Table of MEMS Switch Characteristics (as of late 1997)

This table summarizes some reported characteristics of MEMS switches and relays. The information was obtained via a  query sent to the mems@isi.edu mailing list.   If anyone wishes to add to or update the data listed here, please email me at mpf@ai.mit.edu .
 
 
Source [My estimate 
of MOSFETs]		  Kruglick  Schiele  Zavracky Zavracky Zavracky  YongHong
Technology 
status		 [Best 
commercial 
processes]		 Under 
development		 Published1 Fabricated Projected Projected Best 
available
On 
resistance 
R		 [~ 10 kOhm] ~ 2 Ohm 10 Ohm 1 Ohm     500 mOhm 
- 100 Ohm
Gate 
capacitance 
C		 [30 fF]   5 fF  10 fF    50 fF  0.25 pF 
- 0.5 pF
Actuating 
voltage V		 [1 V] 5 V  20-90 V 100 V   10 V  2 V - 5 V
Maximum 
frequency f		 [~ 3 GHz] >= 1 kHz >= 3 kHz  300 kHz 100 MHz     
Area A [~ 1 µm2]   100 µm 
× 100 µm 
= 104 µm2		 65 µm 
× 40 µm 
= 2600 µm2		 1 µm 
× 3 µm 
= 3 µm2		    
Thickness t [~ 1 µm]   525 µm  4 µm       
Volume v [~ 1 µm3] ~ 9.6 × 105
µm3		 5.35 × 106
µm3		 10,400 
µm3		     500 µm 
× 500 µm 
× 20 µm 
= 5 × 106
µm3
RC time 
constant T		 [~ 0.3 ns]   50 fs 10 fs     125 fs - 
50 ps
CV2 energy 
E		 [3 × 10-14 J]   2 × 10-12 J 
- 4 × 10-11J 		 10-10 J   5 × 10-12 J 10-12 J - 
1.25 × 10-11 J
CV2 RC
energy-delay 
product		 [9 × 10-15 J ns]   10-17 J ns - 
2 × 10-15
J ns		 10-15 J ns     1.25 × 10-16
J ns - 
6.25 × 10-13
J ns
CV2 RC v
quantity q		 [9 × 10-15
J ns µm3]		   5.35 × 10-11
J ns µm3 - 
1.07 × 10-8
J ns µm3		 1.04 × 10-11
J ns µm3		     6.25 × 10-10
J ns µm3 - 
3.12 × 10-6
J ns µm3

Discussion:

The first column of data shows, for comparison, the characteristics of minimum-sized MOSFET switches in the fastest commercially available processes.   These are my own personal rough estimates gleaned through word-of-mouth; I will refine them as I learn more specific details about what's available.

The rest of the columns show data for MEMS switches that was emailed to me by the people listed, and quantities that I derived from that data.

One caveat: The thickness and volume numbers vary widely, in large part because some contributors took the device thickness to be its height above the wafer surface, whereas others took it to mean  the normal wafer thickness.  What I actually intended was the device height plus the minimum thickness of  the wafer, after thinning, that would still be sufficient to support the device structurally.   This would allow calculation of the maximum packing density of devices in 3-D, assuming the substrate were thinned.  The figures in the table have not yet all been adjusted to be consistent with this interpretation.

General observations on the data:

Implications of the derived results:

Sources:

Other Links:

References:

1.  I.Schiele, J.Huber, B. Hillerich, C. Evers, F. Kozlowski. Micromechanical Relay with Electrostatic Actuation. Proceedings Transducers97, 9th International Conference on Solid-State Sensors and Actuators, Chicago/USA, 16-19 June 1997 

Misc. Notes:

Some manufacturers working on MEMS switches:
Michael Frank

Last modified: Sun Dec 7 16:48:40 EST