New Measurement Systems and Methods:
to develop new measurement systems and methods in
support of emerging technologies.
INTENDED OUTCOME AND
BACKGROUND
Through this program, the Division prepares for the future of
time-and-frequency measurements and calibrations. Through interactions and
discussions with NIST constituents, we identify important, emerging industrial
requirements and technologies, and develop measurement systems and methods
designed to meet their needs.
NIST has a long history of metrology expertise and leadership in such areas as:
the non-stationary statistics of clocks and oscillators; optical-frequency
measurement systems; atomic and optical oscillators; time-and-frequency
transfer through satellites; low-phase-noise electronics; and network
synchronization. The ultimate goal is to apply our expertise to carefully
selected projects so as to assure continuity in providing the
highest-performance time-and-frequency services to U.S. industry, science, and
the general public.
The synthesis and measurement of optical frequencies have become a central
focus in this program. The new optical-frequency combs generated by femtosecond
lasers have provided the means for measuring optical frequencies at uncertainty
levels orders-of-magnitude better than could be done previously. Such
measurements will clearly have impact on wavelength standards for fiber
communications systems and on absolute length metrology. These same combs are
being used to generate microwave outputs from optical-frequency standards,
essentially providing the basis for a whole new generation of atomic standards.
Now that the accuracy of these measurement systems has been established, work
involves simplification and noise reduction.
Another component of this program is the measurement of close-to-carrier noise
in oscillators and other electronic components. A closely related effort
involves improving statistical metrics for analyzing such noise. Projects in
this area will have impact not only on improvements in characterization of
clocks and oscillators, but also in telecommunications systems, advanced radars,
and other narrow-band electronic systems.
Finally, smaller, less-accurate oscillators play a major role in a variety of
measurement instruments and electronic systems. We have often contributed to
developing and improving such oscillators. The newest program in this area,
stimulated in large part by some of our fundamental research, is a (Defense
Advanced Research Project Agency) DARPA-funded program to develop a chip-scale
atomic clock, a program involving NIST, several universities, and a number of
companies.
Accomplishments
Improved Frequency-Comb Generator
One of the key stability/reliability issues with the usual frequency-comb
generator involves the microstructure optical fiber used to broaden the output
of the femtosecond laser to a full octave. The very high power density required
to achieve this broadening causes damage to the fiber. It must be adjusted and
replaced periodically.
We recently collaborated with the Institut für
Halbleitertechnik (Germany) to demonstrate a phase-coherent link from
optical to microwave frequencies in a system that does not require a
microstructure fiber for spectrum broadening. This success is based on the use
of a 1 GHz repetition-rate, mode-locked laser and the application of a
scheme that requires coverage of only 2/3 of an octave. In this
self-referencing scheme, the 2nd and 3rd harmonics of the
optical signal of interest are heterodyned with elements of the frequency comb
to ultimately lock the 1 GHz repetition rate of the mode-locked laser to
the optical signal. While our new 1 GHz femtosecond laser just generates a
full octave, further development will be needed to allow reliable locking using
only the fundamental and 2nd harmonic of the optical signal.
The immediate consequence of this development is to increase the uninterrupted,
microwave-frequency output, self-referenced to an optical transition, from tens
of minutes to about a day. Such improvement in reliability is essential to the
future of optical-frequency standards, since, ultimately, such standards will
have to operate unattended for long periods. In the experiments on this new
system, the microwave output frequency from a comb that was phase-locked to a
calcium optical-frequency standard (456 THz) was shown to have a
short-term frequency stability 100 times better than that of a hydrogen
maser. The long-term performance of the device will depend on how well
systematic effects in the optical standard are controlled.
Statistical Measure for Long-Term Stability
The measurement of the long-term stability of clocks and oscillators has been a
long-standing problem for science and industry. The Allan deviation and another
statistic, total deviation, both require data acquisition over twice the period
of the desired averaging interval. For example, to determine the stability at
1 month would require a two-month-long data run.
Recently, we developed a new statistic that yields a measure of stability at
the end-point of the data series. The most significant effect of this advance
is to cut required measurement times in half, thus substantially cutting the
cost of acquiring the most expensive data point.
Moreover, the statistic not only retains all of the desirable features of the
Allan deviation, it has fewer intrinsic biases and much narrower confidence
limits. We found that a more complex combination of frequency sums and
differences could yield an Allan-like statistic clear out to the interval of
the data run itself.
As a test, the statistic was used over the past two years for measuring the
performance of the NIST time scale and the primary cesium-fountain frequency
standard, NIST-F1.
The results clearly demonstrated the efficiency of the statistic. It served to
improve the performance of the time scale and to reduce the time required to
evaluate the accuracy of NIST-F1.
Broadly Tunable Microwave Reference Oscillator
The measurement of phase-modulation (PM) noise and amplitude-modulation (AM)
noise of clocks and oscillators presumes the availability of stable reference
oscillators at the desired measurement frequencies. However, it is too
expensive and cumbersome to maintain dedicated reference oscillators for each
measurement frequency. To date, methods for synthesizing offsets from stable,
fixed-frequency oscillators have been extremely complex. The problem is that
frequency-offset synthesis generally creates unacceptable levels of noise.
We recently developed a simple reference-oscillator system with broad tunability,
but without much of a noise penalty. In the traditional approach, a synthesized
offset is added to a stabilized oscillator. Therefore, the noises of these two
signal are independent and additive.
Figure 6. David Howe and Craig Nelson with the microwave cavity used to
suppress noise in their new reference oscillator. |
The innovation of the new system is in the placement of the offset synthesizer
inside a servo control loop used to stabilize a high-Q microwave cavity.
The frequencies of two oscillators, the reference signal and the offset, must
add up to the fixed resonance frequency of the cavity. Therefore, the
reference-signal frequency will be changed in a direction opposite to changes
in the synthesized offset frequency. The system suppresses the total noise on
the reference signal.
The concept was experimentally demonstrated using a high-Q microwave
cavitywith a resonance frequency of 10.6 GHz, a dielectric resonant
oscillator, and a digital offset-frequency synthesizer. The results
demonstrated the addition of tunability without substantially adding to the
noise. This development advances the art of PM and AM noise measurement.
Chip-Scale Atomic Clock
Earlier fundamental research at NIST on a very small atomic clock, based on the
concept of coherent population trapping, has served to focus broader interest
on the subject of miniaturization. In fact, a new DARPA-funded program was
stimulated by a NIST-hosted workshop held 1  years ago in Boulder. The advantage of the
coherent-population-trapping concept is that the traditional microwave cavity
is eliminated. The need for a microwave cavity has long been an impediment to
miniaturization below the few-centimeter level.
We are now collaborating with a group of eight companies and universities, also
funded by DARPA, on developing a chip-scale atomic clock. Because of our
expertise in miniaturizing atomic clocks, we will study basic aspects of
miniature atomic clocks to help guide developments on this aspect of the
program. Success in this endeavor could have impact on military systems,
particularly GPS, and civilian technology such as wireless telephony.
We are focusing on several projects. The first involves characterizing
miniature vapor cells, both wall-coated and buffer-gas-filled cells. One of the
objectives will be to experimentally establish the scaling (size) laws, thus
defining the limits for miniaturization. We are also supporting studies at the
University of Colorado of the cell-wall coatings deemed to be essential to
miniaturization to the submillimeter level. An auxiliary study, being performed
collaboratively with NIST Electronics & Electrical Engineering Laboratory
(EEEL) staff, is to examine the potential of direct interaction between
magnetically coated Micro Electro Mechanical Systems (MEMS) oscillators and
atoms. Success in this experiment could provide a dramatically simpler approach
to miniaturizing atomic clocks and atomic magnetometers.
First strategic focus |
Second strategic focus |
Third strategic focus |
Fourth strategic focus
"Technical Activities 2002" - Table of Contents |
