Cesium Clocks: Telling Time with Spectroscopymhayden/mcep/portfolio/... · Atomic clocks are the only timing devices with the precision needed to operate GPS. They keep track of time
Post on 08-Jul-2020
2 Views
Preview:
Transcript
Cesium Clocks: Telling Time with Spectroscopy
by
Mark Hayden
A Thesis in Chemistry Education
Presented to the Faculty of the University of Pennsylvania in partial fulfillment of the
requirement of the degree of
Master of Chemistry Education
At
University of Pennsylvania
2009
__________________________
Constance W. Blasie
Program Director
__________________________
Dr. Michael Topp
Thesis Advisor
__________________________
Hayden Page 2
1. Introduction
Due to the nature of science, all scientific disciplines depend on advancements in
metrology, the science of measurement. The quality of data is limited by the accuracy and
precision of instruments used in the data collection process. Communication and peer review
cannot occur without
established physical
quantities that are universally
accepted and accessible.
Time, length, temperature,
and mass are fundamental
quantities from which many
others can be derived.
However, time is unique in that it can be measured thousands of times more accurately than
length, millions of times more accurately than mass, and at a small fraction of the cost of either
(Jones 141). Table 1 shows how accurately the base units of the International System (SI) can be
realized. As it is the most reproducible measurement in all of science, metrologists now aim to
use time to define as many other units as possible.
Since the meter, ampere, candela, and volt are all currently defined in terms of time, the
work of most scientists is at least indirectly connected with precise time measurement. However,
certain scientific fields completely depend
on it. One example of this is radio
astronomy, the branch of astronomy that
studies celestial bodies using radio waves.
Cosmic radiation is detected using large
interferometers as pictured in figure 1. In
the last 30 years, better data has come from
combining the signals from multiple devices
at different locations on Earth. The
technique, called very long baseline
interferometry (VLBI), requires atomic
Table 1- Most accurate realization of SI base units (Jones, 2000)
Measurement SI Base Unit Accuracy
Time Second 3 parts in 1015
Length Meter 1 part in 1012
Mass Kilogram 1 part in 108
Electric Current Ampere 4 parts in 108
Amount Mole 8 parts in 108
Temperature Kelvin 3 parts in 107
Luminous Intensity Candela 1 part in 104
Hayden Page 3
clocks to properly merge this data. Another scientific endeavor that depends on high precision
timekeeping is navigation of spacecraft. Due to relativistic effects, the location of spacecraft
such as Voyager 2 can only be accurately monitored using atomic time (NASA, 2009). One
final example of a scientific application of atomic clocks is researching the nature of time and
space by testing the relativistic effects alluded to above. In his theories of relativity, Einstein
predicted that time would run more slowly when measured with a moving clock and faster when
measured with a clock at high altitudes (Jespersen & Fitz-Randoph, 1999, p. 199). These
phenomena were confirmed in a variety of notable experiments involving atomic clocks.
In addition to the demand for precise time measurement in science, many practical
applications now require exceedingly low levels of uncertainty. In fact, the official United States
time, maintained by the National Institute of Standards and Technology (NIST), is accessed both
manually and electronically over 300 million times each day (NIST Physics Laboratory, 2001).
However, the use that relates to most individuals is its role in the operation of the electronic
power grid. Transmission of power involves frequency and voltage, which are both linked to
time by definition. Since electricity moves at close to the speed of light, atomic timing is the
only way to quickly locate problems on the grid (Jones 151). Synchronization, using precise
timekeeping to operate a system in unison, is also crucial in this case as the United States power
grid is operated by hundreds of different companies. Other industries that provide products and
services that involve frequency and/or synchronization also depend on state of the art
timekeeping devices. These include telecommunication providers and makers of electronic
equipment. Finally, the Global Positioning System (GPS) is another practical application of
atomic clocks. It consists of 24 satellites and 5 ground stations known as the control segment.
Although originally developed by the Department of Defense for military purposes, it is now
most commonly used by civilians for navigation. Anyone who purchases a receiver can access
GPS for free. The theory behind the operation of GPS is well beyond the scope of this
document. However, a basic explanation of how it works and the connection to atomic time is
provided in textbox #1.
Hayden Page 4
Textbox #1: Measurement and the Global Positioning System
All measured values are estimates that involve uncertainty. The time measurements required to operate
The Global Positioning System (GPS) carry some of the lowest uncertainties in all of science. These
measurements are made using atomic clocks.
What is GPS? What does it do?
The Global Positioning System (GPS) is a network of 24
NAVSTAR satellites and 5 ground stations. Initially,
the system was created by the United States Department
of Defense to provide precise coordinates needed to
conduct military operations. In 1983, Ronald Reagan
made the system free for personal use. Since that time,
individuals and companies with a receiver and a clear
view of the sky have used the system for a variety of
purposes. Some applications include determination of
location, navigation, timekeeping / synchronization,
mapping, and tracking.
How does GPS determine the location of a receiver?
The 24 satellites create a constellation 10,900 nautical miles
above the Earth. A receiver can determine its own location
by determining the position of any 4 of the satellites, much
like the way that navigators of ships used the stars to travel
by sea. Specifically, the receiver calculates the distance to 3
of the satellites. With these distances, it creates a sphere that
radiates from the center of each satellite. This narrows down
the receiver location to one of two points where the spheres
intersect. Often, one of these points can be eliminated because it is not on the surface of the Earth. The
final determination and error correction occurs when the receiver calculates the distance to a fourth
satellite.
Using time to calculate distance
To measure the distance between a satellite and the receiver, both begin generating a complex code at the
same time. This code is beamed from the satellite to the receiver as a
pattern of pulsating radio waves. However, because of the distance
involved, the code reaching the receiver will be out of sequence with the
code that is being generated at the receiver. If both the receiver and the
satellite started releasing the pattern at the exact same time, the time-shift
needed to get the code back in sync is the time that it took the radiation to
travel from the satellite to the receiver. Assuming the radiation travels at
the speed of light, the distance can be calculated using unit analysis:
(m) distance (s) times
m103.0 8
Due to the high velocity of light, small differences in the measured time produce large differences in the
Hayden Page 5
calculated distance. Even a difference of 1/100th of a second, many times faster than the blink of an eye,
results in drastically different calculated distances as shown below:
m10 x 1.8 s 0.060 s
m103.0 m10 x 1.5 s 0.050
s
m103.0
s 0.060 t s 0.050t
78
78
1.8 x 107 m – 1.5 x 10
7 m = 3.0 x 10
6 m
Note that a difference of 1/100th of a second resulted in a difference of 3 million meters, just under half of
the radius of the Earth! In addition, large errors occur in the distance calculations if the patterns of pulses
coming from the satellite and the receiver are not synchronized initially.
Time Measurement and Synchronization
Atomic clocks are the only timing devices with the precision needed to
operate GPS. They keep track of time by mimicking periodic electron
transitions within rubidium and cesium atoms. All 24 NAVSTAR satellites
have several atomic clocks on board, each keeping time with an uncertainty
of less than 1 x 10-8
. The 5 ground stations house even better atomic clocks.
They communicate with all of the satellites and each other to ensure that the
entire system remains synchronized. Clocks at these stations have
uncertainties as low as 5 x 10-16
, meaning that they are on pace to not gain or
lose a second for over 60 million years!
References Geocachegirls.com. (2007). Ground station [Data file]. Retrieved from http://geocachegirls.com/images/grnd_station.jpg
Levine, J. (2002, July). Time and frequency distribution using satellites. Reports on progress in physics, 65, 1119-1164.
National Air and Space Museum. (1998). GPS: A new constellation. Retrieved July 26, 2009, from Smithsonian Web site:
http://www.nasm.si.edu/gps/index.htm
Navstar [Data file]. (2009). Retrieved from Enseignement polytechnique Web site: http://www.enseignement.polytechnique.fr/mecanique/
Images/Navstar-2.jpg
NIST-F1 cesium fountain atomic clock [Physics laboratory: Time and frequency division]. (n.d.). Retrieved July 20, 2009, from National Institute of Standards and Technology Web site: http://tf.nist.gov/cesium/fountain.htm
Trimble Navigation Limited. (2009). GPS tutorial. Retrieved July 26, 2009, from http://www.trimble.com/gps/index.shtml
As previously mentioned, science requires established physical quantities that are
universally accepted and accessible. These references are called standards. Today, scientists
rely on the SI units and standards adopted by the General Conference on Weights and Measures
(CGPM). However, before the first version of the current system was created in 1875,
metrologists experimented and debated over various standards. For instance, some suggested
defining the meter as a fraction of the distance of the circumference of the Earth, while others
preferred using the length of a pendulum with a particular period (NIST, 2009). Ultimately, the
Hayden Page 6
former was used due to fluctuations in the pendulum’s period at different places on Earth. Using
this calculated distance, a platinum-iridium bar was made in 1889 to serve as the standard of
length. Copies of this standard were distributed to various countries. A similar standard with
copies was created for the kilogram.
Examples of both standards are shown in
figure 2. However, a similar standard was
not available for the second. The quest to
establish a “physical second” was the
driving force behind much of the
foundational research and developments
made in atomic timekeeping. By 1967,
the international scientific community was sufficiently satisfied with the accuracy, stability, and
reproducibility of the cesium clock and defined the second using the frequency of a specific
electronic transition within cesium. Rapid developments in atomic time continue to this day with
6 of the 7 base SI units, including the meter, now redefined in terms of the second. Although the
kilogram standard is still a platinum-iridium mass, many are in favor of redefining it in terms of
the second using the speed of light in Einstein’s famous equation (Jones 159):
E = mc2
(1)
2. Timekeeping and Frequency
Historical Clocks
Timekeeping involves counting and keeping track of cycles of a repetitive event.
Throughout history, this has been done in a variety of ways involving a broad spectrum of
sophistication. Periodic astronomical occurrences, especially the spin of the Earth and its
rotation around the sun, have long served as the repetitive event used to keep track of time.
Surely, even our earliest ancestors were able to count days simply by acknowledging the
repeating cycle of light and darkness. As early as 3500 B.C., Egyptians had developed obelisks
that could be used to divide the day into 12 parts (Jespersen & Fitz-Randoph, 1999, p. 11).
Solar timekeepers realized that the length of a day varies when measured using the
position of the sun in the sky. The combination of the Earth’s elliptical orbit, varying orbit
speed, and tilted axis produce a complex variation in the length of the solar day. This leads to a
Hayden Page 7
difference in nearly a full minute depending on the month. Although methods were developed to
correct sundials and other timekeeping devices for these factors, time cannot be measured
precisely on the basis of a day that is not constant. For this reason, astronomers developed the
“mean solar day.” It is a more convenient average value for the day that assumes that the sun’s
orbit in the sky as seen from Earth is consistent. It is essentially a mathematical correction for
the variations listed above (Jones, 2000, p. 5). For centuries, the second was defined as
1/86,400 of the mean solar day (Time Services Department, 2003):
(2)
Meanwhile, many were less interested in correcting solar clocks and more interested in
developing accurate timekeeping devices on Earth. Not only were these devices more accurate
than sundials and more useful for monitoring small increments of time, but they also could be
used on cloudy days or at night. The Egyptians, Chinese, Greeks, and Romans built
sophisticated sand and water clocks. However, because of a variety of problems with these
devices, the focus was shifted almost entirely to mechanical clocks by the start of the 15th
century (Jespersen & Fitz-Randoph, 1999, p. 36). Although many types of mechanical clocks
were constructed, few made more of an impact than the pendulum clock.
Galileo is generally credited with coming up with the idea of using a pendulum as the
repetitive event from which to keep time, although he did not construct a working device before
his death in 1642. Christiaan Huygens, known for his work in mathematics and physics, is
credited with developing the first working pendulum clock, which he patented in 1657. Initial
models of this clock were accurate within 10 seconds a day, which was a drastic improvement
over other clocks from the past (Jespersen & Fitz-Randoph, 1999, p. 37). However, it was a
new version of the pendulum clock, called the Shortt clock, which made the biggest impact in
science.
Hayden Page 8
Early versions of the pendulum clock were plagued by structural problems, such as
material expansion with temperature, which led to varying
periods and instability. Although these problems were soon
corrected by constructing pendulums using alloys, intrinsic
energy-related problems were more difficult to overcome
(Jones, 2000, p. 29). Pendulums lose energy due to air
resistance. In addition, energy is lost if the pendulum is
used to drive the clock mechanism. When outside energy is
added to allow the pendulum to continue, its period is
altered. In 1921 William Shorrt, a British engineer, helped
to correct these problems by inventing a clock with two
pendulums- a master and a slave pendulum. The master
pendulum swings freely in a vacuum chamber, while the
slave pendulum operates the time display. Electronic
signals sent between the two pendulums keep both swinging with a stable period. This
development initially produced clocks that were accurate within a few seconds over the course of
five years (Jespersen & Fitz-Randoph, 1999, p. 38). These clocks were widely used as standards
in metrology and astronomy laboratories at the beginning of the 20th
century and had a frequency
uncertainty around 1 x 10-7
(Lombardi, Heavner, & Jefferts, 2007, p. 75).
Clock Performance and Quality Factor (Q)
Every clock has two essential parts, an oscillator and a device for counting and displaying
the time (Jespersen & Fitz-Randoph, 1999, p. 39). The oscillator is also made up of two parts, a
resonator and an energy source. These
components are shown in figure 4. The
only job of the resonator is to move in a
periodic way. However, to maintain
stable periodic motion, energy must be
provided by an external energy source.
Finally, a device is needed to count and
Hayden Page 9
display the time. This device keeps track of the periodic cycles of the resonator and converts
groups of these cycles into an established time interval.
One of the most important characteristics of a clock is the frequency of its oscillator.
Frequency is the rate at which a repetitive event occurs or the number of cycles that occur in a
given unit of time. Every oscillator has a resonance frequency (f0), which is the natural rate at
which it oscillates. However, it is common for the
observed frequency of an oscillator to be one of a
range of values centered about f0. This range of
energy values is measured at the half-energy point,
and is referred to as the resonance width or
linewidth (∆f) and is shown in figure 5. ∆f is
inversely related to the decay time of the oscillator,
so the more narrow the resonance width, the longer
it will run without outside interference. In addition,
oscillators with narrow resonance widths do not respond to outside interference unless it is very
near to f0 (Jespersen & Fitz-Randoph, 1999, p. 47). For these reasons, smaller values of ∆f are
favorable in clocks.
Oscillators can be judged by comparing their quality factor (Q), which is a ratio of the
resonance frequency of the resonator to its resonance width:
f
f 0Q (3)
It follows that Q can be
increased by increasing the
resonance frequency or
decreasing the resonance
width. A clock’s Q value is
related to its accuracy and
stability, with higher Q’s
indicating a higher degree of
accuracy and stability. The most useful clocks are both accurate and stable. Accuracy is “a
Hayden Page 10
measure of the degree to which the clock replicates a defined interval of time,” while stability
indicates how well the oscillator can produce the same frequency over time (Itano & Ramsey,
1993, p. 56). Figure 6 illustrates the difference between the two. The quality factors for some
historical clocks are shown in table 2.
Table 2- Maximum quality factor for various oscillators (Jespersen, 1999)
Oscillator Maximum Quality Factor (Q)
Typical Balance Wheel Mechanical Clock 100-300
Household Pendulum Clock 104
Shortt Pendulum Clock 105
Modern Clocks
Soon after the development of the Shortt clock, focus had shifted away from pendulum
clocks and had moved toward clocks made with quartz-crystal oscillators. Quartz crystals
exhibit a piezoelectric effect, meaning that if a voltage is applied across the crystal, it will
resonate with a characteristic frequency (Kamas & Lombardi, 1990, p. 31). Although the
frequency depends on the size and shape of the crystal used, a common resonance frequency for
these crystals is 32,768 Hz (Jones, 2000, p. 32). This is much higher than the typical f0 of 1 Hz
for pendulums and led to Q values between 105-10
6 (Jespersen & Fitz-Randoph, 1999, p. 42).
By the 1930’s, quartz standards had replaced pendulums in most labs with uncertainties of 1 x
10-7
and 1 x 10-9
by 1950 (Lombardi, Heavner, & Jefferts, 2007, p. 76).
However, quartz clocks and the other clocks made beforehand shared the same
fundamental flaw, a lack of reproducibility. Each is very different depending on its specific
construction and the environment where it is operated. Well before the lowest uncertainties were
achieved in quartz standards, atomic oscillators were thought of as an answer to the problem of
reproducibility. In fact, Lord Kelvin may have been one of the first to make this suggestion in
1879 when he wrote:
“The recent discoveries due to the Kinetic theory of gases and to Spectrum
analysis (especially when it is applied to the light of the heavenly bodies) indicate
to us natural standard pieces of matter such as atoms of hydrogen or sodium,
ready made in infinite numbers, all absolutely alike in every physical property.
The time of vibration of a sodium particle corresponding to any one of its modes
of vibration is known to be absolutely independent of its position in the universe,
Hayden Page 11
and will probably remain the same so long as the particle itself exists” (Jones,
2000, p. 33).
Atomic clocks were only possible after substantial developments in quantum mechanics
and microwave technologies made before, during, and after the second world war (Diddams,
Bergquist, Jefferts, & Oates, 2004, p. 1319). Quantum mechanics suggests that atoms only
absorb and release energy in discrete amounts called quanta. Theoretically, the resonance
frequency of an atom depends on differences between allowed energy levels as shown in
equation 4, where h is Planck’s constant:
h
EEf 12
0 (4)
The first working atomic clock was actually based on a transition in a molecule, not an
atom. Harold Lyons, working for the National Bureau of Standards in 1949, applied microwave
technology used to make radar systems in WWII to cause ammonia molecules to resonate at
23,870 MHz (Jespersen & Fitz-Randoph, 1999, p. 55). This
was done by creating a microwave source connected to a
quartz oscillator, which was tuned to the resonance
frequency of ammonia molecules. The technique resulted in
a frequency uncertainty of 2 x 10-8
, a value in the range of
uncertainties already achieved by quartz oscillators
(Lombardi, Heavner, & Jefferts, 2007, p. 78). In addition,
several inherent frequency shifts that broadened ∆f limited
these clocks and made scientists look elsewhere.
Rabi and others had identified high f0 values in
hydrogen, rubidium, and cesium atoms (approximately 1.4,
6.8, and 9.2 GHz respectively) in the early 1940’s
(Lombardi, Heavner, & Jefferts, 2007, p. 79). As indicated previously, high resonance frequency
values produce higher Q’s. Beam technologies were also being developed that narrowed
resonance widths. Like the ammonia standards discussed above, rubidium and cesium beam
models use an external quartz oscillator to lock on to the resonance frequency. The amount of
time that the atoms interact with the external oscillator is called the interrogation time, τd.
Hayden Page 12
Quality factors are inversely related to τd as shown in the following relationship (Lombardi,
Heavner, & Jefferts, 2007, p. 79):
00
1
ff d
f (5)
The combination of higher f0 and lower ∆f values for atomic resonators produce much higher-Q
values and more accurate and stable clocks.
Table 3- Maximum quality factor for additional oscillators (Jespersen, 1999)
Oscillator Maximum Quality Factor (Q)
Quartz clock 105-10
6
Rubidium clock 106
Hydrogen maser 109
Cesium clock 1010
Although rubidium clocks and clocks built around hydrogen masers have advantages and
disadvantages, it is the cesium transition that is the basis of the best atomic clocks and frequency
standards used today. This is mainly due to some important properties of cesium that are
discussed in section 3. However, both the rubidium and hydrogen models are plagued by
frequency shifts and other problems that limit the time that they can operate effectively.
Defining the Second
With advancements in timekeeping and astronomy came the realization that complexities
surrounding the Earth’s rotation about its axis meant that the mean solar day should no longer
serve as the basis for determining the length of the second. Astronomers had already adopted a
new time scale called Ephemeris Time in 1952, which was based on Earth’s orbit around the sun.
In 1956, the International Committee on Weights and Measures suggested that the second should
be defined as 1/31 556 925.9747 of year 1900
(Bergquist, Jefferts, & Wineland, 2001, p. 37).
This definition was ultimately adopted by the
General Conference on Weights and Measures
in 1960.
Even before the adoption of the
ephemeris second, support was growing for a
definition of a second based on atomic time.
For the first time, this would provide universal
Hayden Page 13
access to a “physical second” that did not require complex and timely astronomical observations
and calculations. In 1955, Louis Essen and Jack Parry created the first operational cesium clock,
shown in figure 8, at the National Physical Laboratory in the UK. Once reliable cesium
standards became more common, effort was directed toward determining the resonance
frequency of cesium in relation to the ephemeris second. In 1958, this was completed and the
frequency was published as 9 192 631 770 ± 20 Hz (Markowitz, Hall, Essen, & Parry, 1958, p.
107). Finally, in 1967 at the 13th
Conference on Weights and Measures, the SI second was
defined as:
“ the duration of 9 192 631 770 periods of the radiation corresponding to the
transition between the two hyperfine levels of the ground state of the cesium 133
atom.”
In 1997, it was amended to state:
“a cesium atom at rest at a thermodynamic temperature of 0K”
3. Properties of Cesium
Physical Properties
Cesium has various physical properties
that make it an optimal choice for use in an
atomic clock. It is unique in that it is one of only
2 metals (along with Hg and Ga) that are liquid at
or close to room temperature. This is critically
important as its low melting and boiling points
(28.44°C and 668.3°C respectively) allow atoms
to be vaporized with minimal amounts of energy (Ball, 2007, p. 2). In addition, operating at the
lowest possible temperature helps to eliminate blackbody frequency shifts addressed in the
conclusion.
In common with most other elements having odd atomic numbers, Cs has only one stable
isotope. Although many other isotopes are possible, they are short-lived radioisotopes. In fact,
natural samples of cesium are isotopically pure, meaning that only one isotope makes up 100%
of its natural abundance. Here lies another advantage of using cesium within atomic clocks.
Hayden Page 14
Any sample used will contain only 133
Cs, which eliminates concern about the presence of
transitions from other isotopes (Jones, 2000, p. 40).
With stable atoms having 55 protons and 78 neutrons, cesium is a heavy atom. This
limits the speed at which the atoms can travel. In fact, cesium atoms move at just greater than
8% of the speed of hydrogen atoms (130 m/s vs. 1600 m/s) at room temperature (Lombardi,
Heavner, & Jefferts, 2007, p. 79). Slower moving atoms help to increase interrogation time in
the microwave cavity. Longer interaction times lead to narrower Δfa, which in turn increase the
Q factor. Higher Q factors are associated with more accurate and stable oscillators as addressed
in section 2 (Lombardi, Heavner, & Jefferts, 2007, p. 77).
Hayden Page 15
Textbox #2: Alkali Metals Alkali metals are the elements in group IA on the periodic table. Recall that
elements in the same vertical group or family have similar properties, because
they have similar electron configurations. As shown in table 1, each has the
same configuration as the preceding noble gas, as well as one additional
valance electron that occupies the s-sublevel.
As with other metals, alkali metals can lose their valence electron to form a
metal cation and gain the stability of an octet. However, since they are found
all the way to the left on the periodic table, these metals have the lowest
ionization energies. As a consequence, alkali metals are particularly reactive and are not found
uncombined with nature.
Reactivity in Water
Alkali metals react vigorously and exothermically with water. In
fact, these metals must be stored under oil (figure 2) to prevent
contact with moisture in the atmosphere. The reaction produces a
metal hydroxide, hydrogen gas, and heat according to the following
general equation:
2M(s) + 2H2O(l) → 2MOH(aq) + H2(g) + heat
It is not uncommon to observe flames or a small explosion (figure 3)
during these reactions due to the ignition of H2:
2K(s) + 2H2O(l) → 2KOH(aq) + H2(g)
Reactivity with water increases going down the group as a result of the decreasing ionization energy
trend. It is commonly believed that this increase in reactivity leads to a proportional increase in the
explosion that occurs during the reaction. However, this belief was called into question in 2006 when
Brainiac, a popular British television show, orchestrated an elaborate hoax involving the reaction of
rubidium and cesium metal in water. The episode is widely available online through media outlets such
as Google Video and YouTube.
During the episode, small samples of both rubidium and cesium were placed into a bathtub filled with
Lithium, Li (Z=3) [He] 2s1
Sodium, Na (Z=11) [Ne] 3s1
Potassium, K (Z=19) [Ar] 4s1
Rubidium, Rb (Z=37) [Kr] 5s1
Cesium, Cs (Z=55) [Xe] 6s1
Francium, Fr (Z=87) [Rn] 7s1
Figure 1: Alkali metals
Table 2: Alkali metal electron configurations
Figure 2: Potassium under oil
Figure 3: Potassium reacting with water
Hayden Page 16
water as shown in figure 4. The rubidium reaction caused most of the water to be expelled and put a hole
in the side of the tub. The cesium reaction completely destroyed the tub in a magnificent explosion.
Within days of the episode, skeptics were already calling for Brainiac to admit that they had used
explosives to dramatize the reaction. In reality, the explosiveness of the rubidium and cesium reactions is
limited by a much smaller output of hydrogen gas. Since these elements have much higher atomic
weights, they produce far fewer hydrogen atoms. Theodore Gray, a chemical enthusiast, created a
website containing videos of alkali metals reacting and an in-depth explanation for why he was unable to
get rubidium or cesium to explode at all.
http://theodoregray.com/PeriodicTable/AlkaliBangs/index.html
Handling alkali metals is extremely dangerous and should only be done under the close supervision of a
trained professional chemist. In addition to the risk of explosion, alkali metal hydroxides that are
produced in these reactions are powerful bases which can cause chemical burns or blindness as shown
below.
References Annenberg Media. (2009). Alkali metals [Data file]. Retrieved from http://www.learner.org/interactives/periodic/images/
alkalimetals_new.gif
Blazius. (2009). NaOH chemical burn [Data file]. Retrieved from http://upload.wikimedia.org/wikipedia/commons/0/09/
Sodium_hydroxide_burn.png
Brown, T. L., LeMay, H. E., Jr., & Bursten, B. E. (2000). Chemistry: The central science (8th ed.). Upper Saddle River, NJ:
Prentice Hall.
Gray, T. (2006). Alkali metal bangs. Retrieved July 26, 2009, from http://theodoregray.com/PeriodicTable/AlkaliBangs/
index.html
Potassium metal [Data file]. (2002). Retrieved from http://image24.webshots.com/24/2/1/90/38520190qdwZyi_ph.jpg
Stattotheblade. (2007). Brainiac alkali metals [Data file]. Retrieved from http://www.youtube.com/watch?v=eCk0lYB_8c0
Zumdahl, S. S. (2005). Chemical principles (5th ed.). Boston: Houghton Mifflan Company.
Another general physical property of cesium that makes it a strong candidate for use in
atomic clocks is its low 1st ionization energy. Cesium has the lowest first ionization energy
(375.7 kJ/mol) of all of the elements on the periodic table. This means that it readily loses an
Figure 4: Rubidium (left) and cesium (right) reacting with water during an episode of Brainiac
Figure 5: Chemical burn from NaOH
Hayden Page 17
electron after an input of only a minimal amount of energy. This has a practical application to
atomic clocks, because the detectors that are used in many older and commercially available
cesium-beam standards use a hot metal filament to ionize cesium atoms and generate an electric
current (Itano & Ramsey, 1993, p. 59). Types of detectors are presented in section 4.
Hyperfine Structure
Another advantage of using cesium
in atomic clocks is that it has a large nuclear
spin (I=7/2) and nuclear magnetic moment
that interact with its lone 6s1 electron
resulting in what is referred to as hyperfine
structure (Ball, 2007, pp. 2-3). Nuclei that
have an odd mass number have a non-zero
spin that results from the spin pairings of
protons and neutrons. These interactions
were studied extensively, particularly at
Columbia University, in the 1930’s by Rabi,
Cohen, Millman, Zacharias, and others
(Millman & Zacharias, 1937, p. 1049).
Zacharias is generally credited with coming
up with concept and design of the first
cesium atomic fountain which will be
discussed in section 5.
Electrons in p-orbitals can contribute to spin-orbit coupling. This leads to an energy
splitting which is referred to as fine structure. Although the 6s only has one fine state,
corresponding to J=1/2 and labeled as 6s 2S1/2 in figure 10, the 6p is split into two fine states that
result from the vector addition of the L and S quantum numbers:
J = L + S (6)
J = 1 + ½ (6p 2P3/2) (7)
J = 1 - ½ (6p 2P1/2) (8)
Hayden Page 18
In the presence of a magnetic field, such the one created by cesium’s magnetic nucleus,i
additional energy splitting called hyperfine structure can be detected at high resolution. This
splitting is the result of the vector addition of the nuclear spin (I=7/2) with J and leads to levels
labeled as F (M. Topp, personal communication, July 16, 2009):
F = I + J (9)
F = 7/2 + ½ = 4 (10)
F = 7/2 – ½ = 3 (11)
Similar hyperfine splitting occurs for the p-orbitals, with additional F values possible because of
the various combinations that result due to the presence of 3 orbital types. Each may also be
split into magnetic (mf) states in the presence of an external field.
It is important to note that the hyperfine energy splitting is often greatly exaggerated so
that it can easily be viewed on diagrams like figure 10. In reality, it is incredibly small in
relation to other energy values associated with the cesium atom as shown in table 4. In fact, the
hyperfine splitting is only a 1000th
of the thermal kinetic energy of a cesium atom at 100°C and a
mere 100,000th
of the magnitude of the ionization energy of the 6s electron (Nave, 2006).
Although hyperfine splitting is exceptionally small from an energetic perspective, it is the
basis for atomic time. Specifically, the ground state transition between the hyperfine F=3 and
F=4 states (which will be referred to as |3,0> and |4,0> respectively) within 6s 2S1/2 has been
used in the definition of the SI second since 1967. This particular transition is unique for a
variety of reasons. First, the energy splitting is particularly large due to cesium’s high nuclear
spin and nuclear magnetic moment (Ball, 2007, p. 3). This makes the corresponding resonance
frequency associated with the transition, 9 192 631 770 Hz, unusually large as well. Greater
resonance frequencies lead to increases in Q and better clock performance, as discussed in
section 2. In addition, the |4,0> ↔ |3,0> transition is insensitive to magnetic fields, which aids in
the selection and detection process utilized within atomic clocks (Lombardi, Heavner, &
Jefferts, 2007, p. 81). Finally, the resonance frequency corresponds to a wavelength of 3.26 cm,
Hayden Page 19
which is part of the microwave region of the electromagnetic spectrum. Equipment is readily
available to induce and detect this transition. In fact, many of the original parts used in clocks
came from advancements made in microwave technology that was used for radar and
communication during the World War II (Diddams, Bergquist, Jefferts, & Oates, 2004, p. 2).
4. Cesium-Beam Frequency Standard
Overview
As discussed in section 2, all clocks consist of two basic components: an oscillator and a
device that counts and displays the time. Once this external source of radiation (usually a quartz
oscillator) has been tuned to match the resonance frequency of a beam of cesium atoms, a
counter keeps track of the cycles. This is very similar to the way the master and slave
pendulums operate in the Shortt clock. Cesium beam standards are no longer used as official
primary frequency standards but are still used in commercially available devices.
The process specifically involves the ground-state |4,0> ↔ |3,0> hyperfine transition in
133Cs. The frequency of the radiation absorbed and released during this transition is referred to
as the resonance frequency and has been identified as 9 192 631 770 ±20 Hz. The corresponding
wavelength is 3.26 cm, which falls in the microwave region of the electromagnetic spectrum.
The goal is to monitor and match the frequency of the external source of microwaves with the
resonance frequency of the cesium atoms.
The source of the cesium beam is an oven where a sample of cesium atoms is converted
to the gaseous state. Since cesium has a relatively low boiling point (28.4°C) and high vapor
pressure, temperatures between 80°C-100°C are enough to produce more than enough vaporized
cesium for the process (Ball, 2007, p. 2). A small slit in the side of the oven releases a beam of
the atoms. Upon leaving the oven, both ground level hyperfine states will be populated
according to the Bolzmann distribution, with F=3 slightly favored (M. Topp, personal
communication, July 21, 2009). However, all 16 possible mf states in the ground level are
occupied, 9 for F=4 and 7 for F=3 (Sullivan et al., 2001, p. 49). Even if the external source of
radiation is properly tuned to the resonance frequency, these atoms will not undergo the desired
|4,0> ↔ |3,0> transition, which makes them useless in their current states.
Hayden Page 20
For this reason, a variety of state selection techniques are used to ensure that as many atoms as
possible are in the |3,0> hyperfine state before exposure to the microwaves. In figure 11,
magnets are used (labeled A magnet) to filter out atoms in undesirable magnetic states before
entrance into the microwave cavity. Magnets were used for state selection in NBS-6, which is
shown in figure 12. It was used as the primary frequency standard from 1975 to 1993. After the
development of optical selection and detection techniques, it was replaced by NIST-7, which is
shown in figure 16. Both state selection techniques are discussed in more detail in the next
subsection.
After state selection, the
beam of cesium atoms enters a
microwave cavity. In this
cavity, atoms are exposed to
microwave radiation from the
external source. If the
microwave radiation is tuned
precisely to the resonance
frequency of 133
Cs, most of the
atoms already in |3,0> will
Hayden Page 21
undergo a transition to |4,0>. These altered atoms are detected outside of the microwave cavity.
The number of |4,0> atoms at the detector reaches a maximum when the resonance frequency is
matched, so the frequency of the microwave radiation must be carefully adjusted until that
occurs. In the past, operators of the standard completed this meticulous process by hand.
However, modern standards employ Servo loops that electronically adjust the frequency in the
microwave cavity based on information provided by the detector. Both microwave cavities and
detectors are discussed in more depth in subsequent subsections.
State Selection Techniques
As indicated in section 3, various levels of fine (J) and hyperfine (F) structure are
detected in cesium at high resolution. In addition to these levels, magnetic (mf) splitting occurs
in the presence of an external
magnetic field. When gaseous
cesium atoms leave the oven,
they occupy all 16 of the
possible mf states within the 6s
2S1/2 ground level state
(Sullivan et al., 2001, p. 49).
These states are shown in figure
13. Both magnetic states
involved in the clock transition,
|3,0> and |4,0>, are insensitive to magnetic fields, which means that magnets can be used to
select for them (Lombardi, Heavner, & Jefferts, 2007, p. 81). In practice, the process is more
about selecting against the other 14 magnetically sensitive states. A device called a Stern-
Gerlach magnet allows atoms in the |3,0> and |4,0> to proceed toward the microwave cavity,
while deflecting atoms in other states toward a getter. However, it is still possible for atoms in
the wrong state to make it past the initial magnet. For this reason, another magnet is used to
select for the |4,0> after the microwave cavity and before detection.
Magnetic selection techniques are no longer used in the most accurate atomic clocks.
Clocks utilizing them experience an excessive amount of “noise” at their detector because of the
number of atoms in the wrong magnetic state that are able to pass by state selection magnets.
Hayden Page 22
This decreases detection efficiency and increases the amount of time needed to make
measurements. In addition, magnets have varying effects on atoms that are traveling at different
velocities, which can make the frequency distribution of the cesium atoms asymmetric (Sullivan
et al., 2001, p. 50). An asymmetric complicates the determination of the resonance frequency.
Currently, the best atomic clocks in operation use
optical state selection techniques. This means that they use
lasers to ensure that the proper states are populated within
cesium atoms before and after entering the microwave cavity.
Figure 14 indicates the specific transitions involved in optical
selection, which is referred to as optical or selective pumping.
The basis of optical pumping is the application of quantum
selection rules which allow only ∆F = 1, ∆mf = 0, ±1 to occur
(Arditi, 1982, p. 60). Atoms are continually pumped to the
specific fine states within 6p 2P3/2 by absorbing
monochromatic laser light. Many of the allowed transitions
result in a return to the |3,0> and |4,0> states (Sullivan et al., 2001, p. 50). Selective pumping
can be utilized to target only the atoms in the |4,0> state. Since |3,0> is a non-absorbing substate
in this situation, it will ultimately become nearly 100% populated. In fact, one of the greatest
advantages of using optical state selection techniques is that signal strength can be greatly
improved since such a large percentage of the atoms can be forced into the appropriate energy
level (Itano & Ramsey, 1993, p. 59).
Microwave Cavities
The purpose of the microwave cavity is to maintain the selected state of the cesium atoms
before and after exposing them to microwave radiation from the external oscillator. In order to
do this, cesium atoms must be shielded from external magnetic fields. This is generally done by
encasing the region with an alloy with high magnetic permeability. A highly uniform magnetic
field, called a C-field, is produced within the cavity to ensure that atoms do not revert back to
undesirable mf states that existed before the selection process (Major, 2007, p. 176).
Asymmetrical frequency distributions can occur if the C-field is not uniform, which result in
additional measurement error when determining the resonance frequency (Teles, Magalhaes,
Hayden Page 23
Santos, Bebeachibuli, & Bagnato, 2002, p. 136). Figure 15 illustrates the difference between
frequencies obtained with and without a uniform C-field.
In early devices developed by Rabi and others, atomic beams were exposed to one strong
microwave field as they passed through the cavity (Sullivan et al., 2001, p. 48). This increased
the interaction time and improved the Q factor as discussed in section 2. However, several
problems accompanied this method of excitation, most notably Doppler shifts, or changes in the
frequency that result due to the motion of the atoms in the beam. Atoms moving with a higher
velocity toward the frequency source appear to absorb a higher frequency than those moving
with lower velocities.
Norman Ramsey, a physicist at Harvard University, eventually developed an improved
cavity which contains a U-shaped design. In fact, most microwave cavities used in atomic
clocks and other applications are now called Ramsey cavities. Instead of exposing the atomic
beam to radiation for the entire length of the cavity, these cavities provide two identical doses of
microwave radiation- one at the beginning and one at the end. The sharpness of the frequency is
related to the travel time between successive exposures (Jones, 2000, p. 49). This technique,
called the separated oscillatory field method, nearly eliminates Doppler shifts and other effects
that contribute to broadened resonance widths (Lombardi, Heavner, & Jefferts, 2007, p. 78).
This is partially because the microwaves are applied perpendicular to the direction that the beam
of cesium atoms is traveling.
Hayden Page 24
Detection
After passing through the Ramsey cavity and
undergoing state selection, it is time for detection. At this
point, only atoms that have been promoted to the |4,0>
hyperfine ground state will still be traveling toward the
detector, as the others have been deflected or eliminated
by state selection. In the older and less accurate atomic
clocks that utilize magnets for state selection, hot-wire
ionizers are used as detectors. Since cesium atoms have a
particularly low ionization energy, contact with a hot
piece of metal ribbon provides more than enough energy
to cause their valence electrons to be released. This
produces an electrical current which provides a relative
indication of the number of |4,0> atoms reaching the
detector. When the electrical current reaches a maximum, the external source of microwaves has
been tuned to match the resonance frequency of the cesium atoms (Diddams, Bergquist, Jefferts,
& Oates, 2004, p. 1320).
More modern atomic clocks and frequency standards, like NIST-7, employ optical
selection and detection systems which result in uncertainties on the order of 10-9
. After
microwave interrogation and state
selection, atoms in the beam are pumped
from the |4,0> of 6s 2S1/2 to the F=5 of
6p 2P3/2, which is called a “cycling
transition” (Sullivan et al., 2001, p. 50).
When atoms are pumped in this way,
photons are scattered as the atoms
fluoresce (Sesko & Wieman, 1988, p.
270). Specific transitions involved are
shown in figure 10 in section 3. The
intensity of the fluorescence is
Hayden Page 25
maximized when the external source of microwaves has been tuned to match the resonance
frequency of the beam of atoms.
As mentioned in the overview, operators of early standards were charged with monitoring
the electrical current or fluorescence intensity and carefully adjusting the frequency of the
external microwave radiation by hand. However, nearly all modern devices contain a servo loop,
which is an electronic mechanism that automatically adjusts the external microwave frequency
until it matches the resonance frequency based on information from the detector. Most of these
devices employ a “square-wave modulation method” that takes measurements on both sides of
the resonance peak to determine the maximum frequency (Sullivan et al., 2001, p. 48).
Often, a device called a frequency divider takes the frequency output of a standard, which
is generally 5 or 10 MHz, and converts it to a smaller value like 1 Hz that can be used in timing
applications (Kamas & Lombardi, 1990, p. 48). In addition, a digital counter and display can
provide a visual output of a variety of time intervals. In the case of a primary frequency
standard, one or more additional oscillators such as quartz crystal oscillators are locked on to the
resonance frequency through the servo loop and control the microwaves that enter the Ramsey
cavity (Jespersen & Fitz-Randoph, 1999, p. 58).
Although the best cesium-beam frequency standards are capable of attaining frequency
uncertainties of 5 x 10-15
, they are limited by two main intrinsic factors- high atomic velocity
(~100 m/s) and short interaction times (~10 ms) (Lombardi, Heavner, & Jefferts, 2007, p. 84).
High atomic velocity leads to an increased number of collisions which broaden the resonance
linewidth. As discussed in section 2, short interaction times lead to lower quality factors (Q),
which are associated with less stability. NIST-7, the last cesium beam used as a primary
standard at NIST, had a linewidth of 65 Hz which corresponds to a Q value of 1.5 x 108
(Sullivan et al., 2001, p. 50).
5. Cesium Fountain Frequency Standard
Overview
Jerrold Zacharias is generally credited with introducing the concept of the cesium
fountain standard during his work at MIT in the 1950’s. His research was in response to a need
to increase interrogation times. Since then, various improvements to Zacharias’s original design
have lengthened interrogation times to ~1s, which has significantly narrowed linewidths to ~1
Hayden Page 26
Hz (Jefferts et al., 2002, p. 332). These and other
developments have resulted in Q values on the order
of 1010
and uncertainties as low as 6 x 10-16
(Bergquist, Jefferts, & Wineland, 2001, p. 40).
The basic goal of the cesium fountain is the
same as the cesium beam standard. NIST-F1, the
primary frequency standard currently used at NIST, is
shown in figure 18. The idea is to excite cesium
atoms from to using microwaves from an
external source, often a quartz oscillator that is to be
locked on to the resonance frequency. However, two
major differences exist between the beam technique and the fountain approach. First, instead of
heating atoms in an oven to create a beam, approximately 107 gaseous cesium atoms in a vacuum
chamber are cooled to nearly 1 µK before state selection and entrance into the microwave cavity
(Jefferts et al., 2002, p. 321). Second, instead of traveling in a horizontal beam, the ball of
cooled atoms is propelled upward through a Ramsey
cavity and allowed to fall back through the cavity under
the influence of gravity. Other minor differences also
exist and will be discussed in subsequent subsections.
Laser Cooling and Control
Cesium fountains generally use lasers to
establish a condition called optical molasses. The term
“optical molasses” refers to the severe reduction in the
diffusion rate of atoms that have been cooled under
these circumstances (Major, 2007, p. 348). The cesium
atoms used are positioned at the center of six laser
beams tuned slightly above 852 nm, which is near-
infrared radiation that is invisible to the eye (Jefferts et
al., 2002, p. 321). These beams trap the atoms and
cause them to be cooled to just above absolute zero
Hayden Page 27
within a fraction of a second. Lombardi, Heavner, and Jefferts describe the process of laser
cooling in two relatively straightforward ways (2007, p. 84). First, it can be understood from a
thermodynamic perspective which is similar to refrigeration. Lasers can be thought of as a
“sink” of extremely low entropy since they are made up of a highly uniform, monochromatic
light. The sample of cesium atoms contains much more entropy due to the random arrangement
and motion of the gaseous particles. When the cesium atoms interact with this laser, they scatter
the light. Since they are unable to escape in any direction, the atomic sample ends up with much
lower entropy than it had at the start. Lombardi, Heavner, and Jefferts suggest that laser cooling
could also be thought of as a mechanical effect (2007, p. 84). Cesium atoms absorb light with a
wavelength of 852 nm. The atoms receive a jolt of momentum in the direction that the photon
was traveling when this occurs. Soon after, each atom releases a photon with 852 nm in a
random direction. When the atom releases the photon, it “recoils” and gets another jolt in the
direction opposite the laser source. Since the atom emits light in random directions, the jolts of
momentum do not cancel out, which allows the atom to be pushed by the laser (Jones, 2000, p.
165). If the frequency of a laser is lowered slightly, atoms that moving toward it able to absorb
the light as a result of the Doppler effect. These atoms still release photons at 852 nm. In this
way, cesium atoms are made to release slightly more radiation than they absorb. This cycle is
repeated continually using three pairs of identical lasers that act in opposite directions as shown
in figure 19. At this point, the atoms are said to be “trapped” at the intersection of the lasers.
Once trapped, the atoms must be directed upward for state selection and microwave
interrogation. This is done by briefly detuning the frequencies of the lasers that are pointing
directly above and below the sample of atoms and simultaneously turning off the other lasers.
This propels the atoms directly upward at speeds around 4 m/s as opposed to the speeds around
100 m/s in beam standards (Bergquist, Jefferts, & Wineland, 2001, p. 40).
Hayden Page 28
State Selection
State selection techniques used in fountain
standards differ from those utilized in beam standards,
because they involve both microwaves and optics.
Figure 20 shows a schematic diagram of the inside of
NIST-F1. After cooling and trapping, almost all
atoms occupy the 9 mf states of |4,0> (Sullivan et al.,
2001, p. 53). These atoms first travel through the
inactive detection region into a magnetically isolated
state selection cavity. This state selection cavity is
identical to the Ramsey cavity. It provides a short
burst of microwaves that drives as many atoms into
|3,0> as possible. Immediately after receiving this
dose of microwaves, a laser is used to remove any
additional |4,0> atoms using the same strategy
employed in beam standards. State selection in NIST-
F1 is able to remove greater than 99% of the |4,0>
atoms (Jefferts et al., 2002, p. 322).
Microwave Cavities
After state selection, the cesium atoms continue their ascent into the Ramsey cavity. The
structure and function of these cavities are the same as those used in the beam standards that
were previously explained. However, interaction time is lengthened as microwaves are applied
once on the way up and then again after the atoms reach the apogee and travel back down.
Detection
Once atoms have gone through the Ramsey cavity, optical techniques are used to
determine how many of the atoms made the desired |4,0> ↔ |3,0> hyperfine transition. As with
the beam standard, fluorescence detectors are used. In the case of the fountain standard, the
|3,0> atoms are also detected, as noise can be reduced by accounting for these atoms (Sullivan et
al., 2001, p. 54). In, NIST-F1, the |4,0> atoms are detected first by using a standing wave that
pumps atoms from F=4 to F=5 as shown in figure #. A mirror, an optical telescope, and a silicon
Hayden Page 29
photodiode comprise a fluorescence detection system that quantifies the number of photons
released. A traveling wave is present
immediately under the standing wave.
It removes the small number of |4,0>
atoms that continue that are not pumped
by the standing wave (Jefferts et al.,
2002, p. 323). The remaining |3,0>
atoms are detected in a similar process
which involves repumping the atoms
into F=4 and detecting the fluorescence.
7. Conclusion
NIST is working on a new standard, NIST-F2, which is ultimately slated to replace
NIST-F1 as the nation’s primary frequency standard. It has a cryogenic vacuum structure that
cools the microwave cavity and flight tubes to temperatures of 77 Kelvin (Lombardi, Heavner,
& Jefferts, 2007, p. 86). The hope is to reduce blackbody frequency shifts, which involves a
transfer of energy from the walls of the apparatus to the cesium atoms. In addition, NIST-F2
utilizes a slightly different fountain technique. Instead of propelling a single ball of cooled
cesium atoms, it will rely on a “multiple ball toss scheme,” where as many as 10 balls of cooled
cesium atoms are interrogated one after another. This helps to eliminate a source of uncertainty
called a density shift, which occurs in NIST-F1 (Lombardi, Heavner, & Jefferts, 2007, p. 86).
Another promising aspect of new clocks is the use of optical transitions rather than the
microwave transitions used in NIST-7 and NIST-F1. Since optical transitions are on the order of
105 higher in their resonance frequencies, they result in much higher quality factors. These in
turn lead to more stable clocks as discussed in section 2. In the past, technology for precisely
determining these frequencies did not exist. This barrier has been eliminated with femtosecond
frequency combs, which is a technique that uses fluorescence signals to observe optical
frequencies at uncertainties comparable to those achieved in traditional beam and fountain
standards (Gerginov, Tanner, Diddams, Bartels, & Hollberg, 2005, p. 1735).
Currently, an optical clock using strontium is among one of the best clocks available.
The instrument which is operated and maintained at JILA, an interdisciplinary collaborative
Hayden Page 30
project between NIST and the University of Colorado, is capable of operating at or below the
uncertainty level of 10-18
(Ludlow et al., 2008, p. 1806). It uses a lattice technique that traps
atoms before interrogation. This and other techniques such as Penning traps, magneto-optical
traps, and chip-scale atomic clocks are driving much of the current research.
In just under 100 years, uncertainty in time-related measurements has decreased by at
least 8 orders of magnitude. It is difficult to determine exactly how accurate the best clock is
today, since the progress seems to be occurring so quickly. In fact, many have compared
research in this area to hitting a moving target. Rapid progress in timekeeping, as a result of the
work of many renown contributors in multiple fields within physical science, is a perfect
example of advancements in science leading to dramatic lifestyle changes for ordinary people.
For this reason, the topic is exigent and relevant for everyone from experienced scientists to high
school learners.
Hayden Page 31
References
Arditi, M. (1982). A caesium beam atomic clock with laser optical pumping, as a potential frequency
standard. Metrologia, 18, 59-66.
Ball, D. W. (2007, January). Atomic clocks: An application of spectroscopy. Spectroscopy, 21(1).
Bergquist, J. C., Jefferts, S. R., & Wineland, D. J. (2001, March). Time measurement at the millennium.
Physics Today, 37-42.
Crowell, B. (2006). Bandwidth [Data file]. Retrieved July 24, 2009, from http://en.wikipedia.org/wiki/
File:Bandwidth2.png
Dennis, S. K. (2007). Cesium metal [Data file]. Retrieved July 15, 2009, from Wikipedia Web site:
http://upload.wikimedia.org/wikipedia/commons/2/2a/Csmetal.jpg.jpg
Diddams, S. A., Bergquist, J. C., Jefferts, S. R., & Oates, C. W. (2004, November 19). Standards of Time
and Frequency at the Outset of the 21st Century. Science, 306, 1318-1324.
Gerginov, V., Tanner, C. E., Diddams, S., Bartels, A., & Hollberg, L. (2004). Optical frequency
measurements of 6s 2S1/2–6p 2P3/2 transition in a 133Cs atomic beam using a. Physical Review
A, 70(042505), 1-8.
Gerginov, V., Tanner, C. E., Diddams, S. A., Bartels, A., & Hollberg, L. (2005). High-resolution
spectroscopy with a femtosecond laser frequency comb. Optics Letters, 30(13), 1734-1736.
Itano, W., & Ramsey, N. (1993, July). Accurate measurement of time. Scientific American, 56-65.
Jefferts, S. R., Shirley, J., Parker, T. E., Heavner, T. P., Meekhof, D. M., Nelson, C., et al. (2002).
Accuracy evaluation of NIST-F1. Metrologia, 32, 321-336.
Hayden Page 32
Jespersen, J., & Fitz-Randoph, J. (1999, March). From sundials to atomic clocks: Understanding time and
frequency (Monograph No. 155). Washington, DC: National Institute of Standards and
Technology.
Jones, T. (2000). Splitting the second. The history of atomic time. Bristol and Philadelphia: Institute of
Physics Publishing.
Kamas, G., & Lombardi, M. A. (1990, September). Time and frequency users manual. Washington: US
Government Printing Office.
Levine, J. (2002, July). Time and frequency distribution using satellites. Reports on progress in physics,
65, 1119-1164.
Lombardi, M. A., Heavner, T. P., & Jefferts, S. R. (2007, December). NIST primary frequency standards
and the realization of the SI second. The Journal of Mesurement Science, 2(4), 74-89.
Ludlow, A. D., Zelevinksy, T., Campbell, G. K., Blatt, S., Boyd, M. M., de Miranda, M. H. G., et al.
(2008, March 28). Sr lattice clock at 1 × 10−16 fractional uncertainty by remote optical
evaluation with a Ca clock. Science, 319, 1805-1808.
Major, F. G. (2007). The quantum beat: Principles and applications of atomic clocks (2nd ed.). New
York: Springer.
Markowitz, W., Hall, R. G., Essen, L., & Parry, J. (1958, August). Frequency of cesium in terms of
ephemeris time. Physical review letters, 1(3), 105-107.
Metcalf, H., & van der Staten, P. (1994, August). Cooling and trapping of neutral atoms. Physics Reports,
244(4), 203-386.
Hayden Page 33
Millman, S., & Zacharias, J. R. (1937). The signs of the nuclear magnetic moments of Li7, Rb85, Rb87,
and Cs133. Physical Review, 51, 1049-1052.
Müller, S. T., Magalhães, D. V., Bebeachibuli, A., Ortega, T. A., Ahmed, M., & Bagnato, V. S. (2008).
Free expanding cloud of cold atoms as an atomic standard: Ramsey fringes contrast. Journal of
the Optical Society of America B, 25(6), 909-914.
NASA. (2009). Voyager location in heliocentric coordinates. Retrieved from http://voyager.jpl.nasa.gov/
science/Vgrlocations.pdf
National Institute of Standards and Technology. (2009). Time and frequency from a to z: An illustrated
glossary. Retrieved June 28, 2009, from http://tf.nist.gov/general/glossary.htm
National Physical Laboratory. (2009). Essen’s Caesium Clock [Data file]. Retrieved from
http://www.npl.co.uk/upload//img_200/atomic_clock.jpg
Nave, C. R. (2006). Quantum physics. In Hyperphysics. Retrieved July 15, 2009, from Georgia State
University Department of Physics and Astronomy Web site: http://hyperphysics.phy-astr.gsu.edu/
hbase/quacon.html#quacon
NIST. (2009). Historical context of the SI. Retrieved August 7, 2009, from http://physics.nist.gov/cuu/
Units/history.html
NIST-F1 cesium fountain atomic clock [Physics laboratory: Time and frequency division]. (n.d.).
Retrieved July 20, 2009, from National Institute of Standards and Technology Web site:
http://tf.nist.gov/cesium/fountain.htm
NIST Physics Laboratory. (2001). Research highlights. Retrieved August 9, 2009, from
http://physics.nist.gov/Divisions/Div840/PLflyer/time.html
Hayden Page 34
NIST virtual museum. (2009). Retrieved July 24, 2009, from http://museum.nist.gov/realmuseum.html
Provechito, J. (2004). The very large array at at Socorro, New Mexico, United States. [Data file].
Retrieved from http://upload.wikimedia.org/wikipedia/commons/6/63/
USA.NM.VeryLargeArray.02.jpg
Ramsey, N. F. (1986). The successive oscillatory field method and the hydrogen maser. In 40th Annual
Symposium on Frequency Control.
Sesko, D. W., & Wieman, C. E. (1988). Observation of the cesium clock transtion in laser-cooled atoms.
Optics Letters, 14(5).
Sullivan, D. B., Bergquist, J. C., Bollinger, J. J., Drullinger, R. E., Itano, W. M., Jefferts, S. R., et al.
(2001). Primary atomic frequency standards at NIST. Journal of Research of the National
Institute of Standards and Technology, 106(1), 47-63.
Teles, F., Magalhaes, D. V., Santos, M. S., Bebeachibuli, A., & Bagnato, V. S. (2002). Characterization
of the brazilian Cs atomic-frequency standard: Evaluation of major shifts. Metrologia, 39, 135-
141.
Topp, M. Personal communication. July 16, 2009
Time services department [Leap seconds]. (2003). Retrieved July 24, 2009, from United States Navy Web
site: http://tycho.usno.navy.mil/time.html
top related