Rf mems theory design and technology pdf
In this application the shunt airbridge switch can be closed in the first step. This reduces the RF power at the Toggle Switch which then can be opened in the next step. Furthermore, the behaviour of the Toggle Switch under RF power for the duration of 60 s was investigated see Tab.
The Toggle Switch reopened directly up to a power of mW. When a power of mW is applied to the closed switch for 60 s, instead of 10 s as used for the RF adhesion tests, the metal cantilever welds to the metal contact pad. This seems to be caused by a temperature rise in the contact area. Power measurements of several different types of Toggle Switches were carried out Tab.
Comparing the different Toggle Switches it can be seen that the switches which needed a higher actuation voltage allow a higher RF power before adhering to the contact pad. The maximum power handling capability is achieved from the Toggle Switch with the highest actuation voltage 50 V.
A power of 2. Four of the six investigated switches could handle powers between 1 W and 2. Power measurements of the open Toggle Switch in the frequency range from 5 GHz to 18 GHz have shown that there is no power induced self actuation up to 2 W input power. A Wiltron B network analyser was used to control the Wiltron B source and the Alessi probe station was equipped with Picoprobe probes.
The switch time measurement results of the single Toggle Switch for the closing and releasing event are shown in Fig. In the static case a voltage of 20 V was needed to close the switch. For the switch cycle measurement a slightly higher maximum voltage of 25 V was applied.
The DC measurement set-up depicted in Fig. It is important to prevent a large potential difference between the cantilever and the contact paddle during the closure of the Toggle Switch.
A too large potential difference will weld the cantilever to the paddle due to a high current running over the contact. To prevent this, the set-up realized in Fig. At the time the pull electrode is switched to ground 0 V the contact will close causing a current flow Imess that can be measured. To open the contact U1 and U2 have to be released simultaneously. Using this set-up more than 2. The Toggle Switch did not show any visual degradation during the switching process.
After 2. Therefore the contact resistance could not be measured. This problem was an adhesion problem between the flexible metal band material in this case Au and the signal line metallisation also Au. It could be solved by inserting a thicker adhesion layer we used 20nm Ti and by extending the size of the flexible metal band in the area which connects it to the signal line. Three different test series were performed at room temperature RT , each including 20 current-voltage measurements.
The gradient was extracted from a linear fit to determine the appropriate resistance Fig. The resistance of the measurement lines and probes is given by R3. Small voltages are required to pull the cantilever down to the contact paddle.
Any further increase in the applied voltage would lead to an instability called pull-in. Pull-in occurs if the electrostatic forces grow faster than the spring forces with respect to the displacements.
As a consequence, the structure snaps down to the counter electrode. Higher pull- in voltages increase the contact force and decrease the contact resistance. The RF measurements Fig. Open Toggle Switches were identified by contacting both sides of the signal line and observing the current flow Imeas due to a voltage Umeas of 0.
Open switches do not show a relevant current at this state. An actuation voltage Uact was applied to close the open switch and the corresponding current Imeas was recorded. This agrees the results from section VII. This contact resistance is not de-embedded and the sum of all resistances the probes, the measurement lines, and the signal line. Higher temperatures increase the resistance in the metal of the signal line.
The wafer is placed on a thermo-chuck for the on-wafer measurements. This is equal to gradient of The decreased actuation voltages are caused by a smaller stress gradient in the gold-nickel-gold cantilever compound. Each gradient leads to a warping of the cantilever. Higher the stress gradients lead to increased distances from the cantilever to the electrode and higher actuation voltages are needed.
This stress gradient influences the temperature dependence on the actuation voltages because the warping of the cantilever changes due to temperature.
The stress gradient depends on the sputtering process parameters gas pressure, the sputter time and sputter power. A reduction of the temperature dependence can be achieved by compensation structures as well as other materials. This corresponds to a gradient of Single switches show excellent RF performance, low switching times and high power capability up to 2. The statistical data shows differences in the measurement results which attribute to variations during the fabrication processes.
Therefore, additional investigation of critical steps during fabrication process are necessary to improve yield and uniformity. Brown, and G. Bonache, J.
Gil, J. Amat, F. Falcone, J. Bonache, I. Gil, T. Lopetegi, M. Laso, A. Marcotegui, M. Sorolla, and R. Falcone, F. Lopetegi, J. Baena, R. Pendry, J. Holden, D. Robbins, and W. Varadian, V. Vinoy, and K.
Rottenberg, X. Brebels, P. Ekkels, P. Czarnecki, P. Nolmans, R. Mertens, B. Nauwe- laers, R. Puers, I. Share This Paper. Background Citations. Methods Citations. Results Citations. Citation Type. Has PDF. Publication Type. More Filters. RF MEMS is a key enabling device technology with the potential for significant performance, cost, and integration benefits to communications and radar applications.
View 2 excerpts, cites methods.
0コメント