Tuneable filters based on dielectric resonators
We present a concept of the fast tuneable dielectric resonator for applications in mobile communication systems. The tuning approach is based on the interaction between two resonant modes: a cylindrical dielectric resonator of TE01d mode and a planar microwave slotline resonator. Variation of the coupling between two resonator modes leads to change of the rf electro-magnetic field distribution in the resonator assembly and as the consequence to variation of the resonance frequency of the resonator. In order to control the intermodal coupling we apply fast microelectromechanical (MEMS) or piezoelectric switches. Based on this concept discrete tuning by 5 MHz in 0.25 MHz frequency steps was demonstrated for a test resonator at 2 GHz. The unloaded quality factor is about 10.000 and the measured switching time is about one millisecond. Within our tuning approach and resonator design a substrate with 4 radial slotlines is symmetrically arranged above a cylindrical dielectric resonator of TE01d mode. One end of each slotline is short circuited while another end is loaded by a MEMS switch. For the case of a switch off-state, each of the 4 slotlines represent a capacitively loaded quarter wave slotline resonator which possesses a strong magnetic field component in radial direction providing strong coupling to the resonant fields of the TE01d mode. In this case, due to interaction between the dielectric resonator mode and the slotline resonator mode the stored electromagnetic energy is distributed between the DR and the slotlines. This leads to an increase of the effective size of the resonator and, as a consequence, to a decrease of the resonant frequency of the TE01d mode. Upon closing one switch, the corresponding quarter wave resonator changes into a halfwave resonator. Due to the sign change of magnetic field in the middle of the slotline, the coupling between slotline mode and TE01d mode is much weaker. In this case there is almost no intermodal coupling and the resonance frequency of the TE01d mode remains almost unaffected by the presence of the slotlines. In order to evaluate the expected tuning range and to choose an optimum slotline configuration, numerical field simulations have been performed employing the commercial code “CST Microwave Studio”. For our resonator design the spatial distribution of resonant fields were calculated for the capacitive (open) and resistive (closed) state of the MEMS for the case of 4 symmetrically arranged slotlines. As expected, the field distribution differs substantially. There is an obvious excitation of the slotlines with open end. For the resistive state, the field is localized primarily inside the dielectric resonator, therefore the effects of slotlines and MEMS are not significant and frequency is about that of the non-perturbed dielectric resonator. Variation of the frequency determines the tuning range, which is about 3% (~60MHz) at quite high quality factors up to 7000. The principle of tuning was proven by an experimental setup based on commercial piezoelectric bimorph actuators. Our setup is based on a cylindrical dielectric resonator (er = 28 - BZT ceramic) with TE01d mode at about f = 1.86 GHz is arranged in a copper cavity. A 0.5 mm thick copper disc with radial slots was assembled symmetrically above the dielectric resonator. As MEMS equivalents, four metallized bimorph piezoelectric actuators control the slotlines states and consequently, the coupling between the dielectric resonator and the slotlines. A large tuning of the resonance frequency is observed with a very challenging values for the figure-of-merit for tuneable resonators being about 160. One of the interesting results, which we found, is that there is no significant intermodal coupling between different slotline resonator modes. Therefore, each of the slotlines can be considered as an independent tuning element. This behaviour can be used as digital frequency tuning. To demonstrate the principle of the digital tuning within a frequency range of 5 MHz in 0.25MHz frequency steps we applied 8 bulk slotlines for a test resonator at 2GHz. The unloaded quality factor was found to be between 10.000 and 12.000 depending on the switching state of this digital tunable resonator. It was shown that a discrete piezoactuator based MEMS tuning concept for dielectric resonators is quite challenging and has a great potential to beat the performance of semiconductor based tuning techniques. Optimisation of the planar resonator structures in order to reach larger tuneability while keeping the highest quality factor is subject of our future activities.
The basic idea for the detector rests on a system of two coupled microwave resonators, one of which acts as a stable reference frequency but the other of which has a strong temperature dependent frequency (in the present prototype this is made of SrTiO3 (STO), which has a very strong temperature dependence of its permittivity. A key property of this resonator is that if the correct mode is chosen the frequency shift of the coupled mode with temperature does not depend on the volume of the temperature dependent resonator, at least to first order. Thus when energy is absorbed by the STO resonator its temperature changes, resulting in a frequency shift in the coupled resonance. Since extremely small frequency shifts are relatively easily detected this temperature sensor can have nK sensitivity. It also possesses other potential advantages over other sensors types. It is essentially non-contacting, in that interrogation of the temperature is done by the microwave field itself, no cables are required to contact the sensor. Secondly the sensor is an insulator. Its thermal capacity is due only to phonon contributions and there is no electronic contribution. At temperatures below 1K the electronic contribution to the specific heat of metals far outweighs the lattice contribution. Thus an insulating sensor of a particular mass can have a much smaller heat capacity than a metal of similar mass at the same temperature so it may have higher resolution than a metallic one or could be operated at a higher temperature with the same resolution.
In many cases, a further advance in the electronics industry is closely related to a reduction in the processing temperatures of functional materials, e.g., the production of a base-metal electrode capacitor, the reduction of Pb losses in piezo-materials, the integration of passive elements in LTCC modules, etc. Because of the lack of fundamental knowledge about low-temperature sintering mechanisms researchers are forced to apply specific empirical principles for each particular material. Only a few attempts to explain the basic mechanisms of low-temperature sintering have been published so far, and no general principles have been described. Here we explain the fundamental elements of a low-temperature sintering mechanism, called reactive-liquid phase sintering, and show that if a few general conditions are ensured then almost any powder can be sintered at temperatures as much as 400°C lower than its initial sintering temperature. We performed the sintering experiments involving BaTiO3 on a variety of different BaTiO3 powders, whose original sintering temperature is from 1250 to 1300°C. A small addition of 0.3wt% Li2O was used as a sintering aid in the form of an either polycrystalline Li2O, Li2CO3 or an acetic solution of Li+ ions. Microstructural investigations of the sintered bodies showed the presence of a small concentration of two secondary phases: Li2TiO3 and Ba2TiO4. A high-resolution TEM was used to check the grain boundaries and the triple points for the presence of the liquid-phase residuals; however, no such evidence was found. To understand this unusual sintering behavior a detailed investigation of the reaction mechanism of the sintering was performed. Based on these experimental results we have developed a general explanation for the low-temperature sintering mechanism, called here reactive liquid-phase sintering. The essential element of reactive liquid-phase sintering is the presence of a low-temperature liquid phase that must be able to directly or indirectly accelerate a reaction with the matrix phase. If we assume the same thermal conditions and the same reaction-limiting process, which is the rate of diffusion through the solid reaction layer then the rate of the reaction depends only on the surface contact areas, which in the case of the reaction between a liquid and a solid is significantly larger. So, the rate of the reaction will be significantly increased when one of the reagents melts. The next important element of reactive liquid-phase sintering is the nature of the reaction with the matrix phase. The reaction must enhance at least one of three mass-transport processes, which are dominant in such a system during sintering. The most direct way is to increase the volume-diffusion coefficients. The volume-diffusion coefficients are proportional to the vacancy concentration; as a result, the sintering rate will increase when the structural vacancies are generated in the matrix phase. The process called a liquid-phase (-assisted) sintering, where the mass transport goes through the liquid phase by a solution-precipitation method, also promotes the sintering. This process can be further accelerated by the increase in the solubility of the matrix phase during or after the reaction. Finally, if during the reaction with the matrix phase a temporary or permanent amorphization occurs a viscous flow from the grain surface to the necks between the grains contributes to the sintering. In the second stage of our investigation we applied the reactive liquid-phase sintering mechanism to successfully sinter a number of powders with very different chemistries. The sintering aids were selected according to the crystal-chemistry of the matrix phase in such a way that they would trigger all the mechanisms required for the reactive liquid-phase sintering. Certainly, the addition of sintering aids has an influence on the physical properties of ceramics. However, the concentration of the sintering aid is small and in many cases it is entirely incorporated into the crystal lattice of the matrix. With the proper selection of sintering aid we can minimize the influence on the particular physical properties. An example is the commercial X7R capacitor formulation, which produces its desired dielectric properties as a result of an inhomogeneous core-shell microstructure. By applying the same method as for the SrTiO3 we managed to reduce the sintering temperature from 1090 to 900°C. Due to the low processing temperature the inhomogeneity was well preserved and consequently the obtained dielectric properties are even better than usual. In addition, due to such a low sintering temperature the expensive palladium can be eliminated from the electrodes of the X7R multilayer capacitors, which makes the production significantly cheaper.
Two magnetically tunable dielectric resonator filters operating at frequency about 2 GHz have been elaborated and constructed. First of one was made on two dielectric resonators tuning by means of a axially magnetized ferrites disks. The Rayleigh-Ritz method was successfully applied to compute the characteristics of tunable resonators. For dielectric resonators employing axially magnetized ferrites tuning range in the order of 40MHz (nearly 5%) was achieved with Q-factor greater then 3000. The tuning speed in the order of milliseconds was obtained. Constructed dielectric resonator filter containing axially magnetized ferrites allowed tuning range of 18 MHz for tuning current limited within 0 - 0.5A. The filter characteristics do not change significantly and losses of the filter are below 0.8 dB. Second filter is based on the dielectric resonator containing circumferentially magnetized ferrite elements. The theoretical analysis of a resonator structure provided to obtain the optimal parameters of a resonator has been made. The resonator was constructed whose resonant frequencies can be switched by short current pulses between two stable values that differ in frequency up to 25MHz with Q-factor value about 4200. The filter is composed with two dielectric resonators.
Compositions, non tantalum containing have been prepared. High Q properties similar to the system BZT have been measured. High Q, commercial dielectrics were originally built around the material system Ba (Zn, Ta) O3. When the TUF project was instigated the price of Ta2O5 was very high and impeding development of ceramic resonator components. The design of new, high Q materials, which did not require Ta2O5 was seen as a key feature. The system developed under TUF, is based upon Ba (Zn, Co) NbO3. Substitution of Nb2O5 for Ta2O5 allows for a high Q, non tantalum conating dielectric. Doping the host BZN material with Co was necessary for tuning the temperature coefficient of resonant frequency (TCf). The control of TCf is one of the three key parameters for commercial realisation of dielectric materials. Materials have been realised which conform to the original TUF specifications. Exploitation applications include: - Manufacture of high Q, non tantalum dielectric resonator components. - Targets for PLD, CVD sputtering. - Compositions suitable for microwave chip capacitors.
The aim of this work package was to develop effective methods for depositing ferroelectric thin films by pulsed laser deposition (PLD). Our key interest was to deposit Ba Sr Ti O films (BSTO). These films were deposited on LAO and MgO substrates using single- and multi-target methods. The main objective was to broaden considerably the Curie transition in order to extend the temperature range of operation. This was achieved by depositing different stoichiometriy layers of BSTO that possessed different Curie transitions. Two routes of thin BaxSr1-xTiO3 (where x = 0.25, 0.50 and 0.75) films deposition have been explored: - Using a single target with corresponding stoichiometry and - From two targets as a periodic structure of BaTiO3 (BTO) and SrTiO3 (STO) layers. It was found that both methods produce samples with stoichiometry close to that expected. Raman spectroscopy showed that the transitional period in the thin films made by multi-target method was wider in the temperature range. The thin film made by the single target method becomes completely cubic at about 360 to 380K range whereas the thin film made from the two target film still shows the tetragonal and cubic phase at the same temperature range. This broadening of phase transition is also evident in the preliminary electrical measurements made. The results prove the possibility of using the multi-target pulse laser deposition as a more flexible method for engineering thin film stoichometry.
A method for fast, electrical tuning of a dielectric resonator is proposed. The results verify its underlying principle. By using a ferroelectric element, fast resonant frequency tuning of a DR while maintaining a useful Q-factor is possible. This is because the coupling between the high Q DR and the low Q ferroelectric element (film) and can be sensitively balanced by altering the distance between the ferroelectric and the DR. This is achieved simply by adjusting the spacer height. Further improvements in the device are possible by using higher quality BSTO films prepared by alternative techniques.
Two methods of characterising thin films have been designed. A 14 GHz Split Post Dielectric Resonator (SPDR) and a Stripline Double Resonator. The 14 GHz SPDR has been obtained from WUT and fully validated for substrate measurements. The SPDR is a non-destructive technique and has been specially designed to accommodate 10x10x0.5mm substrates, the largest size that films can be deposited onto. A generic mode matching code has been developed to enable the complex permittivity of the film to be de-embedded from that of the film substrate composite. The code can model arbitrary radial geometries and is therefore suitable for analysing arbitrary circularly symmetrical TE mode DR geometries. (e.g. both DR measurement cells and DR filters). Providing circular symetry is maintained the code can easily deal with very thin films. This code has been validated against black box software from WUT, Microwave Studios (MWS), and in house finite difference code. A Stripline Double Resonator has been modelled within MWS. Effort was spent on the modelling of the resonator, as this is necessary to de-embed the films complex permitivity from that of the film substrate composite. The difficulty with employing FDTD code is that the ratio of smallest to largest feature size is over 5000. The resonator is capable of measuring the change in complex permittivity of ferroelectric films under tuning. The resonator has been designed to work in the frequency range of the SPDR to allow the two methods to be used in tandem.
A frequency tuneable microwave filter based on a dual mode dielectric resonator has been developed. Tuning of the filter was realised by using metal plate located over the dielectric resonator and moved along the resonator axis by a piezoelectric bimorph. The main parameters of the filter were: central frequency in the middle of the range of tuning 2.35GHz, bandwidth for 1dB level 0.5%, range of tuning 10%. The insertion losses in the whole tuning range were below 1dB.