuclei and cytoplasm can be clearly distinguished from the rest of the cell and its surrounding medium. & 2014 Elsevier B.V. All rights reserved.
Keywords: Carbon nanotube Soft material Near-ﬁeld scanning microwave microscopy Endothelial cell
1. Introduction The developments of the near-ﬁeld scanning microwave microscope (NSMM) over the last two decades have enhanced the NSMM's ability in making local quantitative measurements such as the dielectric constant and to perform surface imaging. The NSMM is used to observe minute changes in the structure and electromagnetic properties of the sample that perturb the electrical interaction between the probe tip and the sample in the microwave near-ﬁeld region. These perturbations change the amplitude of the reﬂected signal S11 and shifts the resonant frequency fr, which can be measured by the NSMM. As a result, the NSMM [1–3] is capable of revealing the distribution of electrical properties of bulk materials and dielectric ﬁlms in ‘soft contact’ mode by applying a tiny force between the probe and the sample. However, a metallic tip is required to have a non-conductive coating when measuring highly conductive materials when the NSMM is operating in contact mode . The NSMM is also capable of producing scanned images in other modes like in ﬁxed-distance mode, where the tip–sample
Corresponding author. Tel.: þ 86 10 62762999. Corresponding author. Tel.: þ 65 65162984. E-mail addresses: [email protected]
(S.-y. Xu), [email protected]
http://dx.doi.org/10.1016/j.ultramic.2014.09.008 0304-3991/& 2014 Elsevier B.V. All rights reserved.
separation is held constant during the scan, or in constant resonant frequency mode, where the tip–sample distance is regulated while the conductive samples are being scanned . Hybrid scanning techniques were explored with the NSMM, and combinations with the scanning tunneling microscope (STM)/ NSMM [6,7] and atomic force microscope (AFM)/NSMM [8,9] were also developed to exploit the tomographic capability of both non-biological  and biological materials [11–13]. Specialized tips like the GaN nanowire probe  were also developed for the NSMM, with the aim of improving the spatial resolution of imaging. The performance of NSMM depends heavily on the material and geometry of the tip. The probes in NSMM are usually hard and both the sample surface and the probe may inevitably be damaged after long periods of scanning. In scanning microscopy, carbon nanotubes have been manufactured as a probe due to their extraordinary thermal conductivity, mechanical and electrical properties [15,16]. Experimental results and simulations demonstrate that multi-walled carbon nanotube (MWCNT) bundles which are used as tips (referred to as ‘CNT tips’) are nearly independent on frequency within the microwave frequency range . Although commercial CNT tips are available for use as an atomic force microscopy (AFM) cantilever probe, the use of the CNT tip in the AFM is still not popular. For the NSMM, the use of the CNT is even rarer.
Z. Wu et al. / Ultramicroscopy 148 (2015) 75–80
In this article, we describe the use of a MWCNT bundle as a tip in a NSMM of our own design. The distribution of the resonant frequency fr and the S11 amplitude of a scanned area were plotted for both the contact mode and tapping mode. The use of the CNT tip on both highly-conducting and dielectric materials were also investigated, and clear images of gold numbers, photoresist stripes and corneal endothelial cells (cell line B4G12) were obtained with CNT tip with the NSMM operating in tapping mode. The CNT tip does not only improve the image quality, but as a ﬂexible probe, it also protects the sample surface from being scratched. The CNT tip in NSMM is also much more durable as compared to metallic probes (e.g. Tungsten probes) and based on our initial experience, the CNT tip (in tapping mode) can repeatedly tap and make contact with the sample for at least 106 times without being damaged. Thus, the use of CNT tips in the NSMM has great potential, especially when imaging soft material.
2. Experimental setup and details Fig. 1(a) shows a near-ﬁeld scanning microwave microscope (NSMM) system of our own design [18,19]. It consists of a high quality factor (Q) λ/4 coaxial resonator , a scanning platform attached to x–y direction motors, a piezoelectric motor, a vector network analyzer (VNA) (Agilent Technologies N5230A), and some lenses mounted to a charge-coupled device (CCD). The resonator is placed in an aluminum holder and is attached to the piezoelectric motor which controls the z-direction, while the sample is placed on the top of the scanning platform which is controlled by the x–y direction motors. Side-view images of the tip and sample are
captured in real-time by the lens and CCD, which are displayed on a LCD monitor. The VNA measures the amplitude of the reﬂection coefﬁcient S11 (in dB) as a function of frequency and displays these values on a graph. The maximum value of the S11 amplitude and the resonant frequency fr are both determined at the maximum of the resonance peak. A program written in LabVIEW is used to control the system and to collect data from the VNA. A tungsten tip with a CNT bundle is mounted on the central conductor of the λ/4 coaxial resonator. The schematic diagram of the CNT tip in the NSMM during scan is shown in Fig. 1(b), with the CNT tip bending as it makes contact with the sample surface. The tungsten tip is fabricated by anode corrosion with 1 M NaOH and the diameter of the tip is approximately 50 mm. A MWCNT bundle (average diameter is 5 mm) is adhered to the tungsten tip using silver paste, and the CNT tip is covered with cyanoacrylate (Loctite 460, Henkel Technologies, εr ¼ 2.75), which helps to bundle the individual CNTs together and to reduce mechanical damage to the CNTs when the scanning is being performed by the NSMM. The cyanoacrylate layer also helps to stabilize the microwave signal when the CNT tip is swept across a dielectric and conductive boundary of the sample during contact mode. The tip can be adjusted vertically in the z-direction by using the piezoelectric motor, while the sample can move horizontally in the x–y direction by controlling the x–y motors. Fig. 1(c) is the equivalent lumped series resonant circuit that has different electrical components representing different parts of the NSMM. The interaction between the CNT tip and the sample is represented by an effective resistance Rtip and effective capacitance Ctip, which may vary as the tip-sample distance is changed. Rsample is the effective resistance of the sample, while the substrate has an effective resistance Rsub and effective capacitance Csub. Any change within the NSMM such as moving to a different measuring point or varying the tip sample distance would cause a response to be generated by the system, which can be seen as a change in the effective resistance or capacitance of a component within the electrical circuit.
3. Results and discussions
Fig. 1. (a) A photograph of the NSMM setup. The NSMM consists of (1) a λ/4 resonator and its holder, (2) a vector network analyzer (VNA), (3) a scanning platform attached to the x–y direction motors, (4) a piezoelectric motor, (5) a light source, and (6,7) a lens attached to a CCD camera. (b) Schematic diagram of the NSMM probe with an attached CNT tip, which bends during scanning when in contact with the sample. The CNT bundle is attached to a tungsten tip and covered with non-conductive glue, while the tungsten tip with the CNT is mounted on the central conductor of the λ/4 coaxial resonator. The tip can be controlled vertically in the z-direction by the piezoelectric motor and the sample on the scanning platform can be moved horizontally with the x–y motors. (c) The equivalent lumped series resonant circuit of the NSMM. Rtip and Ctip are the effective resistance and effective capacitance associated with the tip-sample interaction respectively, Rsample is the effective resistance of the sample, while Rsub and Csub are the effective resistance and effective capacitance of the sample substrate respectively.
The vertical position of the CNT tip can be controlled by changing the input voltage of the piezoelectric motor as seen in Fig. 1(a). Fig. 2(a) and (b) shows the measured resonant frequency fr and the amplitude of the reﬂection coefﬁcient S11 respectively as a function of the tip-sample distance for a glass plate, photoresist ﬁlm (Allresist AR-P 5350) on a glass plate, a silicon wafer and an aluminum plate. The zero point of the tip position is deﬁned to be at the inﬂection of the resonant frequency and S11 amplitude curves and these points are marked with a vertical dashed line as shown in Fig. 2. At the zero point, the CNT tip makes contact with the sample beneath it. The resonant frequency and S11 amplitude changes as the tip-sample distance increases. For a positive tip position as shown in Fig. 2, the tip is not in contact with the sample. Thus, the tip position is related to the distance between the tip and the sample (i.e. tip-sample distance, d). The resonant frequency fr increases while the S11 amplitude decreases with increasing tip-sample distances for all the samples shown in Fig. 2. As the tip-sample distance increases, the tip-sample capacitance Ctip (in Fig. 1(c)) decreases (because C tip p 1=d), which decreases pﬃﬃﬃﬃﬃ ﬃ the total capacitance of the system. Since f p 1= LC , the resonant frequency fr will therefore increase with an increasing tip-sample distance. Meanwhile, the tip-sample distance directly affects the amount of power reﬂected back into the tip and hence the S11 amplitude. The S11 amplitude is at maximum at 0 dB and indicates total reﬂection of the incoming signal towards the tip. As the tip leaves the sample (i.e. increasing d), more power will be radiated into
Z. Wu et al. / Ultramicroscopy 148 (2015) 75–80
Fig. 2. The dependence of (a) the resonant frequency fr and (b) the S11 amplitude on the tip-sample distance respectively for a glass plate, a photoresist ﬁlm on a glass plate, a silicon wafer and an aluminum plate. The CNT tip position is controlled by changing the input voltage of the piezoelectric motor. The zero of the tip position is deﬁned at the point of inﬂection on the fr and S11 amplitude curves and are marked with a dashed line. The tip position is positive (i.e. tipsample distance) when the tip does not make contact with the sample, while a negative tip position indicates contact with the sample and is termed ‘reduced distance’.
free space as a consequence of the inverse square law, causing the S11 amplitude to drop to a higher negative value. Since the tip makes contact with the sample at the zero point, a negative tip position shown in Fig. 2 indicates that the CNT tip is bending and this shall be referred to as the ‘reduced distance’ of the CNT tip. Once the tip makes contact with the sample, the resonant frequency fr and S11 amplitude remain relatively constant, except for small ﬂuctuations which may be caused by an increase in surface area between the CNT and the sample as the ‘reduced distance’ increases. For a ﬁxed tip position, a comparison of the resonant frequency and S11 amplitude curves of the different materials presented in Fig. 2 would be useful in identifying the material composition of a given scanned image. For example, glass has a higher resonant frequency fr and lower S11 amplitude as compared to photoresist. Hence, this result allows the glass and photoresist to be identiﬁed from both the resonant frequency and S11 plots which are generated by the NSMM. The CNT tip as shown in Fig. 1(b) is ﬂexible and durable, which allows our NSMM system to operate in two different scanning modes: contact mode and tapping mode. In contact mode, the CNT tip position makes contact with the sample at a ﬁxed position, which is controlled by the piezoelectric motor. The CNT tip would then be swept across the sample like a broom during the scan. In tapping mode, the CNT tip is driven up and down by the piezoelectric motor mounted in the holder which contains the resonator with the CNT tip. By ﬁne tuning the voltage of the piezoelectric
motor, it can ensure that the CNT tip makes contact with the sample to obtain better images. This also helps to prevent the sample surface from being scratched by the CNT tip. Fig. 3 show images of a gold-patterned number ‘02’ collected by the NSMM while operating in both contact mode and tapping mode. The images obtained in contact mode are shown in Fig. 3(a) and (d) with plots of the fr and S11 values respectively. Although the number ‘02’ can be clearly seen in both images, scan lines and ghosting of the images are also observed. The scan lines may be due to the CNT tip picking up dirt particles as it sweeps across the image, while the ghosting and distortion of the image may be caused by the deformation of the CNT tip. Thus, it may not be preferable to have the CNT tip make contact and be dragged across the sample surface during the scanning process. In contrast, Fig. 3(b) and (e) presents the same image of the gold-patterned number ‘02’ obtained by the NSMM in tapping mode for the fr and S11 values respectively. These images are relatively clearer and have no draggled lines which make the tapping mode the preferred method of scanning. Also, in the optical image of Fig. 3(e), a defect is observed in the number ‘0’. This detail is also clearly observed in the NSMM tapping mode images of Fig. 3(b) and (e). Both the lower fr and higher S11 values of the defect in Fig. 3(b) and (e) respectively indicate that the defect is protruding outwards from the gold-patterned number surface (i.e. the defect is higher than all other parts of the sample). This was deduced from the data in Fig. 2 as a lower fr and a higher S11 value indicate a smaller tip-sample distance. To demonstrate the non-invasive operation of the NSMM with a CNT tip, the NSMM was made to scan stripes of photoresist in tapping mode. A poor image quality may result if the CNT tip is able to scratch the surface of the photoresist and damage it. However, Fig. 4 show clear images of photoresist stripes on a glass plate when the NSMM is operating in tapping mode with a CNT tip. The stripes of photoresist were prepared by photolithography with a width of 10 mm (periodicity of 20 mm) and a thickness of around 1.2 mm. The height difference of the photoresist stripes affects the resonant frequency fr and S11 amplitude during the NSMM measurement. As seen in the Fig. 2 data, the photoresist on glass has a lower resonant frequency fr and a higher S11 amplitude as compared to a bare glass plate. Thus, in Fig. 4(a), the lower resonant frequency corresponds to a photoresist surface, while a higher resonant frequency corresponds to the exposed glass plate which is located between the gaps of the photoresist stripes. Similarly, in Fig. 4(b), a higher S11 amplitude corresponds to a photoresist surface, while a lower S11 amplitude corresponds to the glass plate. Fig. 4(c) is a line scan of the resonant frequency fr along the black line shown in Fig. 4(a), while Fig. 4(d) is the line scan of the S11 amplitude along the black line in Fig. 4(b). The step size of the line scan is 0.24 mm for both Fig. 4(c) and (d). It can be observed from the respective resonant frequency fr and S11 amplitude curves of Fig. 4(c) and (d) that the NSMM can resolve stripes that are 10 mm wide and are spaced 20 mm apart. A clear difference of 0.3 MHz in the resonant frequency fr and about 0.1 dB in the S11 amplitude is observed for the stripes of photoresist and the glass plate in Fig. 4(c) and (d) respectively. From the measurement of the fuzzy fringes in Fig. 4(c) and (d), the lateral resolution of the NSMM with a CNT tip and operating in tapping mode is found to be about 2.5 mm. The NSMM is also capable of imaging biological soft tissue, which we have done for corneal endothelial cells (cell line B4G12) . To prepare the cells for imaging, the glass cover slips that were used to hold the cells were treated with cleansed with plasma for three minutes, sterilized and left overnight with a coat of FNC Coating Mixs to improve cell adhesion. The B4G12 corneal endothelial cells were then seeded with a seeding density of 6000 cells/cm2 and cultured for three days in Dulbecco's Modiﬁed
Z. Wu et al. / Ultramicroscopy 148 (2015) 75–80
Fig. 3. NSMM images of a gold-patterned number ‘02’ on a glass plate with the CNT tip in different scanning modes. The image of the resonant frequency fr when the NSMM is operating in (a) contact mode and (b) tapping mode. (c) A binarized optical image of the gold-patterned number ‘02’. The image of the S11 amplitude when the NSMM is operating in (d) contact mode and (e) tapping mode. All the arrows point to the position of the defect in the gold-patterned number ‘0’ as shown in (c).
Fig. 4. NSMM image of photoresist stripes on a glass plate. The stripe width is 10 mm and has a periodicity of 20 mm. The images shown are for (a) the resonant frequency fr and (b) the S11 amplitude when the NSMM is in tapping mode. (c) The line scan image for the resonant frequency fr along the black line shown in (a). (d) The line scan image for the S11 amplitude along the black line shown in (b). The step size for the line scans in (c) and (d) is 0.24 mm and the lateral resolution is around 2.5 mm.
Eagle's Medium (Sigma Aldrich) supplemented with 10% fetal bovine serum (Gibco, Invitrogen) and 1% Penicillin Streptomycin (Life Technologies). After three days, the cells were ﬁxed with
4% paraformaldehyde. The cover slips with the ﬁxed cells were then taken out of the paraformaldehyde solution and dehydrated naturally in ambient atmosphere for one day.
Z. Wu et al. / Ultramicroscopy 148 (2015) 75–80
Fig. 5 shows the images of endothelial cells (B4G12) that were prepared using the method as described above. Fig. 5(a) shows an optical image of ﬁxed B4G12 cells with an average area of about 600 mm2 , while Fig. 5(b) shows a phase contrast image of a pair of adjacent B4G12 cells which appears ﬂat and tightly bound to the cover slip. It can be observed that each B4G12 cell has a few protrusions along the cell border and while these B4G12 cells are
connected. This pair of adjacent cells will be imaged using the NSMM. Fig. 5(c) and (d) shows the three-dimensional plots of resonant frequency fr and S11 amplitude while Fig. 5(e) and (f) show the contour plots of resonant frequency fr and S11 amplitude for the pair of cells shown in Fig. 5(b). Fig. 5(c)–(f) were obtained by operating the NSMM with a CNT tip in tapping mode. In Fig. 5(c) and (e) [Fig. 5(d) and (f)] of the NSMM images, the red [blue] elliptic
Fig. 5. (a) The optical image of corneal endothelial cells (B4G12) on a glass cover slip. (b) The phase contrast image of a pair of adjacent B4G12 cells. The same pair of cells in are imaged with the NSMM operating in tapping mode with a CNT tip, and are shown in the following plots: (c) resonant frequency fr 3D plot, (d) S11 amplitude 3D plot, (e) resonant frequency fr contour plot, and (f) S11 amplitude contour plot. The red areas in the NSMM image corresponds to the nuclei, the yellow and green areas are the contour of cells, and the dark blue areas correspond to a thin layer of cytoplasm of the B4G12 cell. Red arrows in (e) and (f) point to the nuclei of cells, while black arrows point to one of the long lobes in the cell that is visible and distinguishable. (For interpretation of the references to color in this ﬁgure legend, the reader is referred to the web version of this article.)
Z. Wu et al. / Ultramicroscopy 148 (2015) 75–80
areas correspond to two nuclei of the cells, and they are clearly indicated by red arrows in Fig. 5(e) [Fig. 5(f)]. The yellow and green [light blue] areas correspond to the cytoplasm of endothelial cells of B4G12, which are similar to the contours that appear in the optical image of Fig. 5(b). The lobes of the pair of B4G12 cells can also be identiﬁed from the NSMM images and the longest lobe is marked with black arrows in Fig. 5(e) and (f). Finally, the dark blue [dark red] area in the NSMM images of Fig. 5(c) and (e) [Fig. 5(d) and (f)] correspond to a thin layer of cytoplasm within the B4G12 cells. It may be difﬁcult to distinguish between the cytoplasm and the medium surrounding the cell in the optical image of Fig. 5(b), but it can clearly be distinguished in the NSMM images of Fig. 5(c)–(f). 4. Conclusions MWCNTs were used as tips in a near-ﬁeld scanning microwave microscope (NSMM) of our own design, which were capable of imaging conductive and non-conductive samples in both contact mode or tapping mode. The CNT tip helps to improve the image quality as well as to protect the sample surface from being scratched. Clear images of gold-patterned numbers, photoresist stripes and corneal endothelial cells (B4G12) were obtained by mapping the resonant frequency fr and S11 amplitude of a given scan area by operating the NSMM with the CNT tip in tapping mode. In the imaging of the B4G12 cells, the nuclei and cytoplasm can be clearly distinguished from the rest of the cell and its surrounding medium, which shows the potential of the NSMM in cellular imaging. This work may shed light for further developments and applications of the NSMM technique in imaging soft material in the biological sciences.
Acknowledgment This work was ﬁnancially support by the SERC (Grant no 1021450119), MOST of China (Grant no 2011DFA51450) and NSF of China (Grants 11374016). We would like to thank Prof. Evelyn Yim, Dr. Anh Tuan Nguyen and Mr. Rizwan Muhammad for providing the cells and phase contract images. References  T. Wei, X.D. Xiang, W.G. Wallace-Freedman, P.G. Schultz, Scanning tip microwave near-ﬁeld microscope, Appl. Phys. Lett. 68 (1996) 3506–3508; C. Gao, T. Wei, F. Duewer, Y.L. Lu, X.D. Xiang, Appl. Phys. Lett. 71 (1997) 1872–1874.
 C. Gao, X.D. Xiang, Quantitative microwave near-ﬁeld microscopy of dielectric properties, Rev. Sci. Instrum. 69 (1998) 3846–3851.  C. Gao, T. Wei, F. Duewer, Y.L. Lu, X.D. Xiang, High spatial resolution quantitative microwave impedance microscopy by a scanning tip microwave near-ﬁeld microscope, Appl. Phys. Lett. 71 (1997) 1872–1874.  Z.Y. Wang, M.A. Kelly, Z.X. Shen, L. Shao, W.K. Chu, H. Edwards, Quantitative measurement of sheet resistance by evanescent microwave probe, Appl. Phys. Lett. 86 (2005) 153118.  F. Duewer, C. Gao, I. Takeuchi, X.D. Xiang, Tip-sample distance feedback control in a scanning evanescent microwave microscope, Appl. Phys. Lett. 74 (1999) 2696–2698.  A. Imtiaz, S.M. Anlage, A novel STM-assisted microwave microscope with capacitance and loss imaging capability, Ultramicroscopy 94 (2003) 209–216.  T. Machida, M.B. Gaifullin, S. Ooi, T. Kato, H. Sakata, K. Hirata, Development of near-ﬁeld microwave microscope with the functionality of scanning tunneling spectroscopy, Jpn. J. Appl. Phys. 49 (2010) 116701.  J.C. Weber, J.B. Schlager, N.A. Sanford, A. Imtiaz, T.M. Wallis, L.M. Mansﬁeld, K.J. Coakley, K.A. Bertness, P. Kabos, V.M. Bright, A near-ﬁeld scanning microwave microscope for characterization of inhomogeneous photovoltaics, Rev. Sci. Instrum. 83 (2012) 083702.  M. Farina, A. Di Donato, T. Monti, T. Pietrangelo, T. Da Ros, A. Turco, G. Venanzoni, A. Morini, Tomographic effects of near-ﬁeld microwave microscopy in the investigation of muscle cells interacting with multi-walled carbon nanotubes, Appl. Phys. Lett. 101 (2012) 203101.  V.V. Talanov, C. Del Barga, L. Wickey, I. Kalichava, E. Gonzales, E.A. Shaner, A.V. Gin, N.G. Kalugin, Few-layer graphene characterization by near-ﬁeld scanning microwave microscopy, Acs Nano 4 (2010) 3831–3838.  J. Park, S. Hyun, A. Kim, T. Kim, K. Char, Observation of biological samples using a scanning microwave microscope, Ultramicroscopy 102 (2005) 101–106.  Y.J. Oh, H.-P. Huber, M. Hochleitner, M. Duman, B. Bozna, M. Kastner, F. Kienberger, P. Hinterdorfer, High-frequency electromagnetic dynamics properties of THP1 cells using scanning microwave microscopy, Ultramicroscopy 111 (2011) 1625–1629.  K. Lee, A. Babajanyan, H. Melikyan, C. Kim, S. Kim, J. Kim, J.H. Lee, B. Friedman, R. Levicky, S. Kalachikov, Label-free DNA microarray bioassays using a near-ﬁeld scanning microwave microscope, Biosens. Bioelectron. 42 (2013) 326–331.  J.C. Weber, P.T. Blanchard, A.W. Sanders, A. Imtiaz, T.M. Wallis, K.J. Coakley, K.A. Bertness, P. Kabos, N.A. Sanford, V.M. Bright, Gallium nitride nanowire probe for near-ﬁeld scanning microwave microscopy, Appl. Phys. Lett. 104 (2014) 023113.  H.J. Dai, J.H. Hafner, A.G. Rinzler, D.T. Colbert, R.E. Smalley, Nanotubes as nanoprobes in scanning probe microscopy, Nature 384 (1996) 147–150.  N.R. Wilson, J.V. Macpherson, Carbon nanotube tips for atomic force microscopy, Nat. Nanotechnol. 4 (2009) 483–491.  Y. Yang, C.Y. Tan, W.Q. Sun, W. Li, C.K. Ong, Y. Liu, Y. Li, S.Y. Xu, High frequency resistance of single-walled and multiwalled carbon nanotubes, Appl. Phys. Lett. 98 (2011) 093107.  W. Sun, Y. Yang, Z. Wu, T. Feng, Q. Zhuang, L.-M. Peng, S. Xu, C.K. Ong, Penetrative imaging of sub-surface microstructures with a near-ﬁeld microwave microscope, J. Appl. Phys. 116 (2014) 044904.  Z. Wu, A.D. Souza, B. Peng, W.Q. Sun, S.Y. Xu, C.K. Ong, Measurement of high frequency conductivity of oxide-doped anti-ferromagnetic thin ﬁlm with a near-ﬁeld scanning microwave microscope, AIP Adv. 4 (2014) 047114.  M. Valtink, R. Gruschwitz, R.H.W. Funk, K. Engelmann, Two clonal cell lines of immortalized human corneal endothelial cells show either differentiated or precursor cell characteristics, Cells Tissues Organs 187 (2008) 286–294.  B.K.K. Teo, K.J. Goh, Z.J. Ng, S. Koo, E.K.F. Yim, Functional reconstruction of corneal endothelium using nanotopography for tissue-engineering applications, Acta Biomater. 8 (2012) 2941–2952.