RF characterization of a single wall carbon nanotube bundle

Abstract

Summary form given only. As carbon nanotubes (CNTs) are becoming the most promising material for nanoelectronic devices, interests on their high-frequency properties are being further increased. Recently, many active researches characterizing single-wall or multi-wall CNTs have been reported. Here, we fabricate the device with a bundle of single-wall CNT (SWCNT) captured between two signal electrodes of a coplanar waveguide (CPW), and report its radio-frequency (RF) characterization and equivalent circuit modeling. The article shows an SEM image of the SWCNT bundle captured between two signal electrodes of the CPW with the gap of 700 nm. First of all, the CPW for GSG measurement was fabricated on the high resistivity Si wafer by photolithography and lift-off process. Further electron beam lithography made the CPW has sharp signal electrodes to alleviate the drastic impedance mismatching with the SWCNT and to minimize the parasitic capacitance between two signal electrodes. The bundle of SWCNTs was captured between two signal electrodes of the CPW by dielectrophoresis alignment. Then, we made the ohmic contact between the SWCNT and the CPW by using Au electroplating and subsequent thermal annealing. This electroplating process does not require one more step of lithography. The article shows the measured transmission (S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">21</sub> ) and reflection (S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">11</sub> ) characteristics of the CPWs with/without SWCNT at frequencies of 0.1 ~ 40 GHz. The transmission of the CPW with SWCNT is 1 dB larger than that of the CPW without SWCNT at 10 GHz. The difference denotes the amount of the transmission through SWCNT, and it is added to the transmission through the parasitic capacitance between two signal electrodes of the open CPW. The reflection of the CPW with SWCNT is maximum 1.7 dB smaller than that of the CPW without SWCNT at 38 GHz. Figure 3 shows the equivalent circuit model of the CPW combined with SWCNT. The parasitic parameter values of the open CPW were extracted from the measurement data of the CPW without SWCNT by ADS optimization. The parasitic values are follows; L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> = 0.006 nH, L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> = 0.005 nH, R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> , = 8 Omega, R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> = 10.7 Omega, C <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> = 0.04 pF, C <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> = 0.05 pF, and C <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> = 0.07 fF. Then, we extracted the other parameter values (due to SWCNT) from the measurement data of the CPW with SWCNT, keeping the parasitic values of the open CPW. The values are follows; R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">c</sub> = 4.1 kOmega, C <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">el</sub> = 0.9 fF, R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">CNT</sub> = 1.04 kOmega, and L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">k</sub> = 1.2 nH. We completed the total equivalent circuit model of the CPW with SWCNT by using ADS optimization since the de-embedding of CPW pads may result in overestimation due to contact resistance. The article shows the magnitude and phase of the impedance (Z) obtained from the measurement and from the equivalent circuit of the CPW with SWCNT. The measured data are consistent with the modeled data within a reasonable accuracy. In summary, we have captured SWCNT between two signal electrodes of the CPW and presented its high-frequency characterization. From the de-embedding process using the equivalent circuit, we successfully obtain the resistance (R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">CNT</sub> = 1 -04 kOmega) and the inductance (L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">k</sub> = 1.2 nH) of the SWCNT bundle.