Stefan Marksteiner, J. Kaitila, G. Fattinger, R. Aigner
{"title":"Optimization of acoustic mirrors for solidly mounted BAW resonators","authors":"Stefan Marksteiner, J. Kaitila, G. Fattinger, R. Aigner","doi":"10.1109/ULTSYM.2005.1602861","DOIUrl":null,"url":null,"abstract":"The overall performance of bulk acoustic wave (BAW) filters is dominated by the effective coupling coefficient and the quality factor of the constituting BAW resonators. Whereas the effective coupling coefficient and its dependency on the layer stack is quite accurately modeled with a simple one-dimensional acousto-electric model (e.g. Masonstransmission line model), the prediction and optimization of the resonators quality factor - particularly for solidly mounted resonators (SMR) - completely fails with this model: whereas a calculation of the acoustic reflectance of a standard quarter-wavelength mirror stack leads to theoretical Q-factors well above 10000, experimental SMR devices with this type of mirror show values of typically well below 1000. This discrepancy is commonly explained by either visco-elastic loss in the materials and/or laterally leaking waves leaving the active resonator area. However, we have found a new, far more important loss mechanism relating to shear waves generated in the device. These waves can be created by injection from the resonators border area as well as by reflection/refraction of longitudinal waves at non-perpendicular angle of incidence to a material interface. In this paper, a quantitative methodology for the optimization of the acoustic mirror layer stack will be proposed. The influence of the mirror structure on the trapping of both longitudinal and shear wave energy will be discussed based on this very simple approach. Trade-offs with respect to the other important device parameters, such as effective coupling coefficient, temperature coefficient of frequency (TCF) and purity of the electrical response, are analyzed. The usefulness of this approach for the optimization of resonator Q-values will be proven by experimental results demonstrating Q-factors of 1500 and higher.","PeriodicalId":302030,"journal":{"name":"IEEE Ultrasonics Symposium, 2005.","volume":"43 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2005-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"90","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE Ultrasonics Symposium, 2005.","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ULTSYM.2005.1602861","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 90
Abstract
The overall performance of bulk acoustic wave (BAW) filters is dominated by the effective coupling coefficient and the quality factor of the constituting BAW resonators. Whereas the effective coupling coefficient and its dependency on the layer stack is quite accurately modeled with a simple one-dimensional acousto-electric model (e.g. Masonstransmission line model), the prediction and optimization of the resonators quality factor - particularly for solidly mounted resonators (SMR) - completely fails with this model: whereas a calculation of the acoustic reflectance of a standard quarter-wavelength mirror stack leads to theoretical Q-factors well above 10000, experimental SMR devices with this type of mirror show values of typically well below 1000. This discrepancy is commonly explained by either visco-elastic loss in the materials and/or laterally leaking waves leaving the active resonator area. However, we have found a new, far more important loss mechanism relating to shear waves generated in the device. These waves can be created by injection from the resonators border area as well as by reflection/refraction of longitudinal waves at non-perpendicular angle of incidence to a material interface. In this paper, a quantitative methodology for the optimization of the acoustic mirror layer stack will be proposed. The influence of the mirror structure on the trapping of both longitudinal and shear wave energy will be discussed based on this very simple approach. Trade-offs with respect to the other important device parameters, such as effective coupling coefficient, temperature coefficient of frequency (TCF) and purity of the electrical response, are analyzed. The usefulness of this approach for the optimization of resonator Q-values will be proven by experimental results demonstrating Q-factors of 1500 and higher.