Time and Frequency Standards:
to develop the standards that serve as reference
for time-and-frequency services and research on advanced measurement
systems.
INTENDED OUTCOME AND
BACKGROUND
The intended outcome of this program is the continuous operation of
frequency and time standards with the accuracy and stability essential for
supporting U.S. commerce and trade. This includes coordinating our activities
with those of other national standards laboratories.
The NIST time scale provides accurate time-and-frequency reference for our
services and for research on new standards and measurement methods. It is an
ensemble of commercial cesium-beam standards and hydrogen masers, combined
under computer control. We are advancing the performance of the time scale by
acquiring more stable clocks and improving the electronic systems that read the
clock outputs. Such improvements are critical to the successful evaluation and
use of the next generation of primary frequency standards.
The accuracy of the NIST time scale is derived from the current primary
frequency standard, NIST-F1, a cesium-fountain standard with a relative
frequency uncertainty of 1 fHz/Hz. The Division plans to build an improved
cesium-fountain frequency standard in the near future, designed so as to
simplify the process of accuracy evaluation.
Since the accuracy of frequency standards has improved by more than an
order-of-magnitude every decade since 1950, and practical applications have
historically always followed these advances, an aggressive program of research
and development in this area is easily justifiable. The Division is developing
frequency standards based on optical transitions that have dramatically higher
Q factors. These optical standards already provide much higher
stability and promise orders of magnitude improvement in accuracy. They would
not have been practical choices, however, without the recent development of
extremely stable laser oscillators and the means for directly connecting their
optical outputs to microwave frequencies.
Since the world operates on a unified time system, Coordinated Universal Time
(UTC), highly accurate time transfer (to coordinate time internationally) is a
critical ingredient in standards operations. To achieve this coordination, the
Division is developing and using several satellite-transfer
techniques.
Accomplishments
Improved Steering of UTC(NIST) to UTC
The process of running a national time scale involves periodically steering it
to the international average, that is, to UTC. Since UTC is formed by averaging
clock data submitted by many institutions, this BIPM-generated paper average,
delivered on a monthly schedule to all laboratories, represents the performance
of the time scale at past times. In fact, for a given monthly report, the most
recent BIPM data are at least two weeks old. Each laboratory must have a
strategy for accomplishing this steering.
Figure 1. Judah Levine and Tom Parker discussing strategies for
time-scale steering. |
Until recently, NIST opted to place maximum emphasis on frequency stability.
BIPM data arriving mid-month were analyzed, and a fixed steering rate was
determined for the following month. For example, the decision might have been
made to add (or to subtract) 1 ns per day to (from) the scale for the
following month. A decade ago, the noise on UTC(NIST) and UTC made it difficult
to steer more carefully than this.
Stimulated by the acquisition of five hydrogen masers over the last decade and
the development of a more accurate primary frequency standard, NIST-F1, we
investigated other steering algorithms. Using old clock data, we found that
improved time-offset performance could be achieved by increasing the number of
steers to two each month and by modifying the objectives of the steering. The
philosophy for steering was changed from an emphasis on frequency stability
only to one in which frequency stability could be degraded slightly to achieve
smaller time offsets.
The result is that UTC(NIST) now typically deviates from UTC by no more than
20 ns, an improvement of better than a factor of two.
Reproducibility of Optical Frequency
The mercury-ion optical-frequency standard has the potential for an accuracy
surpassing that of the cesium-fountain standard by two or more orders of
magnitude. With a Q factor greater than 1014 and a
transition that is relatively insensitive to environmental factors, the
potential uncertainty for the standard is as small as 1 aHz/Hz
(0.001 fHz/Hz).
Figure 2. Measurements over two years of the frequency of the mercury
optical standard relative to the cesium primary frequency standard. |
The reproducibility of this standard relative to the present cesium standard
(NIST-F1) was studied over a two-year period. Measurements were referenced to
NIST-F1 through the intermediary of a hydrogen maser. The short-term stability
of the mercury standard is superior to that of NIST-F1, so the maser played a
key role in the comparisons. The variation in the frequency of the S-D optical
transition relative to NIST-F1 was found to be less than ± 10 fHz/Hz
over the two years.
The uncertainty of the absolute measurement (at this same level) is the most
accurate measurement ever made of an optical frequency and a very encouraging
result, since no significant effort had been made to control systematic
frequency shifts. The largest of these shifts is expected to be the atomic
quadrupole shift, which depends on the orientation of the applied magnetic
field relative to ambient, static electric-field gradients. Concepts for
determining and controlling this and other smaller shifts have been developed,
but further studies are needed to test them.
Improved Calcium Frequency Standard
Because the Doppler-cooling limit for calcium is 2 mK, the uncertainty in
frequency measurements of the 657 nm transition are limited at about
10 Hz by the residual motions of the atoms. Thus, second-stage cooling is
needed to improve this as an optical-frequency standard.
| |
Figure 3. The calcium optical-frequency standard being adjusted by Chris
Oates. |
Last year, we reported a concept for achieving sub-Doppler cooling, called
quenched narrow-line laser cooling (QNLC), and demonstrated it in one dimension.
This year, we extended our system to three dimensions.
The three-dimensional cooling relies on the same basic principal, but uses
simultaneous rather than sequential excitation for the three-level system
(clock and quenching transitions). A frequency-doubled diode-laser system
generates the light at 423 nm used for Doppler cooling. The second-stage
cooling requires an additional beam at 552 nm for the quenching
transitions.
With these methods we reduced the average atom velocity by a factor of 15 and
lowered the atom temperature to 10 µK. At this temperature, the frequency
shifts due to motional effects are < 1 Hz at the clock transition,
456 THz. Further improvements should reduce the frequency uncertainty due
to motions to < 0.4 Hz, which corresponds to a relative
uncertainty of less than 1 fHz/Hz.
Having achieved 10 µK, we turned off the three-dimensional cooling lasers
and reapplied a pulsed version of one-dimensional QNLC cooling to the
already-cooled atoms. This reduced the temperature even lower, to 300 nK.
This suggests that even further cooling is possible and that the calcium
optical-frequency standard has much greater potential than previously
recognized.
First strategic focus |
Second strategic focus |
Third strategic focus |
Fourth strategic focus
"Technical Activities 2002" - Table of Contents |
