High Power and Reliable SPST/SP3T RF MEMS
Switches for Wireless Applications
J. Pal, Y. Zhu, Senior Member, IEEE,J.Lu,Senior Member, IEEE,D.Dao,andF.Khan
Abstract—This letter presents novel high power and reliable
radio frequency (RF) microelectromechanical systems switches
with single-pole single-throw (SPST) and single-pole triple-
throw (SP3T) configurations. An in-plane movable structure with
a single layer is made using a simple standard silicon-on-insulator
process, which greatly reduces the micro-fabrication complexity
and cost compared with the previously reported multi-contact
switches with out-of-plane movable structures. The SPST switch
achieves a uniform current distribution through each contact,
thereby increasing the power handling capability of the switch.
The SP3T switch is a derivative of the SPST switch with sepa-
rate individual actuations. The experimental results demonstrate
that the fabricated switches have superior RF performances:
insertion losses are −0.9 and −1.3dBat6GHzforSPSTand
SP3T switches, respectively, whereas isolations are better than
−29 and −37 dB from dc to 6 GHz for SPST and SP3T switches,
respectively. In hot-switching conditions, the SPST switch can
handle RF power up to 2 W for 10 million cycles, whereas the
SP3T switch is capable of handling an RF power of 1 W for
7 million cycles before failure occurs.
Index Terms— RF MEMS, switches, contact resistance,
single-contact and multi-contact micro-switches, SOIMUMPs.
I. INTRODUCTION
M
ICROELECTROMECHANICAL systems (MEMS)
switches have generated considerable interest owing
to their superior electrical performances compared to the
traditional solid-state solutions. For example, MEMS switches
have higher linearity and isolation, with lower insertion loss
and cost. The main challenges for the practical ap plication
of MEMS switches are poor long-term reliability and p ower
handlin g capability [1]–[5] . To address the contact stiction
issue under high current operation, hard metals such as
platinum (Pt) and ruthenium ( Ru) have been utilized as
the contact materials instead of gold. However, hard metals
suffer from high insertion loss due to their high electrical
resistivity and high contact resistance [6]–[8]. Another means
to improve reliability an d power handling cap ability is to
reduce the current through each contact by implementing
multiple contacts in the switch [9]–[11]. Nonetheless,
previous multi-con tact designs are sensitive to the high stress
and poor planarity issues caused b y unequally distributed
electrostatic forces on each contact, increasing the switch
insertion loss. In addition, the earlier reported multi-contact
Fig. 1. Scanning electron microscope images (SEM) of fabricated SPST in
the SOIMUMPs process: (a) Full view of micro-switch, (b) Close-up view of
the bias electrode area.
switches require complex an d costly fabrication processes to
produce out-of-plane movable microstructures.
This letter presents a novel laterally actuated multi-contact
switch based on symmetrically-designed radial contacts with
a shared central post. The switch is actuated with sepa rate
electrodes to control the current density and direction. For
example, all six contacts can be actuated simultaneously in
the SPST switch to lower the current den sity on each contact,
improving switch reliability and power handling capacity.
Moreover, the signal paths can be selected by actuating the
three separate electrodes in the SP3T switch. The actuators
are embedded inside the coplanar waveguide (CPW) to reduce
the discontinuity in the CPW [12]. The proposed devices are
fabricated using a commercially available standa rd fabrication
process (SOI-MEMS
TM
) through MEMSCAP with a single
structure layer and high production yield. Soft metals such
as gold and copper are over-coated on the contact area of
the switche s to reduce the r esistance loss in the material and
contact points.
II. D
ESIGN AND SIMUL ATION
Fig. 1 shows SEM images of the overall device and a
close-up view of the contact area of the fabricated SPST RF
MEMS switch. The design consists of three pairs of contacts
with electrostatically actuated cantilever beam s. Three separate
dc bias electrodes are located inside the signal line for each
contact pair to reduce the discontinuity in the CPW, as shown
in Fig. 1(b). When sufficient dc bias voltages V
1
,V
2
and V
3
are applied between cantilever beams and fixed electrodes,
the cantilever beams are pulled towards the fixed electro des
by the electrostatic force. The cantilever free-ends touch the
central post, resulting in the ON-state of the switch. Once the
dc actuation voltages are removed, the mechanical stresses in
TABLE I
O
PTIMIZED DIMENS IONS OF THE FABRICATED DEVI CE
Fig. 2. Simulated RF current distribution of SPST switch at 2 GHz.
the beams overcome the stiction forces and pull the cantilever
beams back to the original OFF-state.
The RF power handling capability of ohmic contact type
RF MEMS switches is limited by the self-heating effect
and contact temperature, which are generally proportional to
contact resistance [6]. Thus, to improve the power handling
capability of a switch , contact resistance has to be reduced,
which can be achieved by connecting multiple switch co ntacts
in parallel. However, the switch geometry has to be carefully
designed to maintain an even current distribution across mul-
tiple contacts [6] . To analyze the power handling capability
of the proposed design, the RF current distribution at 2 GHz
is simulated using the electromagnetic simulator of CST. The
optimized dimensions of the device are summarized in Table I.
The simulated current distribution for 1 W input power is
shown in Fig. 2. The results indicate that the currents flowing
through each contact are evenly distributed with ±0.2% vari-
ation. The small variation is due to the different paths through
which the RF signal has to travel, before r eaching the contacts
in the centr al part of the SPST. Therefore, a uniform heat
distribution throughout the switch and high power handling
capability can be achieved. Moreover, six-contact topology is
highly reliable since the current will redistribute and remain
uniform if one or more contacts fail. Thus, the switch is robust
to open circuit failures.
III. F
ABRICATION AND CHARACTERIZATION
The switches are micro-fabricated using MEMSCAP’s
Silicon-On-Insulator Process (SOIMUMPs), which offers a
movable 25 µm thick silicon layer, as illustrated in Fig. 3.
An in -plane movable structure with a single lay er can be
made in this process, which greatly reduces the micro-
fabrication complexity and cost compared to the previously-
reported multi-contact switches with ou t-of-plane movable
structures [9]–[11]. A gold layer of 0.65 µm and a copper
layerof1µm are over-coated on the top and sidewall of
Fig. 3. Silicon on insulator MEMS fabrication process flow.
Fig. 4. SEM image of fabricated SP3T switch.
Fig. 5. Measured and simulated insertion losses (ON state) and isolations
(OFF state) of the fabricated SPST switch.
the switching areas by a sputtering process. The over-coated
metal layers can significantly improve the RF performance
by reducing both the series resistance and the contact resis-
tance [3], [12]–[14]. SPST and SP3T MEMS switches were
fabricated in this process, and the SEM images are shown in
Fig. 1 and Fig. 4, respectively.
The RF performances of the fabricated SPST and SP3T
MEMS switches are characterized using a FieldFox RF
Vector Network Analyzer (N9923A) and ground-signal-
ground (GSG) coplanar probes with a pitch of 150 µm.
The system is initially calibrated using the stand ard short-
open-load-through on-wafer technique. S-parameters of the
input (S
1
) and output (S
2
) ports are extracted using the
Vector Network Analyzer. Fig. 5 illustrates the measured and
simulated RF performances of the SPST switch. In the ON
state, the measured insertion loss of the switch is approxi-
mately −0.9 dB at 6 GHz. In the OFF state, the measured
isolation is better than −29 dB for the whole measurement
frequency band. The simulation and measurement results are
well matched.
Fig. 6. Measured and simulated insertion losses (ON state) and isolations
(OFF state) of the fabricated SP3T switch (including 3 quadrants).
Fig. 7. Measured contact resistance and simulated contact force versus
applied voltage for SPST and SP3T switches.
Fig. 6 shows the measured and simulated insertion losses
and isolations in the frequency range of 0-6 GHz for the
fabricated SP3 T switch. The device is symmetric about the
input port (P1) as depicted in Fig. 4, therefore the RF perfor-
mances of the symmetric 3 output ports (P2, P3 and P4) are
almost identical. The measured insertion loss is appr oximately
−1.3 dB at 6 GHz and the measured isolation is better than
−37 dB at 6 GHz. The small variation in measured insertion
losses and isolations are due to the dimensional errors caused
by micro-fabrication uncertainty.
Fig. 7 illu strates measured contact re sistances and the
simulated contact force for SPST and SP3T switches. The
contact resistance varies from 7.61 to 3.03 for SPST and
from 9.78 to 3.81 for SP3T in the bias voltage range of
43-65 V. The corresponding simulated contact force changes
from 23 µNto61µN. High contact force reduces the contact
resistance by increasing the number and area of actual contact
spots [15]. Owing to its six contacts in parallel, the SPST
switch has a lower contact resistance compared to the SP3T
switch.
Fig. 8 shows the measured reliability and power handling
capacity of the fabricated switches with a switching rate of
100 Hz in hot-switching conditions. The contact resistance
was measured every 1 million cycles. No stiction was observed
before each resistance measurement. Because of its six contact
points, the SPST switch is able to handle up to 2 W for 10
million cycles. The RMS voltage of the 2 W RF signal is
estimated to be 10 V, which is m uch less than the measured
self-actuation voltage (pull-in voltage) of 43 V. The test was
stopped at 10 million cycles and not due to switch failure. RF
power over 2 W was not available in this experimental setup.
In comparison with the SPST switch, the SP3T switch can
Fig. 8. Hot-switching reliability and power handling capacity measurements
of (a) SPST switch and (b) SP3T switch.
TABLE II
COMPARI SON OF THI S WORK AND PREVI OUS RF MEMS SWI TCHES
handle less power since two contact points instead of six are
conducting electrical current. The SP3T switch can handle an
RF power of 0.3 W for up to 10 million cycles without failure.
With 1 W RF power, the test was stopped due to switch failure
after 7 million cycles, as shown in Fig. 8 (b).
Table II compares the measured characteristics with previ-
ously reported RF MEMS switches. The present work shows
superior RF performances, high power handling capabilities,
and simple fabrication processes owing to lateral movable
structures.
IV. C
ONCLUSION
This letter presents the design, implementation and char-
acterization of high power and r eliable RF MEMS lateral
switches using a simple SOI process with a single structure
layer. The switches have mu ltiple contacts to increase reli-
ability and power han dling capability. The fabricated SPST
and SP3T switches have superior insertion loss and isolation
ranging from dc to 6 GHz. The design can be scaled to have
more ports or contacts by increasing the radius of the central
post. The switches can be applied in wireless communication
applications, for example compact high-performance switch-
ing devices for dual band Wi-Fi smart antennas, which can
direct radiation at intended recipients and steer radiation nulls
toward the interfering signals [16], [17].
REFERENCES
[1] Y.-Q. Zhu, L. Han, L.-F. Wang, J.-Y. Tang, and Q.-A. Huang, “A novel
three-state RF MEMS switch for ultrabroadband (DC-40 GHz) appli-
cations,” IEEE Electron Device Lett., vol. 34, no. 8, pp. 1062–1064,
Aug. 2013, doi: 10.1109/LED.2013.2269993.
[2] M. Barbato and G. Meneghesso, “A novel technique to alleviate the stic-
tion phenomenon in radio frequency microelectromechanical switches,”
IEEE Electron Device Lett., vol. 36, no. 2, pp. 177–179, Feb. 2015,
doi: 10.1109/LED.2014.2376594.
[3]J.Pal,Y.Zhu,J.Lu,F.Khan,andD.Dao,“Anovelthree-state
contactless RF micromachined switch for wireless applications,” IEEE
Electron Device Lett., vol. 36, no. 12, pp. 1363–1365, Dec. 2015,
doi: 10.1109/LED.2015.2487510.
[4] J. Iannacci, “RF-MEMS: An enabling technology for modern wire-
less systems bearing a market potential still not fully displayed,”
Microsyst. Technol., vol. 21, no. 10, pp. 2039–2052, Oct. 2015,
doi: 10.1007/s00542-015-2665-6.
[5] J. Iannacci, “Reliability of MEMS: A perspective on failure mech-
anisms, improvement solutions and best practices at deve lopment
level, ” Displays, vol. 37, pp. 62–71, Apr. 2015, doi: 10.1016/
j.displa.2014.08.003.
[6] C. D. Patel and G. M. Rebeiz, “A high-reliability high-linearity high-
power RF MEMS metal-contact switch for DC–40-GHz applications,”
IEEE Trans. Microw. Theory Techn., vol. 60, no. 10, pp. 3096–3112,
Oct. 2012, doi: 10.1109/TMTT.2012.2211888.
[7] D. A. Czaplewski, C. D. Nordquist, C. W. Dyck, G. A. Patrizi,
G. M. Kraus, and W. D. Cowan, “Lifetime limitations of ohmic,
contacting RF MEMS switches with Au, Pt and Ir contact materials due
to accumulation of ‘friction polymer’ on the contacts,” J. Micromech.
Microeng., vol. 22, 2012, Art. no. 105005, doi: 10.1088/0960-
1317/22/10/105005.
[8] D. A. Czaplewski, C. D. Nordquist, G. A. Patrizi, G. M. Kraus,
and W. D. Cowan, “RF MEMS switches with RuO
2
–Au con-
tacts cycled to 10 billion cycles,” J. Microelectromech. Syst.,
vol. 22, no. 3, pp. 655–661, Jun. 2013, doi: 10.1109/JMEMS.2013.
2239256.
[9] H. Zareie and G. M. Rebeiz, “Compact high-power SPST
and SP4T RF MEMS metal-contact switches,” IEEE Trans.
Microw. Theory Techn., vol. 62, no. 2, pp. 297–305, Feb. 2014,
doi: 10.1109/TMTT.2013.2296749.
[10] R. Stefanini, M. Chatras, P. Blondy, and G. M. Rebeiz, “Minia-
ture MEMS switches for RF applications,” J. Microelectromech.
Syst., vol. 20, no. 6, pp. 1324–1335, Dec. 2011, doi: 10.1109/
JMEMS.2011.2170822.
[11] B. Lakshminarayan, D. Mercier , and G. M. Rebeiz, “High-
reliability miniature RF-MEMS switched capacitors,” IEEE Trans.
Microw. Theory Techn., vol. 56, no. 4, pp. 971–981, Apr. 2008,
doi: 10.1109/TMTT.2008.919076.
[12] M. Sterner, N. Roxhed, G. Stemme, and J. Oberhammer, “Static zero-
power-consumption coplanar waveguide embedded DC-to-RF metal-
contact MEMS switches in two-port and three-port configuration,” IEEE
Trans. Electron Devices, vol. 57, no. 7, pp. 1659–1669, Jul. 2010,
doi: 10.1109/TED.2010.2048239.
[13] A. Q. Liu, M. Tang, A. Agarwal, and A. Alphones, “Low-loss
lateral micromachined switches for high frequency applications,”
J. Micromech. Microeng., vol. 15, no. 1, pp. 157–167, 2005,
doi: http://dx.doi.org/10.1088/0960-1317/15/1/023.
[14] J. Park, E. S. Shim, W. Choi, Y. Kim, Y. Kwon, and D.-I. Cho,
“A non-contact-type RF MEMS switch for 24-GHz radar applications,”
J. Microelectromech. Syst., vol. 18, no. 1, pp. 163–173, Feb. 2009,
doi: 10.1109/JMEMS.2008.2011124.
[15] S. Majumder, N. E. McGruer, G. G. Adams, P. M. Zavracky,
R. H. Morrison, and J. Krim, “Study of contacts in an electrostati-
cally actuated microswitch,” Sens. Actuators A, Phys., vol. 93, no. 1,
pp. 19–26, 2001, doi: 10.1016/S0924-4247(01)00627-6.
[16] W. H. Chen, J. W. Sun, X. Wang, Z. H. Feng, F. L. Chen, Y. Furuya,
and A. Kuramoto, “A novel planar switched parasitic array antenna with
steered conical pattern,” IEEE Trans Antennas Propag., vol. 55, no. 6,
pp. 1183–1187, Jun. 2007, doi: 10.1109/TAP.2007.898643.
[17] R. Schlub, J. Lu, and T. Ohira, “Sev en-element ground skirt monopole
ESPAR antenna design from a genetic algorithm and the finite element
method,” IEEE Trans. Antennas Propag., vol. 51, no. 11, pp. 3033–3039,
Nov. 2003, doi: 10.1109/TAP.2003.818790.