Search

This paper describes two types of novel compact low-loss millimeter-wave filters using an overlay coplanar waveguide (OCPW) line for lowpass filters and an inverted overlay coplanar waveguide (IOCPW) line for bandpass filters with periodic defected ground structures (DGSs). The OCPW and IOCPW lines are utilized to reduce the conductor and the substrate dielectric loss of the transmission lines. Also, these lines can provide a wide impedance range by controlling the overlapped area. The DGSs implemented on the transmission line have the role of size reduction and harmonic suppression of the filters. The combination of transmission lines having suspended structures and DGSs can provide great potential in terms of compact size, band rejection property and low-loss characteristics. The proposed filters are fabricated by using a micromachining technology and their RF performances are measured and analyzed. One of the proposed filters is a lowpass filter with a five-section stepped impedance topology having a periodic high?low impedance section. In this filter, the low impedance section is composed of the OCPW line with dumbbell-shaped DGSs and its length can be reduced by 52.7% compared to the OCPW line without DGSs. The measured insertion loss of this filter is lower than 0.2 dB up to 9 GHz, and significant spurious bands are not observed up to 40 GHz, resulting in a wide stop band. The other work that is presented here deals with a pi-shaped bandpass filter. This filter consists of three high impedance sections and two low impedance sections, and the line length of the high impedance section can be reduced by using the IOCPW line with spiral-shaped DGSs. The measured center frequency of the fabricated filter is 19 GHz, and the passband loss is 2.3 dB at the measured center frequency. The use of DGSs on the transmission line also enables us to suppress effectively the third-order harmonic component at 57 GHz.

This paper reports two types of compact dual-band bandpass filters for wireless local area network (WLAN) applications operating at two frequency bands of 2.4 and 5.2 GHz. The two types of filters (contact-type and capacitive-type) are focused on their digital mode operations in order to alternatively select the operational frequency band of the WLAN. Different types of micromachined frequency-tuning elements having only two states are implemented on each filter to control the center frequencies of the filters. Two digitally operated WLAN filters are fabricated using a surface micromachining technology, which enables us to make simple planar structures of the filters, resulting in easy integration with other planar circuits. Besides, external switching devices for frequency selection are not required, since they are already implemented on the filter's own structures. One of the proposed filters is a contact-type filter using switched inductors with direct-contact MEMS switches. In this filter, a precise and stable frequency-tuning ratio can be obtained because the center frequency is changed by total inductance change due to ON/OFF actuations of the switch. The measured center frequencies of the filter were 2.5 and 5.1 GHz (49% tuning ratio), and the passband insertion loss and return loss were 4.7 dB and 21.5 dB at 2.5 GHz, and 5.2 dB and 21 dB at 5.1 GHz, respectively. The other is a low-loss capacitive-type filter using micromachined variable capacitors. The frequency tunability of this filter is controlled by discrete change of a capacitance according to simultaneous actuations of a total of 12 variable capacitors. This filter shows lower loss compared to the contact-type filter, since loss due to contact resistance of the switch is not included in this filter configuration. The initial center frequency was measured to be 5.4 GHz and it was shifted to 2.6 GHz (48% tuning ratio) with the applied voltage. The total insertion loss and return loss were 2.8 dB and 24.4 dB at 5.4 GHz and 3.3 dB and 20.1 dB at 2.6 GHz, respectively.

In this paper, a packaging method utilizing an LTCC (low temperature co-fired ceramic) substrate and a BCB (benzocyclobutene) adhesive layer has been developed for RF MEMS devices, and the RF performance and characteristic parameters of the package have been evaluated. LTCC substrates have good RF characteristics in high-frequency applications, and via feedthroughs can be easily incorporated during the manufacturing process. In this paper, an LTCC substrate is used as a capping wafer to reduce the complex processes for vertical interconnections. A layer of BCB, in the form of sealing rims, is used as an adhesive to bond the MEMS substrate with the LTCC cap due to the excellent properties of BCB as a packaging material. A CPW (coplanar waveguide) line has been fabricated on a quartz substrate and packaged to demonstrate the performance of the proposed packaging method. After forming the CPW lines, a 28 µm thick BCB layer is patterned by double-coating photolithography for an adhesive bonding. On the backside of the LTCC cap, a 150 µm deep cavity is formed to improve the RF characteristics. The CPW and the contact pad are connected electrically through the silver via-post in the LTCC substrate by screen-printed silver epoxy. The RF characteristics of the CPW line have been measured before and after packaging. The insertion loss of a bare CPW is 0.047 dB at 2 GHz and 0.092 dB at 20 GHz. After packaging, the insertion loss of the packaged CPW is 0.091 dB at 2 GHz and 0.312 dB at 20 GHz. A leak test has been performed using both IPA (isopropyl alcohol) soaking and the He leak tester. Most of the samples show no leakage for the IPA test, and a measured leak rate of 10−8 atm cc s−1 for the He leak test. In addition, the shear strength of the package was measured to be 25–35 MPa. From the experimental results, we showed the feasibility of a low-loss RF MEMS package from dc to 20 GHz with acceptable package performances.

This paper describes a single-aperture MEMS probe and a MEMS probe array for the measurement of biological properties. We designed and fabricated the single-aperture MEMS probe using surface micromachining and verified it by measuring the permittivity of a standard liquid before introducing the MEMS probe array. The actual aperture size of the single-aperture MEMS probe is only 390 µm × 80 µm, which is very small in comparison with the conventional laser-machined coaxial probe. In order to show the feasibility of the proposed single-aperture MEMS probe for permittivity measurements, we performed in vitro measurements of 0.9% saline. Once the single-aperture probe was verified, we proposed the concept of a probe array for biological measurements and experimentally showed the suitability of the MEMS probe array for biological applications through experiments using pork. The MEMS probe array consists of five microstrip feed lines, each of which is followed by open-ended strip lines, and the permittivity measurement of each port is separately performed through the use of a conventional multiport coaxial switch (Agilent, HP 87106 C), followed by a network analyzer (HP 8510 C). Through broadband measurements of 0.9% saline and pork using the MEMS probe array, we were able to discriminate the muscle and fat of pork through just one contact by placing the MEMS probe array on the boundary of muscle and fat. This newly proposed MEMS probe array has great potential in terms of disposability, low cost, integration with planar circuits and a short detection time for biological measurements.

We developed a silicon MEMS probe for permittivity measurements using an adhesive bonding process. Only two photolithographic masks are required to fabricate the probe, which can be implemented through simple bonding processes using silicon substrates and a benzo cyclo butene (BCB) adhesive layer. Undoped silicon substrates with thicknesses of 300 ?m are used as the dielectric layers of the proposed probe. BCB layers, which have good electrical properties at high frequencies as well as adhesive properties for the bonding process, play the role of bonding materials between the two silicon substrates. The length of the probe is 30 mm, and the aperture located at the tip of the probe is 1.1 mm ? 0.62 mm. The permittivity of 0.5% saline was measured, and the results agreed with the values obtained through the Cole?Cole equation. To validate the feasibility of this probe for practical biological applications, we also performed in vivo measurements of the muscle, skin and blood of mice. Due to the simple fabrication process, the cost of the probe can be reduced in comparison with the previous micromachined probe (Kim et al 2005 J. Micromech. Microeng. 15 543?50) as well as the conventional laser machined probe. Low cost leads to disposability, which is an important factor for practical biomedical applications; and thus, coupled with the probe's capabilities of MMIC integration and CMOS compatibility, this probe has excellent potential in the field of microwave permittivity measurements.

This paper reports the first demonstration of a single crystalline silicon(SCS)-based vertical gap-tuning MEMS variable capacitor with high mechanical reliability using the SiOG (silicon-on-glass) process. The proposed variable capacitor was fabricated by silicon-based micromachining technology, and its mechanical and electrical performances were tested. By adopting SCS, which has nearly zero stresses and superior thermal characteristics as a structural material, we can improve the manufacturability, reproducibility and mechanical reliability of the fabricated devices compared with a conventional metallic variable capacitor. The variable capacitor described in this paper is fabricated on a coplanar waveguide (CPW) transmission line and the top electrode in the driving part is incorporated with a SCS actuating membrane, resulting in a highly rigid structure without structural deformations. The designed SCS variable capacitor is actuated by an electrostatic force, and the total chip size is 1.05 mm × 0.72 mm. The measured pull-in voltage was 37 V, and the RF characteristics of the fabricated SCS variable capacitor from dc to 40 GHz were measured. With and without an applied dc bias voltage, the measured S11 and S21 were changed from −15.6 to −5.1 dB, and from −0.49 to −2.3 dB at 30 GHz, respectively. The measured capacitance at the up and the down states was 30 and 140 fF, respectively, which corresponds to the capacitance ratio of 4.67. The mechanical lifetime of the fabricated variable capacitor was also measured. RF performance degradation and stiction problems were not observed, even after 108 cycles of repeated actuations.

We implemented a micromachined probe for the measurement of biological properties using MEMS technology, and experimentally showed the suitability of the micromachined probe in biological applications. The micromachined probe was fabricated on a silicon substrate, and to remove wave transmission through the silicon substrate, we etched the silicon substrate from beneath a lower ground and made the etched silicon surface conducting by using thermal evaporation of Cr/Au and a coating of conductive epoxy. The micromachined probe consists of a CPW and strip line between benzo cyclo butene (BCB) layers, which is known to be a material with high resistivity, low loss tangent, and low permittivity at high frequency. We measured the permittivity of a number of well-known liquids—0.5%, 0.9% and 1.3% saline, acetone, ethanol, and muscle and fat of pork—as biological samples using the micromachined probe after liquid calibration. The measured permittivity of 0.9% saline agreed well with the expected value of the Cole–Cole equation. In this paper, we first demonstrate that the micromachined probe can provide broadband measurement of measurable solid materials, such as biological samples, and also of well-known liquids at microwave frequencies. The size of the micromachined probe is 2000 µm (width) × 580 µm (thickness) × 30 000 µm (length), and the aperture size of the micromachined probe is only 650 µm × 70 µm. Therefore, we can extract the biological information from very small biological tissues and reduce radiation effects. Thus we show the feasibility of low-cost, small and portable permittivity measurement systems using a micromachined open-ended coaxial RF MEMS probe.

We have experimented with using continuous self-assembled monolayer (SAM) coatings on different surfaces, such as gold and silicon nitride, as anti-stiction coatings for radio-frequency microelectromechanical systems (RF MEMS). SAMs are coated on gold and silicon nitride surfaces in order to fabricate stiction-free gold structures after the wet release process. SAM coatings on gold and silicon nitride surfaces change the contact angle of a water droplet from 70.6° to 112.6° and from 36.7° to 115.8°, respectively. From atomic force microscopy, we can see that there is little change in the roughness of the gold and silicon nitride before and after the SAM coating. Using the electroplating technique and the wet release process, we have fabricated 4 μm thick gold cantilevers and bridges on a gold surface with a 6.5 μm gap from the substrate. Cantilevers and bridges are fabricated with different lengths in the range of 100–1000 μm with separations of 100 μm. It is observed that cantilevers and bridges with a length of 1000 μm on a gold substrate are released and standing free. A 1000 μm long gold cantilever has only 16.5 μm tip end deflection after the wet release process and SAM coating, much smaller compared with the 120 μm tip end deflection with the dry release process. Gold cantilevers, 100–500 μm long, on silicon nitride are released free and gold bridges on silicon nitride are released with a length of 1000 μm with continuous SAMs. The SAMs are coated on gold and silicon nitride surface in order. This method enables the gold cantilever and bridge on the gold or silicon nitride surface to be released with low deflection.

This paper presents a theoretical and experimental strategy for thermal de-isolation of silicon microstructures during a plasma etching process. Heat sinking blocks and thin metal layers are implemented around a thermally isolated mass to avoid severe spring width losses by a steep temperature rise. Thermal de-isolation significantly reduces the fabrication errors from −51.0% to −9.0% and from −39.5% to −6.7% for spring widths and resonant frequencies, respectively. Thermal de-isolation also reduces the standard deviation of resonant frequencies from 8.7% to 1.5% across a wafer, which clearly demonstrates the proposed method.

A metal/insulator microtube with micromesh walls was constructed using stress-assisted self-rolling technology. The mesh-sidewall Pt/Ti/SiO2 microtube was self-formed by a tensile-stressed metal Pt/Ti film deposited onto a pre-patterned SiO2 micromesh layer. The microtube measured about 25 µm in diameter and was longer than 7 mm. The sidewall of the microtube was a square mesh, 5–20 µm long, and was electrically connected to electrical pads for electrical conductance measurement. The electrical resistance of the rolled-up microtube was measured to be 250–350 Ω when the microtube resistor's length was around 540 µm. The real-time measurement of the conductance change of the microtube with a Pt resistor could monitor the temperature change generated by heat injection. The microtube with micromesh walls is expected to be an interesting structure that has promising potential for use in electronics, chemical and biological applications.

A pattern width loss during the fabrication of lateral silicon resonators degrades resonant frequency reliability since such a width loss causes the significant deviation of spring stiffness. Here we present a design guide for etch holes to obtain reliable resonant frequencies by controlling etch holes geometries. The new function of an etch hole is to generate the comparable amount of the width loss between springs and etch holes, in turn to minimize the effect of the spring width loss on resonant frequency shift and deviation. An analytic expression reveals that a compensation factor (CF), defined by the circumference (Cu) of a unit etch hole divided by its silicon area (Au), is a key parameter for reliable frequencies. The protrusive etch holes were proposed and compared with square etch holes to demonstrate the frequency reliability according to CF values and etch hole shapes. The normalized resonant frequency shift and deviation of the protrusive etch hole (?13.0% ? 6.9%) were significantly improved compared to those of a square etch hole with a small CF value (?42.8% ? 14.8%). The proposed design guide based on the CF value and protrusive shapes can be used to achieve reliable resonant frequencies for high performance silicon resonators.

We firstly propose the concept of a frequency and bandwidth fine-tuning method using an RF MEMS-based suspended tunable band-stop resonator. We experimentally show the feasibility of the continuously tuned resonator, including a second-order filter, which consists of cascaded resonators to achieve center frequency and bandwidth fine tuning. The structure consists of a freestanding half-wavelength (λ/2) resonator connected to a large displacement comb actuator. The lateral movement of the λ/2 resonator over the main transmission line produces different electromagnetic decoupling values from the main transmission line. The decoupled energy leads to continuous center frequency and bandwidth tuning using the band-stop resonator circuit for fine-tuning applications. The freestanding λ/2 resonator plays the role of a variable capacitor as well as a decoupling resonator in the proposed structure. The fabricated tunable filter shows suitability for Ku-band wireless communication system applications with continuous reconfiguration.

This paper reports a hybrid RF MEMS probe array system for permittivity measurements. A single-pole triple-throw (SP3T) RF MEMS silicon switch and a novel surface micromachined transmission line with three planar apertures were designed, fabricated and measured. These MEMS devices were connected using wire bonding and bond wire effects were cancelled as a consequence of in-liquid calibration of the whole system. A permittivity measurement of 0.9% saline was made by operating each RF MEMS silicon switch. The measured result showed good agreement with the reference value from the Cole?Cole equation up to 20 GHz. This result shows the feasibility of the proposed hybrid RF MEMS probe array system for permittivity measurements.