Time and Frequency Division

Technical Highlights

• Accuracy of NIST-7. The Atomic Beam Standards Group has now completed extensive evaluation of the systematic errors of NIST-7. The results shown in table 1 clearly establish it as the most accurate primary frequency standard in the world. The traditional approach to combining these errors indicates a combined uncertainty of 6 × 10^-15.
  
  

  Table 1. Systematic Errors in NIST-7

  Effect	Bias (× 10-15)	1 σ Uncertainty (× 10-15)
  Second-order Zeeman	≈ 100,000	< 1
  Second-order Doppler	≈ 350	3
  End-to-end phase shift	733	4
  Distributed phase shift	—	< 1
  Cavity pulling	7	< 1
  Fluorescence light	—	< 1
  Line overlap	< 0.1	< 1
  Electronics	—	< 3

  To further assure confidence in the evaluation process, the Group has, wherever possible, evaluated each systematic error using at least two different methods. For example, the velocity profile needed to calculate the second-order Doppler shift was measured through inversion of Ramsey patterns taken at different power levels. It is also measured using a pulsed laser method. In every case, excellent agreement between methods has been achieved. Particular emphasis was also placed on evaluating errors due to cavity pulling, magnetic-field inhomogeneity, and Rabi pulling.

  While the root-mean-square uncertainty indicated above should be used to compare the performance of this standard with other world standards, the Group points out that the use of this method could result in an estimate of uncertainty that is too optimistic. This is because the various systematic errors may not be completely independent. That is, some measurements are used to establish more than one of the errors. In this situation, there will likely be some correlation, so caution must be used in estimating the total combined uncertainty. It is clear that it can be no worse than the simple sum of errors. This same concern must be addressed by other laboratories operating primary frequency standards. (R. Drullinger)
• Automation of Primary Standards. David Lee, John Lowe, Fred Walls, Jon Shirley, and Bob Drullinger have developed substantial automation and measurement techniques that will change the way that NIST-7 and all future NIST primary standards are operated. At the heart of this development is a digital servo-control system. While the primary function of this system is to lock a microwave source onto the cesium resonance, its frequency agility under computer control allows for automated control and evaluation of a number of systematic errors. In addition, David Lee has generated substantial computer routines that permit automated operation, evaluation, and remote monitoring of subsystems and the evaluation process. The remote monitoring allows the staff to look in on the standard on weekends and nights (from home) resulting in more nearly continuous operation of the standard.
  
  The servo-control system locks onto the cesium resonance with a very long time constant, so the system can periodically check the frequency of one of the Zeeman lines. Since the frequencies of these lines are directly related to the strength of the C (magnetic) field, this information is used to servo control the C field. Figure 1 shows the effect of putting the C field under servo control. Between runs the system automatically takes scans of the Ramsey pattern at predetermined microwave power levels. These provide the basis for determining the atomic velocity profile, important input to evaluation of several systematic errors. As a result of this automation, full evaluation of the standard can be accomplished in two days, a vast improvement over the month-long evaluations of NBS-6. The group is now working on a concept for operating counter-propagating beams simultaneously, thus continuously measuring end-to-end cavity phase shift, an error that is now determined by several independent frequency determinations. This could reduce evaluation time to one day. The digital servo-control system is critical to the success of this effort. (D. Lee)
• Cryogenic Linear Ion Trap. Jim Bergquist, Wayne Itano, and Dave Wineland of the Ion Storage Group, along with former postdoc Martin Poitzsch, have observed linear "crystals" of tens of laser-cooled ¹⁹⁹Hg⁺ ions in a linear rf ion trap that operates at liquid helium temperature. Operation at low temperatures provides very high vacuum through cryopumping. This substantially reduces chemical-reaction-induced ion loss and frequency shifts caused by background gas collisions. Superconducting circuits are also employed to reduce power dissipation in the trap rf drive and magnetic-field correction coils. The system is designed for use as a prototype 40.5 GHz frequency standard with a potential accuracy of several orders of magnitude beyond that of the current best cesium beam standards (uncertainties ≅ 10^-14). (J.C. Bergquist)
• Raman Cooling. Members of the Ion Storage Group report experiments in which a bound atom is cooled to the zero-point energy using resolved-sideband Raman cooling. The cooling process involves an interaction between the vibrational states and the internal states of the atom through Raman transitions. These experiments set the stage for future work that is important in several contexts. First, it may be possible to exploit this interaction to create squeezed and other nonclassical states of particle motion. Second, if the cooling technique is applied to the center-of-mass of two or more ions, it may be possible to (1) prepare correlated internal atomic states which can reduce quantum noise in spectroscopy (or atomic frequency standards), or (2) construct quantum logic gates applicable to the realization of a quantum computer.
  
  In the experiments reported by the group, a single beryllium ion bound in a rf (Paul) trap is cooled to the zero-point energy. In one dimension the zero-point energy (n_x=0) is achieved 93% of the time. In three dimensions, the zero-point energy (n_x=n_y=n_z = 0) is achieved 58 % of the time. Further improvements in the system, including the application of additional cooling cycles, should improve the 3-D Raman cooling. (C. Monroe)
• Bragg Scattering from Laser-Cooled Atomic Ions. Members of the Ion Storage Group have observed Bragg scattering from relatively large samples (N ≅ 10⁵) of laser cooled ions which are confined in a Penning trap (nonneutral ion plasmas). A typical scattering image is shown in Figure 2. This marks the first demonstration of Bragg scattering from a laser-cooled atomic sample. The Bragg patterns can be transformed to give information on the local ordering within the plasma (pair-correlation function) and reveal details about the liquid and solid properties. The Bragg patterns can be observed in real time which also gives rather immediate and direct temperature information that is important for Doppler shift calibrations in atomic clocks. (J.J. Bollinger)

  

  Figure 2. Bragg scattering image observed with approximately 3 × 10^-4 Be⁺ ions. The first two Bragg peaks are visible. The ion density here was about 9 % of the Brillouin density which produced a scattering angle for the first peak of 1.5°. The square shadow near the edge of the image is from a wire mesh on the top of the upper experimental trap endcap. The dashed line outlines a beam stop.

• Coaxial Ion Trap for Strong Confinement. Staff of the Ion Storage Group have developed a coaxial-resonator driven rf trap capable of strongly localizing an ion within the trap. The trap is a variant of the quadrupole rf trap successfully used by the group for a variety of measurements over a number of years. The strong confinement allows the attainment of the Lamb-Dicke condition whereby the extent of the atoms motion is less than λ/2π where λ is the wavelength of the exciting radiation. Attainment of this condition is important because, among other things, it suppresses both the broadening due to the first order Doppler effect and the fluctuations of the "carrier" from measurement to measurement.
  
  With the attainment of strong confinement, the Group should be able to achieve resolved sideband cooling very rapidly using allowed electric-dipole transitions. This will provide for lower noise in spectroscopic studies and studies aimed at the development of advanced atomic frequency standards. (S.R. Jefferts)
• A New Timing Distribution Amplifier. Fred Walls and Craig Nelson of the Division, along with Marco Siccardi of the Technical University of Torino, have developed a timing distribution amplifier with exceptionally low environmental sensitivity. The amplifier, operating at 5 MHZ and 10 MHZ, provides 5 outputs for 1 input and is used for distributing timing signals within a measurement laboratory. The characteristics assure that the performance of the standard driving the amplifier is realized at the many application ports within the laboratory.
  
  The sensitivity to temperature shown in Figure 3 is typically less than 0.3 ps/K at the normal operating temperature. This results in a fractional frequency error of less than 3 × 10^-18/K for a measurement time of 1 day. The isolation from channel to channel and front to back is typically 120 dB. This isolation means that shorting or disconnecting one channel changes the phase of another channel by less than 33 fs. (F. Walls)
• Steered-Clock Algorithm. Fred Walls and Judah Levine have developed an improved method for steering the output of the NIST time scale. They combined the current algorithm (AT1) output with that of a hydrogen maser to realize both the long-term stability of the time scale and the excellent short-term stability of the maser. This is of great interest for two-way time transfer, since such transfer is more nearly real time and comparisons with other international time scales have been limited by the short-term noise of the NIST time scale.
  
  This system steers a phase micro-stepper to follow the combination of AT1 and maser. However, the micro-stepper also contributes short-term noise, so this is reduced by having the micro-stepper servo control an oscillator of high spectral purity with a time constant of several seconds. Such measures were not needed in the past, because time transfer measurements using GPS were averaged over periods ranging from hours to days, and the short-term noise of the time scale averaged to zero. The new system reduces the short-term noise by a factor of five. (J. Levine)
• Experimental Time Scale Using Masers. Jim Gray, Judah Levine, Tom Parker, Trudi Peppler, and Fred Walls have developed an experimental time scale that incorporates two hydrogen masers. Preliminary data indicate a stability of 7 × 10^-15 for τ = 1 day and 2.5 × 10^-15 for τ = 10 days. This is significantly better than the performance of the present AT1 time scale when only cesium frequency standards are used. The work is driven by the need to reduce short-term noise in the time scale and improve long-term predictability. The concern in using masers is their long-term drift. Special precautions must be taken to make sure that this drift does not adversely affect the performance of the time scale.
  
  The Division has owned one maser for several years, and a second maser was delivered this year. A third maser is on order. The availability of three masers will allow complete characterization of the frequency stability of the individual masers. This is critical to application of the masers in the time scale. The new system will be studied further before considering any switch from the present system. (J. Gray)
• Two-Way Time Transfer. Christine Hackman, Steve Jefferts and Tom Parker of the Time Scale and Coordination Group have partially completed a careful evaluation of the performance of the NIST 3.7  meter earth station used for two-way time transfer. A full characterization of this system is needed to assure optimum coordination of the NIST time scale with the BIPM and other international time scales. They are also gaining operational experience with the system through extensive two-way transfers with six sites in Europe and two in North America.
  
  They have completed an absolute calibration of the transmit and receive time delays to one nanosecond. This will contribute to the reduction of the uncertainty in absolute time maintained by NIST. They have also evaluated instabilities in the two-way equipment by performing measurements between two co-located stations referenced to the same clock. Typical results of these measurements are shown in Figure 4. Time transfer instabilities are flicker phase modulation in nature and are on the order of 200 ps to 300 ps for 2 day averaging. These measurements are continuing with a focus on identifying and eliminating (or reducing) sources of instability. However, certain common-mode environmental sensitivities may not be visible with this approach, so the group is initiating two-way time transfer measurements between two masers located at different sites. These should uncover instabilities not observed in the common-clock experiments. (C. Hackman)
• New Network Time Server. A second INTERNET time server was developed and installed by Judah Levine at the National Center for Atmospheric Research (NCAR) in Boulder. NCAR is one of the main nodes of the network, so this location is ideal. Current usage of the service is 8000 calls/day and the calling rate seems to be doubling about every two months. This service operates under a special NIST-developed algorithm that takes the properties of the net into account. A paper on the algorithm used in these systems was recently published.
  
  The original server is still in operation at NIST Boulder. It is currently handling more than 5000 calls per day. It has become clear that service demands will require establishing additional servers in other locations. (J. Levine)
• New ACTS Server. A new generation server for the Automated Computer Time Service (ACTS) has recently been demonstrated by Judah Levine. The server, based on standard PC hardware, will replace the original hardware-oriented system that has been in service for more than five years. It should be easier to maintain and operate, and should be readily adaptable to new computers as they evolve. The system is fully compatible with the existing servers that run at 300 bits/s and 1200 bits/s, but extends the capability of the service by handling all speeds up to 14.4 kbits/s.
  
  In a related project, Victor Zhang of the Time Scale and Coordination Group has developed a monitor system that assures integrity of all signals broadcast by ACTS. The monitor, also based on PC technology, takes inputs from all ACTS servers and regularly compares them with a time signal of completely independent origin. (J. Levine)
• Time Stamp Service. Judah Levine has completed development of an encrypted time-stamping method for computer-generated records. His system will permit authentication of dates of records important in patent applications as well as other documents which must remain proprietary. The development was funded through a special Pioneer Fund Grant from the Department of Commerce. The service is based on a "hashing" algorithm applied to a computer-generated file with the addition of an encoded NIST time stamp. The enciphered output is infeasible to duplicate without the original text, which can be produced by the owner in cases of disputes of dating. The method provides an electronic means of validating dates and maintaining confidentiality, with the eventual elimination of notaries and witnesses for records that can be converted to digital format. It is compatible with NIST standards for encryption including FIPS 180 and FIPS 186.

  The system is now undergoing testing, and discussions with management concerning the details of operation of the system as a regular NIST service have been initiated. The development was prompted by the realization that many NIST scientists are now keeping lab records on computer rather than a hard-bound notebook. This raised the question about validity of computer stored documents in resolving any claims about inventions. (J. Levine)

• Second-Generation Frequency Measurement Service. The Services Group has recently developed a second generation version of its Frequency Measurement Service. The new system is easier to use, more reliable, and assures higher accuracy traceability to NIST. It will provide an improved reference frequency for use by industry and other government agencies.

  The system was completely redesigned by Mike Lombardi and Lisa Nelson. They replaced the computer, receiver, and time interval counter, and rewrote all software. The reference signal is now taken from the Global Positioning System rather than LORAN-C. The system also generates two-sample (Allan) variance analysis of the performance of the users local standards, and it can perform calibrations with no other equipment required. (M. Lombardi)

• Evaluation and Upgrade of WWVB. Radio station staff in Fort Collins along with Wayne Hanson and Roger Beehler have embarked on a study of the WWVB transmitter system with the objective of increasing the power output, minimizing icing problems, and increasing reliability of operation. An increase in power output will enable U.S. manufacturers to develop mobile and consumer products that can receive signals anywhere in the U.S. At the present power output, only more sophisticated receivers can pick up the signal in certain urban areas on the east coast.

  The study was initiated by hiring a consultant to evaluate the condition of the current equipment and antennas and recommend improvements. Following extensive measurements, the preliminary report of the consultant indicates that present antennas are reasonably sound but need some improvement. The transmitters, matching networks, and balanced lines connecting transmitters to the antennas all require replacement or major rebuilding. The Division is now studying options for funding and implementing these recommendations. (R. Beehler)

• Preliminary Studies of Noise Conversion in RF and Microwave Devices. Eva Ferre Pikal and Fred Walls have identified two mechanisms that appear to explain much of the 27 year old mystery of up-conversion of baseband noise to produce AM and PM noise about rf or microwave signals. These are the limiting noise processes in many amplifiers and oscillators. This understanding should lead to improvements affecting a wide range of applications in electronic measurement and control. They have shown that these mechanisms depend on frequency, the specific transistor, and details of dc and rf portions of the circuit. Relatively small changes can improve the close-to-carrier PM and AM noise by as much as 20 dB. Only two circuit types have been studied, so substantial additional study is needed. There is also a need to extend the measurements to higher frequencies. (F. Walls)
• Training for Foreign Metrologists. The Division notes a recent surge in training visits by scientists from other countries. Three visitors are currently in the Division and three additional visitors are scheduled to arrive in 1995. The Division has also performed an extensive study for the Singapore Institute of Standards and Industrial Research (SISIR) outlining options for the development of time and frequency dissemination to government, commercial, and industrial interests in that country. This increase in activity is presumably tied to ISO 9000 requirements that must be met in order to have products accepted in international markets.

  Dave  Howe of the Services Group performed the study for SISIR traveling to Singapore to confer with officials to determine the exact nature of their requirements. His written report, which takes the size of the country into account, details available options along with their various advantages and disadvantages.

  Current visitors are from Ecuador and Saudi Arabia. Federico Puma from Ecuador is working under Fred Walls of the Phase Noise Measurements Group to develop time and frequency systems that will be used for metrology in that country. John Lowe of the Atomic Beam Standards Group traveled to the Saudi Arabian Standards Organization (SASO) in Riyadh to assist that organization with measurement methods and the use of GPS for international time coordination. He is responsible for the training of two SASO engineers, Khalid Salith, A. Al-Dawood, and Khalid Khamis Al-Dossary, who are currently spending 3 months in Boulder.

  The Division has agreed to host visitors from Mexico and Brazil during the coming year. Dr. J. Mauricio López and Eng. Francisço García of the Centro Nacional de Metrologia (CENAM) will each spend 9 months with the Division while Eng. Ricardo José de Carvalho of Brazils Conselho Nacional de Resenvolvimento Científico e Technólgico (CNPq) will visit for 3 months. (D.B. Sullivan)
• Noise in Diode Laser Spectroscopy. Hugh Robinson of the Optical Frequency Measurements Group has developed a computer model that reproduces very complex noise spectra observed using solitary (broad-linewidth) diode lasers tuned near atomic resonances. The computer results for rubidium shown in Figure 5 demonstrate typical output of his model. Measurements on rubidium closely mimic the fine detail of this plot. This indicates a very good understanding of the source of noise in such systems, a fact that should prove useful in a wide range of systems involving the interaction of diode-laser radiation with atoms.

  

  Figure 5. Modeled noise profile for rubidium probed by a solitary diode laser. f_N is the Fourier noise frequency (in GHz) and f_L is the frequency detuning (also in GHz) from the rubidium transition at 780 nm. The vertical scale is noise amplitude in arbitrary units. The theoretical model agrees remarkably well with experimental results.

  The theoretical model includes the effects of diode laser frequency noise, all hyperfine structure within the rubidium absorption, atomic saturation effects, and Doppler broadening. The effects observed and explained here are not specific to rubidium or to diode lasers, but will occur in any spectroscopy done with lasers that have excess frequency noise. (H. Robinson)
• Ultrasensitive Detection of Atoms and Molecules. During the last year, Richard  Fox of the Optical Frequency Measurements Group has demonstrated trace detection of materials important in water and the atmosphere using diode-laser systems. Precisely tunable diode lasers have tremendous potential for high sensitivity detection of specific atomic and molecular species. Because they are compact, relatively low cost, and operate with very high efficiency, they will be suitable for transportable instrumentation that could be used in field applications such as pollution monitoring, process monitoring/control, and even medical applications.

  In a continuing collaboration between the Group and NOAAs Aeronomy Laboratory, two molecular species important in atmospheric photochemistry, especially relating to atmospheric ozone, have been studied. NO₃ has been detected with red diode lasers operating at 662 nm and IO has been detected with 425 nm light generated by frequency doubling an infrared diode laser.

  In separate experiments, frequency doubled laser light has been used to detect trace quantities of Pb, Cr, and Ca in water. This method has the potential for both higher detection sensitivity and much faster detection than traditional methods. Both Pb and Cr levels in drinking and waste water are now monitored and controlled by EPA regulations. (L. Hollberg)

• Revised Frequency Standards for the Infrared Based on the CO₂ Laser. In a recent paper in the Journal of Molecular Spectroscopy, NIST researchers report substantially improved frequencies for the CO₂ molecule used to stabilize the CO₂ laser. This revision of the frequencies of 780 lines of this laser will serve as an important standard for frequency and wavelength measurements in the infrared region of the spectrum. The CO₂ molecule was also observed in absorption spectroscopy, so these frequencies serve as state-of-the-art standards for Fourier transform spectroscopy. A number of the measurements upon which this revision is based were made in the Time and Frequency Division in Boulder by Ken Evenson and Lyndon Zink of NIST and Che-Chung Chou and Jow-Tsong Shy of the Republic of China. Art Maki, a molecular theorist, now retired from the NIST Molecular Spectroscopy Division, performed the new least-squares fit of the data. Aside from the many measurements of CO₂ frequencies made at NIST, this revision relies critically on two highly accurate (absolute) measurements made in France and a revised value of the 88 THz methane frequency.

  The CO₂ laser now stands as the most used secondary frequency standard in the far infrared. This standard supports accurate spectroscopic measurements useful for identification of molecular species in space, in the upper atmosphere, and in industrial processes. These results combined with the development at NIST of a substantially improved CO₂ laser, puts NIST in an excellent position to support scientific and industrial measurements in this important spectral region. The new NIST laser operates in a continuous-wave mode on more laser lines than any other CO₂ laser ever built. (K.M. Evenson)

• Accurate FIR Spectral Measurements on HO₂. Ken Evenson and Lyndon Zink of the Laser Spectroscopy Group with Guests Kelley Chance of the Harvard Smithsonian Center for Astrophysics and K.J. Park of the University of Oregon have made accurate measurements of 17 transitions in the far infrared spectrum of the hydroperoxyl radical (HO₂). Combining these with previous microwave and millimeter wave transition frequencies, they have developed a revised set of molecular constants that yield a highly accurate set of transition frequencies for this important radical.

  HO₂ is one of the most abundant free radicals in our upper atmosphere, and its frequencies are very important in analyzing the far infrared spectrum observed at high altitude. HO₂ frequencies are also of considerable interest in radio astronomy for the possible detection of this species in interstellar space. (K. Evenson)

• IR Source Based on Difference Frequency Generation. In a paper recently accepted by the Journal of the Optical Society of America, researchers at Rice University and the Optical Frequency Measurements Group of NIST describe collaborative development of a system for generating tunable infrared radiation near 3 µm by difference-frequency mixing the outputs of two solid-state lasers. The specific experiments reported involve the mixing of radiation from a diode laser and a diode-laser-pumped Nd:YAG laser.

  They report an output of more than 2 µW of cw radiation in the range from 3.155 µm to 3.423 µm. This spectral region has not been accessible using simple, compact radiation sources, and it is of very high interest because it covers many important carbon-hydrogen stretch absorption lines. To demonstrate the spectroscopic capabilities, the group used their system to detect fundamental stretch vibration modes of methane by both direct and wavelength-modulation absorption spectroscopy. These experiments used high resolution scans of the difference frequency over about 30 GHz (~1 cm^-1).

  This type of radiation source should prove useful in a variety of applications including the trace detection of important organic molecules. There are very few sources available in this spectral region, and those that do exist tend to have poor spectral quality and be expensive and large. This work demonstrates the feasibility of one of a class of future sources exhibiting simplicity, low power consumption, small size, and wavelength tunability. The concepts will most certainly be extended to other frequency regions. (L. Hollberg)

• Optical Frequency Standard of Reduced Linewidth. Rich Fox and Leo Hollberg have achieved very high resolution in the observation of the narrow calcium inter-combination line at 657 nm. In their uncooled calcium-beam experiments, they demonstrated optical Ramsey fringe linewidths as narrow as 10 kHz. They also demonstrated one dimensional transverse cooling of this calcium atomic beam using 423 nm light generated by frequency doubling the output of a laser diode operating in the infrared. These experiments are important because the 657 nm calcium line is now internationally accepted for use in realizations of the meter. The calcium transition is particularly interesting because of its 400 Hz intrinsic linewidth. This is 4 orders of magnitude narrower than the more commonly used lines in iodine. Furthermore, the transition is not substantially perturbed by ambient electric or magnetic fields, even at the 10^-14 level.

  In the experiments, diode laser light is converted from the infrared to the blue with a conversion efficiency as high as 30 % while maintaining precise single-frequency tunability. More than 40 mW of blue light is produced. This is sufficient for three-dimensional trapping and cooling of calcium. The blue source will have numerous applications beyond the calcium project. (L. Hollberg)

• N₂O Laser. Ken Evenson and Guest Researcher Maki Tachikawa of the University of Tokyo have developed a practical 2 m N₂O laser that lases on more than 40 lines in the 9 µm and 11 µm regions of the spectrum. Typical output of this laser is shown in Figure 6. Lasing in the 9 µm region had been observed once in 1969, but with a laser tube that was 6 meters long. The observation of the 11 µm hot bands is the first ever for this gas. Frequency measurements of these lines will begin soon. The laser can be saturation stabilized to N₂O in a manner analogous to the way in which the CO₂ laser is stabilized. Hence, N₂O should soon become another infrared frequency standard similar to CO₂. The laser also serves as a pump laser for far infrared lasers; in fact, a new far infrared laser line in methanol pumped by a 9 µm N₂O line has already been discovered. (K.M. Evenson)