1. Field of the Invention
This invention relates generally to a microstrip-to-microstrip RF transition circuit and, more particularly, to a microstrip-to-microstrip RF transition circuit for a semiconductor wafer that employs an RF microstrip to a co-planar waveguide (CPW) transition.
2. Discussion of the Related Art
Microelectro-mechanical switches (MEMS) used for RF applications is a technology area that has potential for providing a major impact on existing RF architectures in sensors and communications devices by reducing the weight, cost, size and power dissipation in these devices, possibly by a few orders of magnitude. Key devices for existing RF architectures include switches in radar systems and filters in communications systems. However, while MEMS technology has demonstrated the potential to revolutionize such devices, MEMS devices have not been specifically designed for performance in harsh environments, typically required for military applications, such as unmanned aerial vehicles (UAV) and national missile defense (NMD) systems. Particularly, MEMS technology requires further development in order to be able to provide effective performance under large temperature variations, strong vibrations and other extreme environmental conditions.
An appropriate packaging scheme that combines the properties of traditional high-speed packages and compatibility with planar technology offers a solution to this issue. Packaging is one of the most critical parts of the RF and MEMS fabrication process. Packaging is the most expensive step in the production line and will ultimately determine the performance and longevity of the device.
A large number of publications exist relating to RF MEMS based circuits, including phase shifters, single-pole multiple-through circuits, tunable filters, matching networks, etc. Many of these circuits have been designed based on a microstrip line configuration. Therefore, in order to develop a compatible on-wafer packaging scheme, a microstrip-to-microstrip transition needs to be provided. Such a transition for a MEMS is disclosed in U.S. Pat. No. 6,696,645 describing a coplanar waveguide (CPW)-to-CPW transition. The transition needs to be as broadband as possible, with minimum insertion loss and no parasitic resonance up to 50 GHz.
It is an important design consideration for a broadband microstrip-to-microstrip transition to maintain a characteristic impedance of the transition at 50Ω, especially at high frequency (>5 GHz). The 50Ω characteristic impedance through the transition is necessary to minimize signal reflections that would otherwise provide signal loss and degrade device performance. The design problem occurs because of the need for a wider microstrip line, which provides a lower impedance, in order to accommodate for the anisotropic etching of the vias through a semiconductor silicon wafer. When silicon is etched in potassium hydroxide or tetramethyl ammonia hydroxide, the etch rate of the <100> crystal plane is much higher than the etch rate of the <111> plane. This means that the final etched structure has a pyramidal shape found in the <111> planes of the silicon crystal. The angle between the <111> and the <100> planes is 54.74°. Other semiconductor wafer materials, such as GaAs, InP, etc., have similar anisotropic etching profiles. Therefore, in order to get a 20×20 μm square at the bottom of the via, a 160×160 μm square at the top of the via is required. This means that the width of the microstrip line needs to be at least 200 μm at the top of the via to accommodate for the size of the top of the vias. The wider microstrip line has a decreased characteristic impedance (approximately 25-30Ω). This mismatch increases the return loss of a back-to-back transition, thus reducing the overall bandwidth of the structure.
A signal via 20 is formed through the top semiconductor wafer and is in electrical contact with the microstrip lines 12 and 14. Two ground vias 22 and 24 are formed through the bottom semiconductor wafer and provide an electrical contact between the upper ground plane 16 and the lower ground plane 18. The vias 20-24 have a “pyrmidical shape” from top to bottom because of the anisotropic etch rates through the crystal planes of silicon when the opening for the vias 20-24 are formed, as discussed above. The microstrip lines 12 and 14 and the vias 20-24 would be made of a suitable metal, as would be well understood to those skilled in the art.
Typically, the thickness of the semiconductor wafers is about 100 μm because this is the minimum wafer thickness for current wafer fabrication processes. It is desirable that the semiconductor wafers be as thin as possible so that the parasitic inductances generated by the vias 20-24 is as minimal as possible. When the openings for the vias 20-24 are etched for a wafer of this thickness, the timing of the etch produces about a 160×160 μm metallized square at the top end of the vias 20-24 so that the etch produces about a 20×20 μm square at the bottom end of the vias 20-24. The size of the top end of the vias 20-24 ensures that the openings for the vias 20-24 will be formed all the way through the thickness of the wafer.
The width of the microstrip line 12 is about 80 μm to provide the desired 50Ω. However, a widened portion 26 of the microstrip line 20 that makes electrical contact with the top end of the via 20 is wider than the metallized square at the top of the via 20 to provide a suitable electrical contact and the proper orientation and alignment. For the dimensions being discussed herein, the width of the widened portion 26 would be between 200 and 220 μm. Because the wider portion 26 is wider than the rest of the microstrip line 12, it has a different characteristic impedance, typically 25-30Ω. A tapered transition 28 between the widened portion 26 and the rest of the microstrip line 12 minimizes reflections provided by the change in the characteristic impedance, but does not eliminate them. Thus, significant signal loss occurs at the transition between the microstrip line 12 and the via 20, especially at high frequencies. The microstrip line 14 includes the same size transition to a widened portion 46 at the bottom end of the via 20 so as to maintain the 25-30Ω characteristic impedance.
In accordance with the teachings of the present invention, a microstrip-to-microstrip RF transition circuit is disclosed that employs a wide microstrip line transition to a short co-planar waveguide section. In one embodiment, a first microstrip line and a first ground plane are patterned on a top surface of a semiconductor wafer, and a second microstrip line and a second ground plane are patterned on a bottom surface of the semiconductor wafer. A signal via is formed through the wafer and makes electrical contact with the first and second microstrip lines. Likewise, at least one ground via is formed through the wafer and makes electrical contact with the first and second ground planes. The microstrip lines include a widened portion where the microstrip line makes electrical contact with the signal via. The widened portion of the microstrip line is positioned between extended portions of the respective ground plane so that a slot is provided between the widened portion and the extended portion. The widened portion of the microstrip line and the slot between the ground plane defines a CPW, where the width of the widenend portion and the width of the slot is selected to provide a characteristic impedance equal to the characteristic impedance of the rest of the microstrip line.
Additional advantages and features of the present invention will become apparent from the following description and appended claims, taken in conjunction with the accompanying drawings.
The following discussion of the embodiments of the invention directed to a microstrip-to-microstrip transition circuit employing a short CPW is merely exemplary in nature and is in no way intended to limit the invention or its applications or uses.
The combination of the waveguide portions 40 and 48 of the ground plane 38, the widened portion 34 of the microstrip line 32 and the slots 44 and 52 define a short CPW that has a certain characteristic impedance. The narrow portion of the microstrip line 32 has a characteristic impedance defined by the width of the microstrip line 32. The characteristic impedance of the CPW is defined by the width of the widened portion 34 and the width of the slots 44 and 52. The width of the widened portion 34 is defined by the diameter of the top end of the signal via 36, and the width of the slots 44 and 52 are selected so that the CPW has a characteristic impedance that matches the characteristic impedance of the narrow part of the microstrip line 32 for the width of the widened portion 34.
The CPW is wide enough to accommodate the anisotropic etching of the vias 36, 42 and 50. By utilizing this approach, the RF signal sees the minimum mismatch, and therefore the return loss can be kept below −10 dB for a wider bandwidth of operation.
After the via holes are etched, the co-planar waveguide ground planes are connected forming the microstrip ground plane, while the signal line transitions at the backside of the semiconductor wafer. This design can also be used for a microstrip line-to-co-planar waveguide transition because for some integrated RF circuits it is preferable to use a different type of interconnect on the inside and outside of the package.
The circuit 60 further includes a first CPW 80 defined by the widened portion 64 of the microstrip transition line 62 and two extended portions 82 and 84 of the ground plane 70, and the slots therebetween. Likewise, the transition circuit 60 includes a second CPW 90 defined by the widened portion 68, two extended portions 92 and 94 of the ground plane 72, and the slots therebetween. The CPW 80 and the CPW 90 have an effective characteristic impedance that matches the characteristic impedance of the narrow portion of the microstrip lines 62 and 66. Therefore, signals propagating on the microstrip lines 62 and 66 and through the wafer have a minimal return loss. In one embodiment, the characteristic impedance is 50 Ω.
A first microstrip transition line 108, a second microstrip transition line 110 and a first ground plane 112 are patterned on a top surface of the wafer 106. A second ground plane 114, a third ground plane 116 and a third microstrip line 118 are patterned on a bottom surface of the wafer 106. Four CPWs 120 and 122 transfer the signal from the microstrip line 108 to the microstrip line 118 and then to the microstrip line 110 in the manner as discussed above through signal vias 124 and 126.
One advantage of the design of the present invention is that in the case of an on-wafer packaging architecture, the ground plane of the microstrip can be used for forming a sealing ring. This means that the ring will always be connected to the RF line, and therefore the parasitic resonance due to its length will be reduced or even eliminated. Moreover, because most of the developed RF MEM switches are suspended over microstrip lines, this proposed packaging architecture can be of great interest in the industry.
The foregoing discussion discloses and describes merely exemplary embodiments of the present invention. One skilled in the art will readily recognize from such discussion, and from the accompanying drawings and claims, that various changes, modifications and variations can be made therein without departing from the spirit and scope of the invention as defined in the following claims.
This application claims the benefit of the priority date of U.S. Provisional Patent Application No. 60/584,328, titled Microstrip-to-Microstrip RF Transition Including Co-Planar Waveguide, filed Jun. 30, 2004.
Number | Date | Country | |
---|---|---|---|
60584328 | Jun 2004 | US |