exchange process, a glass substrate was placed in a mixed melt of AgNO 3 and KNO 3 at 400°C. The thickness of the glass substrate, time of the ion-exchange process, and weight concentration of AgNO 3 in the melt determined the concentration and distribution of Ag + ions in the glass. Thermal annealing of the ion-exchanged glass in a H 2 reduction atmosphere, typically at 400-450°C, resulted in the formation of spherical silver nanoparticles [14].Macroporous silicon with lattice constants of 500 nm or 2 lm were grown in a photoelectrochemical etching process of lithographically prestructured <100>-oriented n-type silicon wafers. In each case, the front side of the wafer was in contact with hydrofluoric acid (concentration c HF = 5 wt.-%; T = 10°C) whereas the backside was illuminated generating electron-hole pairs. An external anodic bias then consumed the electrons, and the electrons/holes diffused through the whole wafer to the silicon electrolyte interface, promoting the silicon dissolution there. The pores with very flat surfaces and high aspect ratios grew straight along the (001) direction of the silicon single crystal [15,16]. The macroporous silicon samples were then sputtered with 10 nm chromium film to avoid anodic bonding during the experiments [17,18]. because of their unique one-dimensional (1D) electronic properties and high tensile strength.[2] Nanotubes were first produced mostly by the arc discharge of carbon, the pyrolysis of hydrocarbons, and the laser ablation of graphite. The selective growth of aligned nanotubes was later developed, using metal-coated Si wafers as growth templates, which produce so called nanocarpets. [3] The intertube binding within carpets is relatively weak compared with the covalent character of the in-plane C-C bonds. Consequently, aligned nanotube films currently focus mainly on electronic field emission. [4] To date, the fabrication of aligned nanotube films into microelectromechanical systems (MEMS), such as suspended microbeams and -plates, remains a challenge, mainly because the intertube weakness makes them unable to carry a load in high-frequency bending maneuvers. If nanotube composites are to be made as MEMS devices, their vibration damping factor must be equivalent to that of metals. In this paper, we report on the successful development of a novel method for reinforcing intertube binding with a polymer: the nanotube composites were then fabricated into MEMS devices. Mechanical tests reveal an increase of the composite modulus by a factor of 20 compared with bare nanotubes. The damping and quality factors for nanotube MEMS devices are 12 and 0.042 at 1 atm (1 atm = 101.325 kPa), respectively, which are comparable with those of typical MEMS components. Aligned multiwalled nanotube (MWNT) films have been grown as microgripper structures (Fig. 1b), which are well defined and commonly seen in MEMS devices. These structures were established in step II (Fig. 1a). MWNT grippers were then filled by polymer in step III, and typical examples are COMMUNICATIONS
