MEMS technology for timing and frequency control

TL;DR: As vibrating RF MEMS devices are perceived more as circuit building blocks than as stand-alone devices, and as the frequency processing circuits they enable become larger and more complex, the makings of an integrated micromechanical circuit technology begin to take shape, perhaps with a functional breadth not unlike that of integrated transistor circuits.

Abstract: An overview on the vise of microelectromechanical systems (MEMS) technologies for timing and frequency control is presented. In particular, micromechanical RF filters and reference oscillators based on recently demonstrated vibrating on-chip micromechanical resonators with Q's > 10,000 at 1.5 GHz are described as an attractive solution to the increasing count of RF components (e.g., filters) expected to be needed by future multiband, multimode wireless devices. With Q's this high in on-chip abundance, such devices might also enable a paradigm shift in the design of timing and frequency control functions, where the advantages of high-Q are emphasized, rather than suppressed (e.g., due to size and cost reasons), resulting in enhanced robustness and power savings. Indeed, as vibrating RF MEMS devices are perceived more as circuit building blocks than as stand-alone devices, and as the frequency processing circuits they enable become larger and more complex, the makings of an integrated micromechanical circuit technology begin to take shape, perhaps with a functional breadth not unlike that of integrated transistor circuits. With even more aggressive three-dimensional MEMS technologies, even higher on-chip Q's are possible, such as already achieved via chip-scale atomic physics packages, which so far have achieved Q's > 107 using atomic cells measuring only 10 mm3 in volume and consuming just 5 mW of power, all while still allowing atomic clock Allan deviations down to 10-11 at one hour

Summary

Introduction

• The performance of their electronic systems is generallylimited by the accuracy and stability of the clocks or frequency references they use.
• Unfortunately, their best clocks and frequency references (e.g., atomic clocks, oven stabilized crystal oscillators) are often too large or consume too much power to be used in portable applications.
• Fig. 1 presents the simplified system block diagram for an example handset receiver targeted for multiband operation, clearly showing that it is the high-Q RF filters, not the IF filter, that must be addressed.
• It particularly considers mechanical circuit concepts based on this technology, first presenting early mechanical circuit examples, and then attempting to suggest the MEMS technologies and attributes most suitable to enabling a generalized integrated micromechanical circuit platform.

II. MEMS Technology

• After achieving the cross section of Fig. 2(a), the whole wafer is dipped into an isotropic etchant, in this case hydrofluoric acid, which attacks only the oxide sacrificial layer, removing it and leaving the structural polysilicon layer intact and free to move.
• On the other hand, mechanical elements can be cascaded into long chains because of their extremely low loss—a benefit of their high Q. Note, however, that the ingredients required for a micromechanical circuit technology comprise much more than just small size.
• The resonators and other elements in the repertoire of a micromechanical circuit design environment should have frequencies or other characteristics definable by lateral dimensions easily specifiable by CAD.

A. High Frequency and Q

• On the macro-scale, a guitar string made of nickel and steel, spanning about 25′′ in length, and tuned to a musical “A” note, will vibrate at a resonance frequency of 110 Hz when plucked.
• As with nanoelectronics in the electrical domain, there are issues in the mechanical domain that might hinder the use of nanomechanical vibrating resonators (at least in their present form) for today’s communication purposes.
• Since the center of the disk corresponds to a node location for the radial contour vibration mode shape, anchor losses through the supporting stem are greatly suppressed, allowing this design to retain a very high Q even at this UHF frequency.
• Use of a materialmismatched stem maximizes the Q, allowing this design to set the record in frequency-Q product for any on-chip UHF resonator at room temperature.
• >1 GHz; unlimited w/scaling and use of higher modes Series Resistance, Rx ∼ 50–5,000 Ω∗ ∗Small values of Rx can be achieved using large dc-bias values and very small gaps, albeit at the cost of linearity, also known as Range.

B. Capacitive Transduction

• Note that Table I contains all capacitively transduced devices, which in general offer the best frequency-Q products among micromechanical resonator types, since they generally are constructed in single high quality materials, and thus suffer less from the material interface losses that can encumber other transducer types (e.g., piezoelectric).
• The voltage VP generated by the charge effectively amplifies both the force imposed by the ac excitation signal vi (applied to port 1) and the output motional current io generated (at port 2) by the dc-biased timevarying electrode-to-resonator capacitor that results when the disk vibrates.
• Rx goes to infinity, making this device an effective open circuit, as depicted by Fig. 4(b).
• Note that this can now be done via a simple transistor switch (e.g., a pass gate), since this switching function is out of the signal path, making switch loss a non-issue.

C. Thermal Stability, Aging, and Impedance

• Besides frequency range and Q, thermal stability, aging/drift stability, and impedance are also of utmost importance.
• Table II presents some of the micromechanical resonator devices designed specifically to address these parameters.
• In particular, the fixed-fixed beam device of row 1 in Table II utilizes a temperature-tailored top electrodeto-resonator gap spacing to attain a total frequency deviation over 27–107◦C of only 18 ppm, which actually betters that of AT-cut quartz.
• And is amenable to CAD specification, the piezoelectric device of row 3 in Table II still sacrifices the important high Q, on/off self-switching, and temperature stability attributes offered by capacitive transducers.
• Solid-dielectric capacitively transduced resonators employing a vertical-tolateral drive, and thereby not requiring a nanoscale lateral gap, have also been successfully demonstrated [36].

V. Micromechanical Circuit Examples

• Given that they satisfy all of the attributes listed in Section III, it is no surprise that capacitively transduced resonators have been used to realize the most complex micromechanical circuits to date.
• But there is another, perhaps more elegant, circuit-based remedy based on arraying.
• With an output equal to the sum of its resonator outputs, an N -resonator array composite exhibits an N times lower motional resistance and a substantially larger power handling capability than a stand-alone resonator, while still maintaining a comparable Q.
• In particular, the composite array filter in row 5 uses more than 43 resonators and links, which approaches a medium-scale integrated (MSI) micromechanical circuit.

VI. Micromechanical Resonator Oscillators

• Leeson’s equation [45] indicates that the stability of an oscillator, as measured by its phase noise, is inversely proportional to the Q of its frequency-setting tank element.
• Interestingly, the theory of [49] and [50] actually do not always completely describe all of the 1/f3 noise behavior, in that oftentimes more 1/f3 noise is present than can be theoretically generated by aliased transistor 1/f noise alone, especially at large resonator displacement amplitudes.
• It also reduces the vibration amplitude of the resonator and, hence, reduces noise aliasing through nonlinearity in its capacitive transducer, which then reduces 1/f3 noise [12], [49].

A. Micro-Oven Control

• In addition to good short-term stability, MEMS technology has great potential to achieve oscillators with excellent thermal stability.
• In particular, the tiny size and weight of vibrating micromechanical resonators allow them to be mounted on micro-platforms suspended by long thin tethers with thermal resistances many orders of magnitude larger than achievable on the macro-scale.
• Such a large thermal resistance to its surroundings then allows the platform and its mounted contents to be heated to elevated temperatures with very little power consumption (e.g., milliwatts).
• Fig. 11 presents the first such platformmounted folded-beam micromechanical resonator, where the nitride platform supporting the resonator also included thermistor and heating resistors that, when embedded in a feedback loop, maintained the temperature of the platform at 130◦C with only 2 mW of total power consumption [51].
• Much better reduction factors are expected with a more refined platform design.

VII. Towards Chip-Scale Atomic Clocks

• And microoven control might soon allow them to reach temperature stabilities down to 10−9 (and maybe even better), their aging rates will likely not match those of atomic clocks.
• An atomic clock, on the other hand, derives its frequency from the energy difference between the hyperfine states of an alkali metal atom, which is a constant of nature and, thereby, much more predictable and stable.
• Here, a cell containing the alkali metal in a sufficiently dense vapor state is interrogated by a laser at a wavelength absorbed by the vapor (i.e., that excites the single outer orbital electron to the next orbital; 894 nm for the cesium D1 line).
• A microwave oscillator capable of delivering the needed output power is then locked to the (very accurate) hyperfine splitting frequency via a feedback circuit that controls the oscillator frequency so that the photodetector intensity is maximized at the hyperfine peak.

A. Reducing Power Consumption Via Scaling

• As mentioned, the alkali metal atoms must be maintained at a sufficient density in a vapor state to operate the atomic clock, which means power must be consumed to heat the vapor cell that contains the atoms.
• For a tabletop atomic clock, this can take tens of watts of power.
• Once again, as with the vibrating resonators of Section IV, smaller is better.
• It should be noted that the above illustration considered only conductive heat loss for simplicity.
• This radiation heat loss, however, is also small, so that the actual micro-atomic tether-supported cell by the Symmetricom/Draper/Sandia team in the CSAC program, shown in Fig. 14, still requires only 5 mW of power to maintain a temperature of 80◦C in its vacuum enclosure [52].

B. Scaling Limits

• But along with its benefits, scaling also introduces some potential disadvantages.
• In particular, among the more troublesome disruptors of stability in a gas-cell atomic clock are collisions between the atomic gas species and the walls, which can dephase the atoms, disrupting their coherent state.
• The thermal isolation for this physics package is not quite as good as that of Fig. 14, so its power consumption is on the order of 75 mW, but its tiny vapor cell so far has permitted measured Allan deviations better than 10−11 at one hour.

C. Tiny Atomic Clocks

• Fig. 16 presents the Allan deviation plot and photo of a completely self-contained atomic clock by the Symmetricom/Draper/Sandia team in the CSAC program that occupies only 9.95 cm3, yet achieves an Allan deviation of 5 × 10−11 at 100 s, while consuming less than 153 mW of power.
• This is the smallest, lowest power atomic clock in existence to date.
• But it won’t remain so long; in particular, if things go as planned in the CSAC program, 1-cm3 versions consuming only 30 mW while attaining 10−11 at one hour Allan deviation will likely surface soon.

VIII. Toward Large-Scale Integrated Micromechanical Circuits

• Again, to fully harness the advantages of micromechanical circuits, one must first recognize that due to their microscale size and zero dc power consumption, micromechanical circuits offer the same system complexity advantages over off-chip discrete passives that planar IC circuits offer over discrete transistor circuits.
• Again, as with transistor circuits, LSI (and perhaps eventually VLSI) mechanical circuits are best achieved by hierarchical design based on building block repetition, where resonator, filter, or mixer-filter building blocks might be combined in a fashion similar to that of the memory cell or gate building blocks often used in VLSI transistor ICs.
• Unfortunately, such Q’s have not been available in the sizes needed for portable applications.
• In addition, high-Q often precludes tunability, making RF channel selection via a single RF filter a very difficult prospect.
• In addition, since RF channel selection relaxes the overall receiver linearity requirements, it may become possible to put more gain in the LNA to suppress noise figure (NF ) contributions from later stages, while relaxing the required NF of the LNA itself, leading to further power savings.

IX. Practical Implementation Issues

• Some of these issues were already described in Sections II and V, including aging and drift stability, temperature stability, motional impedance, and power handling (i.e., linearity).
• Most of the demonstrated evidence was at frequencies below 100 MHz.
• While there is presently little reason to doubt they will come, demonstrations of adequate aging, drift, and temperature stability are still needed at GHz frequencies, as are demonstrations of antennaamenable impedances past 1 GHz.
• But beyond device-centric performance issues, there are a multitude of practical implementation issues that also must be overcome before vibrating RF MEMS technology can enter mainstream markets.
• Among the more important of these are absolute and matching fabrication tolerances, packaging, and (hybrid or fully integrated) merging with transistor circuits, all of which must be solved with the utmost in economy, given that cost is generally paramount in wireless markets.

A. Absolute and Matching Tolerances

• Before embarking on this topic, it is worth mentioning that the advent of fractional-N synthesizers [57] now alleviates to some degree the accuracy requirements on high-volume reference oscillators for portable wireless devices.
• At the time of this writing, there are several companies endeavoring to commercialize vibrating RF MEMS technology, including Discera1 and SiTime2, both of which are pursuing timekeepers as initial products.
• Unfortunately, for obvious reasons, these companies do not publish manufacturing statistics.
• Reference [58] presents one of the first published investigations on the absolute and matching tolerances of radialmode disk resonators using polysilicon and polydiamond structural materials.
• The fabrication tolerances of [58] are not, however, sufficient to realize the RF channel-select filter bank of Fig. 17 without trimming or some other mechanism to null offsets.

B. Packaging

• Packaging has historically been an impediment to commercialization of many MEMS-based products.
• These include packages based on wafer-level glass-frit bonding of caps [60] and low-pressure chemical vapor deposition sealing of fully planar encapsulations [10], [61].
• Such package stresses not only can shift the absolute center frequencies of micromechanical resonators, but also can often degrade their temperature sensitivity, induce hysteretic behavior in their frequency versus temperature curves, and cause undue frequency drift.
• A resonator anchored to the substrate at a single point, such as the disk resonator of row 4 in Table I, would be much more resilient against package stresses than the clamped-clamped beam of row 1, which is anchored to the substrate at two points, and so directly absorbs any package-derived substrate strains.
• It is, however, more difficult to envision complete LSI mechanical circuits anchored at only one point.

C. Merging With Transistors

• Often called the 1st level of packaging (where resonator encapsulation comprises the 0th level), merging with transistors via bonding, flip-chip bonding, or direct planar integration, might soon become the bottleneck to realization of the larger proposed mechanical circuits.
• To date, a more practical variant of the process of Fig. 2 using poly-SiGe structural material LPCVDdeposited at 450◦C [15] has been demonstrated that should allow fully integrated merging of MEMS structures with 0.18-µm-channel-length CMOS using conventional metallization.
• Thus, a new structural material that can be deposited at a very low temperature (e.g., less than the 320◦C melting temperature of low-k Teflon dielectric), yet still retain very high Q at high frequency, is highly desirable.
• Recent literature suggests that nickel metal, which can be electroplated at 50◦C, might be a strong candidate structural material.
• Nevertheless, its VHF performance coupled with the low temperature of its deposition makes nickel a very intriguing prospect for modular posttransistor integration of vibrating RF MEMS with next generation nm-scale CMOS.

X. Conclusions

• MEMS-based realizations of timing and frequency control functions, including 0.09% bandwidth filters with less than 0.6-dB insertion loss, GSM-compliant low phase noise oscillators, and miniature atomic clocks posting 5 × 10−11 at 100 s Allan deviation (so far) and consuming only 153 mW have been described with an emphasis on the performance benefits afforded by scaling to micro dimensions.
• In particular, via scaling, vibrating RF MEMS devices have now reached frequencies commensurate with critical RF functions in wireless applications and have done so with previously unavailable on-chip Q’s exceeding 10,000.
• Given present transistor scaling trend towards lower dynamic range digital devices, such a relaxation in dynamic range requirements may be arriving at an opportune time.
• In fact, with knowledge of the micromechanical circuit concepts described herein, perhaps a reconsideration of the numerous ongoing research efforts to make transistors out of nanowires is in order.
• In the meantime, micromechanical resonator technology is making its way into commercial markets through several companies (e.g., Discera and SiTime) which are now sampling low-end timekeeper products based on this technology.

Figures

TABLE III Summary of Vibrating Micromechanical Circuits.

Fig. 4. Perspective-view schematic of the disk resonator of row 4 in Table I, emphasizing the degree to which it is reconfigurable via the voltage applied to its bias port 3. Here, with V3 = VP , the device is a high-Q LCR; with V3 = 0 V, the device is an open circuit (i.e., is “off”); and with V3 = a local oscillator signal vLO, the device becomes a mixler [28].

Fig. 6. Equivalent circuit and explicit mechanical-to-electrical equivalence for the bridged micromechanical filter of row 3 in Table III.

Fig. 7. Illustration showing that, in the same footprint, an array of small 1-GHz disks can achieve a larger sidewall surface area (hence, larger electromechanical coupling and smaller filter impedance) than a single 1-GHz ring made large to minimize impedance.

Fig. 14. Tiny physics package by Symmetricom/Draper/Sandia, featuring a tiny atomic vapor cell supported by long thin thermally isolating tethers that allow heating to 80◦C with less than 5 mW of heating power.

Fig. 13. Schematics summarizing the analytical determination of the powers required to maintain macro and micro vapor cells at 80◦C in vacuum. Here, the equivalent thermal circuit used also yields warm up times.

Fig. 11. (top) SEM of a fabricated folded-beam micromechanical resonator mounted on a micro-oven platform. (bottom) Measured power required to achieve a given temperature. Here, only 2 mW are required to achieve 130◦C [51].

Fig. 12. Schematic describing the basic topology of a vapor cell-based atomic clock, this one exciting the hyperfine resonance via a lasermodulated approach dubbed coherent population trapping (CPT).

Fig. 1. System block diagram of a future multimode receiver front end, showing how the number of RF passives will continue to increase, regardless of whether or not a direct-conversion architecture is used.

Fig. 10. Comparison of measured phase noise density versus carrier offset for the 62-MHz array composite disk oscillator circuit of row 7 in Table III against that of a 61-MHz stand-alone wine-glass disk oscillator, with an extrapolation for the former down to 13 MHz for evaluation against GSM specs.

Fig. 3. SEM of a fully integrated 16-kHz watch timekeeper oscillator that combines CMOS and MEMS in a single fully planar process [14].

Fig. 2. Cross sections (a) immediately before and (b) after release of a surface-micromachining process done directly over CMOS [14].

Fig. 17. Schematic diagram for an RF channel-select micromechanical filter bank, with an example showing how various input frequencies can be simultaneously selected via mere application or removal of resonator dc-biases. In the bottom plots, filters 2, 4, 5, and n are on, while all others are off.

Fig. 8. Schematic of a single chip of high-Q passives made possible by lateral-mode vibrating MEMS resonator technology. Here, a resonator technology that allows many different frequencies to be defined (e.g., via CAD layout) in a single film layer is assumed. The 0.25 × 0.5-mm2 chip includes the receive path high-Q passives of Fig. 1, plus additional oscillator tanks for improved phase noise and injection locking suppression. Although only receive path high-Q filters are included here, note that transmit filters, as well as other lateral-dimension-defined passives described in this paper, could also be added while still retaining a tiny chip size.

Fig. 9. Circuit schematic of the 62-MHz series resonant reference oscillator of row 7 of Table III using a 9-wine-glass disk array composite resonator frequency-setting element with a Q of 118,000 in vacuum [11].

TABLE II Thermal Stability and Impedance of Microresonators.

Fig. 16. Measured Allan deviation plot for the output of the 9.95- cm3 atomic clock by the Symmetricom/Draper/Sandia CSAC team. (The points are the measured data.)

Fig. 15. Chip-scale atomic physics package by NIST, occupying less than 10 mm3 in volume.

Full text

UC Berkeley
UC Berkeley Previously Published Works
Title
MEMS technology for timing and frequency control
Permalink
https://escholarship.org/uc/item/19m6207z
Journal
IEEE Transactions on Ultrasonics Ferroelectrics and Frequency Control, 54(2)
ISSN
0885-3010
Author
Nguyen, CTC
Publication Date
2007-02-01
Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California

ieee transactions on ultrasonics, ferroelectrics, and frequency control, vol. 54, no. 2, february 2007 251
MEMS Technology for Timing and Frequency
Control
Clark T.-C. Nguyen, Fellow, IEEE
Abstract—An overview on the use of microelectrome-
chanical systems (MEMS) technologies for timing and
frequency control is presented. In particular, microme-
chanical RF filters and reference oscillators based on re-
cently demonstrated vibrating on-chip micromechanical
resonators with
Q
’s
>
10,000 at 1.5 GHz are described as
an attractive solution to the increasing count of RF compo-
nents (e.g., filters) expected to be needed by future multi-
band, multimode wireless devices. With
Q
’s this high in on-
chip abundance, such devices might also enable a paradigm
shift in the design of timing and frequency control functions,
where the advantages of high-
Q
are emphasized, rather than
suppressed (e.g., due to size and cost reasons), resulting
in enhanced robustness and power savings. Indeed, as vi-
brating RF MEMS devices are perceived more as circuit
building blocks than as stand-alone devices, and as the fre-
quency processing circuits they enable become larger and
more complex, the makings of an integrated micromechan-
ical circuit technology begin to take shape, perhaps with
a functional breadth not unlike that of integrated transis-
tor circuits. With even more aggressive three-dimensional
MEMS technologies, even higher on-chip
Q
’s are possible,
such as already achieved via chip-scale atomic physics pack-
ages, which so far have achieved
Q
’s
>
10
7
using atomic
cells measuring only 10 mm
3
in volume and consuming just
5 mW of power, all while still allowing atomic clock Allan
deviations down to 10
;
11
at one hour.
I. Introduction
T
he performance of our electronic systems is generally
limited by the accuracy and stability of the clocks
or frequency references they use. For example, the ability
and speed with which a global positioning system (GPS)
receiver can lock to a GPS satellite’s pseudorandom sig-
nal and obtain position is dependent heavily upon how
well synchronized its internal clock is to that of the satel-
lite; here, the better the internal clock, the more likely
and faster the lock. Unfortunately, our best clocks and
frequency references (e.g., atomic clocks, oven stabilized
crystal oscillators) are often too large or consume too much
power to be used in portable applications. This forces us
to keep our best electronic systems on tabletops and out
of the hands of users, who must then access them through
Manuscript received April 7, 2006; accepted August 23, 2006.
Much of this work was supported by the Defense Advanced Re-
search Projects Agency (DARPA) and the National Science Foun-
dation (NSF).
C. T.-C. Nguyen was with the Department of Electrical Engineer-
ing and Computer Science, University of Michigan, Ann Arbor, MI
48109-2122. He is now with the Department of Electrical Engineering
and Computer Sciences, University of California at Berkeley, Berke-
ley, CA 94720 (e-mail: ctnguyen@eecs.berkeley.edu).
Digital Object Identifier 10.1109/TUFFC.2007.240
sometimes unreliable remote channels. Indeed, a technol-
ogy capable of miniaturizing and lowering the power con-
sumption of our best timekeepers and frequency references
to the point of allowing insertion into truly portable ap-
plications would be most welcome.
Among the many applications enabled by timing and
frequency control, portable wireless communications stand
to benefit most from miniaturization technology, since an
ability to shrink high-Q passives can potentially change the
premises under which wireless subsystems are designed. In
particular, today’s wireless transceivers are designed un-
der a near mandate to minimize or eliminate, inasmuch as
possible, the use of high-Q passives. The reasons for this
are quite simple: cost and size. Specifically, the ceramic fil-
ters, surface acoustic wave (SAW) filters, quartz crystals,
and now film bulk acoustic resonator (FBAR) filters, ca-
pable of achieving the Q’s from 500–10,000 needed for RF
and IF bandpass filtering and frequency generation func-
tions, are all off-chip components that must interface with
transistor functions at the board level, taking up a sizable
amount of board space and comprising a sizable fraction
of the parts and assembly cost.
Pursuant to reducing the off-chip parts count in mod-
ern cellular handsets, direct-conversion [1] or low-IF [2]
receiver architectures have removed the IF filter, and in-
tegrated inductor technologies are removing some of the
off-chip L’s used for bias and matching networks [3]. Al-
though these methods can lower cost, they often do so at
the expense of increased transistor circuit complexity and
more stringent requirements on circuit performance (e.g.,
dynamic range), both of which degrade somewhat the ro-
bustness and power efficiency of the overall system. In ad-
dition, the removal of the IF filter does little to appease the
impending needs of future multiband, multimode reconfig-
urable handsets that will likely require high-Q RF filters
in even larger quantities—perhaps one set for each wire-
less standard to be addressed. Fig. 1 presents the simplified
system block diagram for an example handset receiver tar-
geted for multiband operation, clearly showing that it is
the high-Q RF filters, not the IF filter, that must be ad-
dressed. In the face of this need, and without a path by
which the RF filters can be removed, an option to reinsert
high Q components without the size and cost penalties of
the past would be most welcome, especially if done in a
manner that allows co-integration of passives with transis-
tors onto the same chip.
In this regard, microelectromechanical systems (MEMS)
technology, with its ability to shrink mechanical features
and mechanisms down to micron (and even nano) scales,
0885–3010/$25.00
c
 2007 IEEE

252 ieee transactions on ultrasonics, ferroelectrics, and frequency control, vol. 54, no. 2, february 2007
Fig. 1. System block diagram of a future multimode receiver front end, showing how the number of RF passives will continue to increase,
regardless of whether or not a direct-conversion architecture is used.
already provides substantial size and power reduction for
applications spanning displays, sensors, and fluidic sys-
tems, and is now emerging to provide similar advantages
for timekeepers and frequency control functions in portable
wireless devices [4], [5]. In particular, vibrating microme-
chanical resonator devices based on silicon micromachining
technologies have now been demonstrated with on-chip Q’s
greater than 10,000 at GHz frequencies [6], [7], frequency-
Q products exceeding 2.75×10
13
[6], temperature stability
better than 18 ppm over 27 to 107
◦
C [8], and aging sta-
bilities better than 2 ppm over one year [9], [10]. They
have also been embedded into oscillator circuits to achieve
phase noise performance satisfying global systems for mo-
bile communications (GSM) specifications for reference os-
cillators [11]–[13]. In addition, a combination of MEMS
and microphotonic technologies have now achieved 10-cm
3
complete atomic clocks consuming less than 200 mW of
power, while still attaining Allan deviations better than
5 ×10
−11
at 100 s. Continued scaling of such devices that
take advantage of compressed control time constants and
lower heating power consumption is expected to soon yield
complete atomic clocks in less than 1 cm
3
and consuming
less than 30 mW of power, while still achieving an Allan
deviation of 10
−11
at one hour.
But the benefits afforded by vibrating RF MEMS tech-
nology go far beyond mere component replacement. In
fact, the extent to which they offer performance and eco-
nomic benefits grows exponentially as researchers and de-
signers begin to perceive these devices more as on-chip
building blocks than as discrete stand-alone devices. In
particular, by mechanically linking vibrating mechanical
structures into more general networks, “integrated mi-
cromechanical circuits” can be conceived capable of imple-
menting virtually any signal processing function presently
realizable via transistor circuits, and with potential power
and linearity advantages, especially for functions that
involve frequency processing. In essence, micromechani-
cal linkages might form the basis for an integrated mi-
cromechanical circuit technology with a breadth of func-
tionality not unlike that of transistor integrated circuits.
Among the application possibilities are reconfigurable RF
channel-selecting filter banks, ultra-stable reconfigurable
oscillators, frequency domain computers, and frequency
translators. When further integrated together with other
microscale devices (e.g., transistors, micro-ovens, micro-
coolers, atomic cells), system-level benefits for portable
applications abound, particularly those for which archi-
tectural changes allow a designer to trade high Q for lower
power consumption and greater robustness, with poten-
tially revolutionary impact.
This overview paper describes not only the MEMS tech-
nologies that have achieved some the abovementioned per-
formance marks, but also what other capabilities their in-
tegration density might enable for timing and frequency
control in the coming years. It particularly considers me-
chanical circuit concepts based on this technology, first
presenting early mechanical circuit examples, and then
attempting to suggest the MEMS technologies and at-
tributes most suitable to enabling a generalized integrated
micromechanical circuit platform. The paper ends with a
discussion of some of the key practical implementation is-
sues that must be overcome if MEMS technology for timing
and frequency control is to be commercialized on a large
scale.

nguyen: mems technology for timing and frequency control 253
Fig. 2. Cross sections (a) immediately before and (b) after release of
a surface-micromachining process done directly over CMOS [14].
II. MEMS Technology
There are now a wide array of MEMS technologies capa-
ble of attaining on-chip microscale mechanical structures,
each distinguishable by not only the type of starting or
structural material used (e.g., silicon, silicon carbide, glass,
plastic, etc.), but also by the method of micromachining
(e.g., surface, bulk, three-dimensional (3-D) growth, etc.),
and by the application space (e.g., optical MEMS, bio
MEMS, etc.). For the present focus on timing and portable
communications, MEMS technologies amenable to low ca-
pacitance merging of micromechanical structures together
with integrated transistor circuits are of high interest.
To this end, Fig. 2 presents key cross sections describing
a polysilicon surface micromachining process done directly
over silicon CMOS circuits in a modular fashion, where
process steps for the transistor and MEMS portions are
kept separate, in distinct modules. (Modularity is highly
desirable in such a process, because a modular process can
more readily adapt to changes in a given module, e.g.,
to a new CMOS channel length.) As shown, this process
entails depositing and patterning films above a finished
CMOS circuit using the same equipment already found in
CMOS foundries until a cross section as in Fig. 2(a) is
achieved. Here, the structural polysilicon layer has been
temporarily supported by a sacrificial oxide film during its
own deposition and patterning. After achieving the cross
section of Fig. 2(a), the whole wafer is dipped into an
isotropic etchant, in this case hydrofluoric acid, which at-
tacks only the oxide sacrificial layer, removing it and leav-
ing the structural polysilicon layer intact and free to move.
Fig. 3 presents the SEM of a watch timing oscillator
that combines a 16-kHz folded-beam micromechanical res-
onator with a Q of 50,000 together with sustaining CMOS
transistor circuits using the very process flow of Fig. 2,
but with tungsten as the metal interconnect in order to
accommodate 625
◦
C structural polysilicon deposition tem-
peratures [14]. Although the use of tungsten metallization
instead of the more conventional copper and aluminum
Fig. 3. SEM of a fully integrated 16-kHz watch timekeeper oscillator
that combines CMOS and MEMS in a single fully planar process [14].
prevents the process of [14] from widespread use, other
variants of this modular process have now been demon-
strated that allow more conventional CMOS metals [15]. In
addition, other non-modular merging processes [16] have
been used in integrated MEMS products for many years
now. Whichever process is used, the size and integration
benefits are clear, as the complete timekeeper of Fig. 3
measures only 300 ×300 µm
2
, and could even be smaller if
the transistors were placed underneath the micromechan-
ical structure.
III. Integrated Micromechanical Circuits
The MEMS-enabled integration density illustrated in
Fig. 3 has the potential to shift paradigms that presently
constrain the number of high-Q components permissible in
the design of present-day timing or frequency control func-
tions, and instead allow the use of hundreds, perhaps thou-
sands (or more), of high-Q elements with negligible size
or cost penalty. In particular, MEMS technology sports
the attributes and ingredients to realize a micromechanical
circuit technology that could reach large-scale integrated
(LSI), or even very large-scale integrated (VLSI), propor-
tions, the same way integrated circuit (IC) transistors had
done over recent decades, and with potential for advances
in capabilities in the mechanical domain as enormous as
those achieved via the IC revolution in the electrical do-
main.
Like single transistors, stand-alone vibrating microme-
chanical elements have limited functionality. To expand
their functional range, micromechanical elements (like
transistors) need to be combined into more complex cir-
cuits that achieve functions better tailored to a specific
purpose (e.g., frequency filtering, generation, or transla-
tion). Given that the property that allows transistors to
be combined into large circuits is essentially their large
gain, it follows that mechanical elements can be combined
into equally large circuits by harnessing their large Q.As
a simple example, transistor elements can be cascaded in

254 ieee transactions on ultrasonics, ferroelectrics, and frequency control, vol. 54, no. 2, february 2007
long chains, because their gains compensate for the noise
and other losses that would otherwise degrade the signal
as it moves down the chain. On the other hand, mechanical
elements can be cascaded into long chains because of their
extremely low loss—a benefit of their high Q. In essence, if
an element has an abundance of some parameter (i.e., gain,
Q,...), then this can generally be used to build circuits of
that element.
Note, however, that the ingredients required for a mi-
cromechanical circuit technology comprise much more
than just small size. For example, the piezoelectric FBARs
[17], [18] that have already become a successful high-
volume product in the wireless handset arena, although
small, are perhaps not suitable for circuit design, since
their frequencies are governed almost entirely by thick-
ness, which is not a parameter that can be specified via
computer-aided design (CAD) layout. Given how instru-
mental CAD has been to the success of VLSI transistor IC
design, one would expect CAD amenability to be equally
important for micromechanical ICs. In this respect, the
resonators and other elements in the repertoire of a mi-
cromechanical circuit design environment should have fre-
quencies or other characteristics definable by lateral di-
mensions easily specifiable by CAD.
Continuing on this theme, a more complete set of at-
tributes needed to effect a micromechanical circuit design
environment can be listed as follows:
• CAD-amenable design. For example, frequencies
should be determined by lateral dimensions, which can
be specified via CAD, not just vertical dimensions,
which cannot. This makes possible an ability to attain
many different frequencies in a single layer on a single
chip.
• Geometric flexibility. Here, a given (often high) fre-
quency should be attainable in a wide variety of shapes
(e.g., beams, disks, etc.) and modes.
• Q’s > 1,000 from 1–6000 MHz, with Qs > 10,000
much preferred, if possible. Q’s this high are needed
to allow cascading of circuit blocks without accumu-
lating excessive loss, and to allow channel selection
at RF.
• Thermal and aging stability to better than 2 ppm,or
at least amenable to compensation or control to this
level.
• On/off switchability. Here, the overriding preference
is for vibrating micromechanical devices that can
switch themselves, i.e., that do not require extra se-
ries switches to do so, and that thus avoid the extra
cost and insertion loss.
• Massive-scale interconnectivity. In some cases, several
levels of both mechanical and electrical interconnect
are desired.
• Nonlinear characteristics that enable such functions as
mixing, amplification, limiting, and other useful signal
processing abilities.
• Amenability to low-capacitance single-chip integration
with transistors. This not only eliminates the issues
with high impedance (to be described), since the
tiny magnitudes of on-chip parasitic capacitors allow
impedances in the kΩ range, but also affords design-
ers a much wider palette of mechanical and electrical
circuit elements.
A mechanical circuit technology with the attributes
mentioned above might make possible filter banks capable
of selecting channels (as opposed to just bands) right at RF
with zero switching loss; oscillators using multiple high-Q
resonators to attain improved long- and short-term stabil-
ity; oscillators with oven-control-like temperature stability,
but consuming only milliwatts of power; ultra-low power
completely mechanical RF front ends for wireless handsets;
and all of these realized on a single silicon chip.
Before expanding on some of the above, the next section
first summarizes the basic resonator building blocks that
might be used to implement such mechanical circuits.
IV. Vibrating Micromechanical Resonators
Among the attributes listed above, the first two, re-
quiring CAD amenability and geometric flexibility, are
perhaps the most basic and the most difficult to achieve
if constrained to macroscopic machining technologies. In
particular, if GHz frequencies were required for flexural
mode beams (for which the frequency is governed in part
by length—a lateral dimension), then a macroscopic ma-
chining technology would be hard-pressed to achieve such
high frequencies. Fortunately, a major impetus behind
MEMS technology stems from the fact that mechanical
mechanisms benefit from the same scaling-based advan-
tages that have driven the IC revolution in recent decades.
Specifically, small size leads to faster speed, lower power
consumption, higher available complexity (leading to in-
creased functionality), and lower cost. And it does so not
only in the electrical domain, but in virtually all other
domains, including and especially mechanical. Although
many examples of this from all physical domains exist, vi-
brating RF MEMS resonators perhaps provide the most
direct example of how small size leads to faster speed in
the mechanical domain.
A. High Frequency and Q
For example, on the macro-scale, a guitar string made
of nickel and steel, spanning about 25

in length, and
tuned to a musical “A” note, will vibrate at a resonance
frequency of 110 Hz when plucked. In vibrating at only
110 Hz, and no other frequency, this guitar string is actu-
ally mechanically selecting this frequency, and is doing so
with a Q on the order of 350, which is ∼50 times more fre-
quency selective than a typical on-chip electrical LC tank.
Of course, selecting a frequency like this is exactly what
the RF and IF filters of a wireless phone must do, but they
must do so at much higher frequencies, from tens of MHz
to well into the GHz range. To achieve such frequencies

Frequently asked questions

Q1. How many ppm would be needed to match the RF channel?

In particular, since RF channel selection requires filters with percent bandwidths less than 0.2%, a resonator-to-resonator frequency matching tolerance better than 190 ppm would be needed if mismatch-induced passband ripple is to be held to less than 0.3 dB. 

Q2. What are the contributions in "Mems technology for timing and frequency control" ?

An overview on the use of microelectromechanical systems ( MEMS ) technologies for timing and frequency control is presented. 

Q3. How much power is needed to maintain the atoms in a vapor cell?

In particular, for the case of a vapor cell-based atomic clock, scaling a Cs- or Rb-filled atomic cell to millimeter or even micron dimensions greatly reduces the power required to maintain the cell at the elevated temperature needed to keep the atoms in a sufficiently dense vapor state. 

Q4. Why is modularity desirable in a CMOS process?

(Modularity is highly desirable in such a process, because a modular process can more readily adapt to changes in a given module, e.g., to a new CMOS channel length.) 

Q5. Why can mechanical elements be cascaded into long chains?

On the other hand, mechanical elements can be cascaded into long chains because of their extremely low loss—a benefit of their high Q. 

Q6. What is the main reason for the noise reduction in the resonator?

Needless to say, a better understanding of noise generation mechanisms in micromechanical resonator oscillators could lead to significantly better oscillator performance, so research in this area is expected to be abundant over the next few years. 

Q7. What does it mean that mechanical elements can be combined into equally large circuits?

Given that the property that allows transistors to be combined into large circuits is essentially their large gain, it follows that mechanical elements can be combined into equally large circuits by harnessing their large Q. 

Q8. How much cooling was achieved with the oven-controlled feedback?

With oven-controlling feedback engaged, the temperature coefficient of the platform-mounted micromechanical resonator was reduced by more than 5 times, with thermally induced warping of the platform the performance limiter. 

Q9. What is the current state of micromechanical circuit complexity?

At present, micromechanical circuit complexity is nearing medium-scale integration (MSI) density, as exemplified by the composite array filter in row 5 of Table III, which uses more than 43 resonators and links. 

Q10. What is the design strategy for a polysilicon ring resonator?

With this design strategy, this polysilicon ring resonator achieves a Q of 15,248 at 1.46 GHz, which is the highest Q to date past 1 GHz for any on-chip resonator at room temperature [7]. 

Q11. What is the importance of CAD for micromechanical ICs?

Given how instrumental CAD has been to the success of VLSI transistor IC design, one would expect CAD amenability to be equally important for micromechanical ICs. 

Q12. What is the inverse of the stability of an oscillator?

Leeson’s equation [45] indicates that the stability of an oscillator, as measured by its phase noise, is inversely proportional to the Q of its frequency-setting tank element and to the power circulating through the oscillator feedback loop. 

Q13. How many ppm of RF filter tolerances are needed for the die of Fig?

In other words, all of the filters in the die of Fig. 8 could be wafer-level fabricated with the needed ∼3% bandwidth filter specifications without any need for costly frequency trimming. 

Q14. What is the transfer function of a dc-bias VP?

while other resonators require a (lossy) switch in series to be switched in or out of an electrical path, a capacitively transduced micromechanical resonator can be switched in or out by mere application or removal of the dc-bias VP applied to its resonant structure. 

Q15. How many ppm of RF channel selection tolerances are needed for the Fig.?

In the meantime, one should not rule out the possibility that a production wafer-level fabrication facility might actually be able to achieve matching tolerances on the order of 190 ppm, which would allow trimless (hence, low-cost) manufacturing of the RF channel selector of Fig. 17. 

Q16. What are the attributes that require CAD amenability and geometric flexibility?

Vibrating Micromechanical ResonatorsAmong the attributes listed above, the first two, requiring CAD amenability and geometric flexibility, are perhaps the most basic and the most difficult to achieve if constrained to macroscopic machining technologies.