Time and Frequency Division

Technical Highlights

• Mercury-Ion Microwave Frequency Standard. Using laser-cooled ¹⁹⁹Hg⁺ ions, staff of the Ion Storage Group have recently demonstrated operation of a microwave frequency standard. D. Berkeland, J. Miller, J. Bergquist, W. Itano, and D. Wineland have locked a microwave source to the ions’ ground state hyperfine transition at 40.5 GHz. In this work 7 ions contained in a cryogenic linear ion trap with a Ramsey free-precession time of 100 s were used to achieve a fractional frequency stability of 3 × 10^-13 τ^-1/2 and a reproducibility (uncertainty) of 3 × 10^-15. This rivals the performance of today’s best primary frequency standards. Still better performance is achievable using a larger number of ions (up to about 100 ions appears feasible). Trapped and cooled ions are useful for frequency standards for several reasons. The ions can be benignly held in a well-controlled environment, significantly reducing systematic effects such as Stark, Zeeman and Doppler shifts. Also, the same ions can be observed repeatedly and for very long times, so that narrow transitions can be observed with extremely high resolution.

  Since fractional frequency stability improves with increasing frequency, it is estimated that substantially higher performance can be achieved using an optical, electric-quadrupole transition at 282 nm in the same ion. Figure 1 is an energy level diagram showing both the 40.5 GHz hyperfine transition and the 282 nm transition. Experiments on the optical frequency standard, using methods similar to those described above, have been initiated. (J. Bergquist)

Figure 1. Energy-level diagram for 199Hg+. The 194 nm, electric-dipole transition is used for both laser cooling and state detection. The two clock transitions are the hyperfine transition at 40.5 GHz and the electric-quadrupole transition at 282 nm.

• Studies of Micromotion in RF Traps. D. Berkeland, J. Bergquist, W. Itano, and D. Wineland of the Time and Frequency Division have recently developed a method for sensing micromotion velocity and have applied this to the ¹⁹⁹Hg⁺ microwave clock described above. This allows them to place an upper uncertainty limit on their determination of the second-order Doppler shift due to micromotion of 2 parts in 10¹⁷.

  Laser cooling can substantially reduce the time-dilation or second-order-Doppler shift in frequency standards. Unfortunately, for trapped ions, part of the ions' velocity depends only on trap fields and is not directly affected by laser cooling. In the Penning trap the velocity in the rotation can be precisely controlled and determined by phase locking the rotation of the ions to a low-frequency reference. In the Paul rf trap used for the ¹⁹⁹Hg⁺ microwave clock, the "micromotion" caused by the trapping rf fields is minimized by confining a "string" of ions along the axis of a linear Paul trap as first demonstrated at NIST. However, sensitive diagnostic techniques are needed to establish when this condition is achieved. By sensing micromotion velocity through induced changes in the ion fluorescence caused by the first-order Doppler shift, the micromotion velocity can be minimized by applying compensation potentials to the trap electrodes. This has been done by sensing the velocity in three non-coplanar directions using three separate laser beams. (D. Wineland)

• Miniature-Ion-Trap Development. In collaboration with J. Beall of the Electromagnetic Technology Division in EEEL, staff of the Ion Storage Group have constructed traps whose electrodes are formed from laser-machined alumina substrates upon which metal electrodes are deposited using lithographic techniques. The objective of this effort, involving C. Myatt, D. Wineland, and C. Monroe of the Ion Storage Group, is the fabrication of very small ion traps with very high dimensional tolerances.

  For frequency standard applications and experiments on quantum logic, it is desirable to strongly confine the ions. Basically, this requires large potentials and small trap electrodes. Potential strengths are limited by vacuum breakdown, whereas making small trap electrodes (< 0.5 mm spacings) with very good tolerances requires nonconventional machining techniques.

  This design also allows the incorporation of passive trap circuit elements (filters, etc.) using small surface mount components. A first trap has been constructed (≅ 200 µm spacings) and will be tested in the ⁹Be⁺ quantum-logic project. In parallel, new traps using the same construction techniques are being developed for the Hg⁺ frequency-standard experiments. (D. Wineland)
• Phase-Locked Rotation of Atomic-Ion Plasmas. Scientists in the Physics Laboratory have phase locked the rotation of strongly coupled nonneutral plasmas to a well-controlled rotating electric field. This work was performed in Boulder by P. Huang, J. Bollinger, T. Mitchell, and W. Itano of the Time and Frequency Division.

  The plasmas, consisting of up to 10⁶ positive beryllium ions, are contained and laser cooled in a Penning trap to near absolute zero where the ions form a crystal whose shape is elongated by a rotating electric field and then rotates as a rigid body in synchronization with this field. Figure 2 shows a top view of the ion crystal and the electrodes that generate the rotating electric field. Bragg diffraction peaks from the rotating crystal show that the crystal lattice can remain stable for longer than 30 min or 10⁸ rotations, and the rotation is phase locked to the rotating electric field without any slippage.

  Figure 2. Electrode system used to rotate the ion crystal. This view is along the B-field axis of the Penning trap. The crystal is distorted by the electric field as shown in the figure. With alternating (sine-wave) electric fields applied with 120º phase differences to the three pairs of opposing electrodes, the electric field rotates and the ion crystal rotates in synchronization with the field.

  While the methods promise possible improvement for studies of plasma crystallization, the most significant practical implication is for future atomic frequency standards. The magnitude and uncertainty of the second-order Doppler shift for ions stored in a Penning trap places a severe limit on the accuracy of Penning-trap frequency standards. With this new method, the plasma can be phase locked at an optimum rotation rate (minimum value for the mean Doppler shift), and additional measurements should allow determination of the minimum Doppler shift to perhaps 1 %. Using 10⁶ ions, a standard operating in this mode would have a potential accuracy in the range of one part in 10¹⁶ to one part in 10¹⁷. (J. Bollinger)

• Cesium-Fountain Frequency Standard. The Division has made significant progress in the development of a cesium-fountain frequency standard. The standard is assembled and cesium atoms have been laser cooled, trapped, tossed upward, and detected as they fall back under the influence of gravity. The next phase of the project will involve operation of the microwave systems and preliminary assessment of the accuracy of the standard. The near-term objective is to achieve an uncertainty of at least one part in 10¹⁵, a factor of five better than the uncertainty of NIST-7. This development continues a progression in improvement of accuracy that has averaged better than a factor of ten every decade for the last 45 years, and assures that NIST will be able to maintain its metrology support for important timekeeping applications in navigation, telecommunications, and basic science.

  The majority of the work on this project has been done by D.M. Meekhof and S. Jefferts, but contributions have also come from R.E. Drullinger, L. Hollberg, W.D. Lee, C. Nelson, T. Parker, and F.L. Walls. The performance of this standard will be further enhanced by improved cooling methods that are being developed by W.M. Golding, W.M. Klipstein, W.D. Phillips and S.L. Rolston of the Atomic Physics Division.

  This new standard operates by launching laser-cooled cesium atoms vertically through a microwave cavity. Then, under the influence of gravity, they fall back through the same cavity allowing Ramsey-type interrogation without the usual end-to-end cavity phase shift. The atoms move more slowly than those in an atomic beam standard, so the Doppler shifts are much lower. Furthermore, the long observation time results in a narrower linewidth. Figure 3 shows the detected fluorescence of atoms that have been tossed to various heights above the trapping region of the standard and then fallen to a point about 10 cm below the trapping region where their fluorescence is detected. The width of the lines in this figure provides a measure of the vertical temperature of the atoms. The reduction in signal with increasing toss height is a measure of their horizontal temperature. For these data, both temperatures are about 6 µK. In normal operation, the atoms will be tossed vertically about 1.2 m. (S. Jefferts)

  Figure 3. Relative fluorescence as a function of time for various toss heights. At 33 cm, the signal is distorted, because it has just hit the top of the system. The atoms will be tossed much higher when the upper section of the standard is added.

• Improved Version of NIST-7. In an effort to upgrade NIST-7 and to add confidence to the evaluation of its systematic frequency shifts, NIST has joined with the Communications Research Laboratory (CRL) of Japan to build a copy of NIST-7 for CRL. This project, led by R.E. Drullinger, is fully funded by Japan, but involves staff of both NIST and CRL. W.D. Lee, C. Nelson, J. Shirley, and F.L. Walls are also contributing to the effort. The standard will remain in Boulder for comparison with NIST-7 for more than six months following its completion.

  Aside from providing NIST with the opportunity to compare two like standards assuring that the same results can be replicated, this project affords the opportunity to improve a number of the clock subsystems. For example, the pump laser system has been completely redesigned to be simpler and more robust, and optical fiber is now used to transmit laser beams to the state-preparation and state-detection regions of the standard. Also, the software for all control systems, including the digital servosystem, has been completely rewritten using object-oriented-programming methods. As these subsystems are developed for the new standard, they are replicated to replace subsystems of NIST-7. Many of the improvements developed in this project are also being adapted for the new cesium-fountain standard. (R.E. Drullinger)

• Improved Rubidium Frequency Standard. A collaboration between R.E. Drullinger and F.L. Walls of the Time and Frequency Division in Boulder and guest researchers from both Switzerland and the People's Republic of China has produced a major improvement in the short-term stability of a rubidium-cell frequency standard. With the short-term stability now within a factor of three of the stability of commercial hydrogen masers, this new class of rubidium standards should meet the growing need for a higher performance local oscillator in the next generation of laser-cooled, primary frequency standards. Hydrogen masers can also meet this requirement, but they are much more complex and costly. These new frequency standards can be miniaturized and should be sufficiently low cost that they might find application in such areas as telecommunications synchronization.

  This success is a result of several changes in design and careful attention to noise within the optical pumping system and the microwave frequency synthesizer. One of the key changes involves the use of a diode laser, rather than a discharge lamp, for optically pumping the rubidium cell. This reduces noise associated with the broad spectrum of pump light generated by a discharge lamp. However, amplitude and frequency noise on the diode-laser output, and phase-modulation noise on the synthesized microwave probe, had to be suppressed substantially to achieve the improved performance. This microwave-probe noise was carefully analyzed, and it was found that noise well beyond the modulation frequency contributes to system noise. This analysis guided the noise suppression used to achieve improved performance. The present stability of 3 × 10^-13 at 1 s is an order of magnitude better than that of the best conventional rubidium-cell standards available today. The use of a wall-coated cell and a new modulation scheme will lead to further improvement in performance. (R. Drullinger and F.L. Walls)

• Improved AT1 Time Scale. J. Gray, J. Levine, and T. Parker have improved the stability of the AT1 time scale through the addition of higher performance clocks and adjustment of time-scale parameters. AT1 is the basic time scale generated by appropriately averaging the outputs of all clocks in the time scale. UTC(NIST) is derived from AT1 through steering to international UTC. The performance of AT1 is thus critical to all aspects of the Division's work.

  During the last year two hydrogen masers and one high-performance cesium standard were added to the system so that it now contains a total of three masers and four high-performance cesium standards. In adding the masers to the time scale, work had to be done on methods for treating maser frequency drift, an issue that is not relevant for cesium standards. Figure 4 shows the stability of the current time scale relative to its stability a year ago. This improvement in time-scale performance will be particularly useful in the evaluation of the performance of the new cesium-fountain frequency standard. (T. Parker)

  Figure 4. Stability of the AT1 time scale as a function of averaging time.

• Environmental Sensitivities of Hydrogen Masers. In order to better understand the short-term and medium-term stability of the NIST time scale, T. Parker has done a study on the environmental sensitivities of the hydrogen masers that now have a major share of the weight in the time scale. This study is important because the time scale serves as reference for all Division research and all services offered by the Division.

  These studies were done on the effects produced by varying temperature, relative humidity, atmospheric pressure, line voltage, and magnetic field. The temperature and humidity studies were performed simply by changing set points on the environmental chambers housing the masers. The line-voltage and magnetic-field studies were similarly straightforward. However, it was more difficult to determine the effects of pressure on performance. This required construction of a special environmental chamber that could be pressurized while maintaining good control of both temperature and humidity. For realistically encally encountered variations in line voltage, humidity, pressure and magnetic field, perturbations of the frequency for these masers were found to be less than 1 × 10^-16 out to a period of 20 days. At this level, there is no noticeable impact on the performance of the time scale. However, the temperature fluctuations within the environmental chambers, at a level of about 50 mK rms, produce deviations of about 3 × 10^-16 at 20 days. As can be seen in Fig. 4 above, this is a noticeable effect, but still not much larger than the size of the dots in the figure. While environmental fluctuations do not appear to contribute significantly to instability, it is clear that very careful attention must be paid to the reliability of temperature control to assure long-term stability. (T. Parker)

• Multichannel GPS Receiver for Common-View Time Transfer. J. Levine of the Time and Frequency Division has developed a multi-channel, GPS, common-view receiver based on a commercial GPS "engine." The system runs under the control of a simple PC with an internal board used as a counter. One objective of this project is to replace the original (1985) NBS common-view receivers that are still in use by a number of laboratories, since these are becoming increasingly difficult to maintain. A second objective is to move to multichannel technology and a higher level of automation for data handling. The increased volume of comparison data provided by the multichannel technology should somewhat improve the precision of comparisons, and the improved automation will allow handling of a larger number of comparisons. Since common-view time transfer continues to be the de facto standard for international time coordination and atomic-clock performance continues to improve, it is important to continue pushing the performance of this method.

  Figure 5 shows a preliminary common-clock test of the receiver. With two receivers in the same location and connected to the same clock, the common-view difference becomes a very good measure of the receiver performance. The variance for these data is 2.5 ns, somewhat better than is achieved with single-channel receivers. (J. Levine)



Figure 5. Common-clock test of the multichannel receiver. Each point represents a 5 min average of the difference between collocated receivers referred to a common clock and using all space vehicles in view. The fixed offset simply reflects a fixed differential delay in coaxial cables.

• Time Transfer Using GPS Carrier Phase. In collaborative experiments designed to test a potentially more precise approach to international time coordination, J. Levine of NIST and K. Larson of the University of Colorado have been using common-view, GPS carrier phase methods to compare clocks at NIST in Boulder and the U.S. Naval Observatory (USNO) in Washington, D.C. While it is currently cumbersome to resolve the ambiguity in such measurements (both sites must be looking at the same cycle of the carrier), the problem will likely become easier as more laboratories participate in the comparisons, since multiple data sets can be processed to determine the ambiguity. One advantage of the common-view, carrier-phase method is that it is not limited by the deliberate, pseudo-random GPS clock variations associated with Selective Availability (SA). In addition, the fact that the carrier has a frequency about 1000 times that of the C/A code gives this method a theoretical advantage over traditional code-based designs.

  Short-baseline experiments used two commercial geodetic receivers at NIST with antennas separated by about 40 m. The receivers were driven by two different clocks in the time scale and the RMS difference between the carrier-phase data and the time-scale data was 125 ps. Preliminary evaluations of long-baseline experiments using receivers located at NIST and USNO yielded comparable results. (J. Levine)

• Improved Internet Time Distribution. Substantial improvements in software used for receiving time signals from the Time and Frequency Division's Internet Time Service have recently been made by J. Levine. This Physics Laboratory service, now receiving more than 2,000,000 calls per day, involves 10 servers operating in Boulder and on the east and west coasts of the United States. The new client software employs a method, called the "interlock method," that improves performance by better estimating the Internet path delay. Typical performance is 1 ms to 2 ms or better over long network paths and 100 µs to 200 µs over local or short-haul networks. This software requires some manual tuning to obtain the best possible performance, and is considered as an intermediate step in the development of more powerful software, using a method called "autolock," which will provide automatic characterization of the performance of the local clock and the network delay. This software, which will be released shortly, eliminates the manual tuning, thus achieving near optimum performance in a manner that is transparent to the user. (J. Levine)
• Progress on WWVB Upgrade. Staff at the Division's Fort Collins broadcast site, headed by M. Deutch and assisted by D.W. Hanson, J.P. Lowe, and A.J. Clements, have completed the tasks required to implement the first phase of the upgrade of radio station WWVB. The low-frequency signal broadcast by this station is used throughout the continental United States as a frequency reference and to automatically set clocks. The key objective of the upgrade is to increase the power output allowing for dramatic simplification and reduction in cost of receiver systems. This is expected to stimulate the development of products ranging from radio-controlled watches and home clocks to simple frequency standards that can provide intrinsic time-base accuracy to a range of electronic instrumentation.

  The completed system will involve two higher-power transmitters delivering signals phase coherently to two existing antenna arrays. Substantial modifications and upgrades of the antennas, antenna matching networks, and transmission lines are required. By the time this report is published, the first of the two systems should be in operation resulting in a factor of two increase in broadcast power. Once modest reliability of this system is assured, work will commence on the second transmitter/antenna system. A major objective of this upgrade program is an increase in reliability of operation of these broadcasts. Planned improvements that will assure a higher level of reliability are expected to take several years to complete. (W. Hanson)

• IEEE Standard Definitions for Time and Frequency. NIST has played a leadership role in the revision of IEEE Standard 1139-1988, Standard Definitions of Physical Quantities for Fundamental Frequency and Time Metrology - Random Instabilities. The IEEE standards committee was led by E.S. Ferre-Pikal of the Time and Frequency Division and included F.L. Walls of the Division as well as representatives from the Aerospace Corporation, Hewlett-Packard Company, EG&G Inc., Timing Solutions Corporation, the U.S. Army Communications-Electronics Command, the Jet Propulsion Laboratory, the U.S. Naval Research Laboratory, and BIPM. The draft was largely written at NIST and accepted with only minor revision. This revision incorporates recent NIST work 1) on improving the confidence interval for measurements that are a substantial portion of the data length and 2) on quantifying the confidence intervals for fully overlapped measurements.

  IEEE Standard 1139-1997 provides the basis for specifications and acceptance testing for clocks, oscillators, and a range of other equipment used in timing applications and low-phase-noise systems such as are encountered in telecommunications, navigation, and high-performance radars. (E.S. Ferre-Pikal)

• Low-Noise, Regenerative Frequency Dividers. E.S. Ferre-Pikal and F.L. Walls of the Division have demonstrated regenerative frequency dividers that have exceptionally low noise and operate at frequencies up to 40 GHz. This may be the highest-frequency, direct electronic division achieved to date. Such high-frequency, high-performance dividers are important components in frequency synthesizers within atomic frequency standards and in frequency synthesis chains linking the microwave and optical regions. They might also prove useful in high-data-rate telecommunications systems.

  These dividers achieve good stability through use of a filter (within the regenerative loop) that rejects the image frequency, and their noise properties are insensitive to the loop phase. This makes them practical devices for application in the microwave region. Figure 6 compares the performance of these new dividers with other, lower-frequency devices. The regenerative divider noise is so low that it would support synthesis with a fractional frequency stability of 10^-18 for long averaging times. (F.L. Walls)

  Figure 6. Plot of the fractional frequency stability versus averaging time for three different frequency dividers. The regenerative divider described here can be seen to have much better noise performance than either gallium-arsenide or emitter-coupled-logic (ECL) dividers.

• PM and AM Noise Measurement Using Carrier-Cancellation Techniques. F.L. Walls of the Time and Frequency Division has recently analyzed the advantages and disadvantages of using carrier-cancellation techniques for measuring PM and AM noise in oscillators, amplifiers and passive devices. While this is not a new concept, earlier work was hampered by a lack of suitable rf and microwave components. The development of needed components prompted him to revisit and evaluate the potential of the technique. He has now shown that for many measurements this technique can achieve good accuracy more quickly than conventional single-channel or two-channel techniques.

  Although the method requires tighter control of the average phase and amplitude of the input signals, it might prove to be advantageous in production measurements where measurement time is of critical concern. It should be especially useful for making AM and PM noise in amplifiers. Despite these benefits, to obtain unbiased measurements of PM noise in oscillators, two-channel cross-correlation measurements against two independent references are still required. (F.L. Walls)

• PM Noise in Phase Detectors. In collaboration with F. Garcia-Nava of the Centro Nacional de Metrologia in Mexico, E.S. Ferre-Pikal and F.L. Walls of the Division have measured the PM noise floor of a number of phase detectors for both homodyne (same-frequency) and heterodyne (different-frequency) detection. These measurements then guided them in their development of a model for noise in the commonly used phase detectors.

  In the homodyne case, they found that the baseband noise added to the noise floor. They demonstrated that for most mixers the low frequency (1/f) portion of the noise was 3 dB to 6 dB lower in the heterodyne than in the homodyne configuration. Their results should have impact on the design of nearly all low-phase-noise equipment used in applications such as frequency synthesis, telecommunications and radar. Since low-noise, high-frequency homodyne detectors are more expensive than heterodyne detectors operating at the same frequency, the circuit designer can reduce cost and maximize performance by using heterodyne techniques as much as possible within higher-frequency parts of a circuit. (F.L. Walls)

• Low-Frequency PM Noise in Commercial Amplifiers. In a recent collaboration with M. Delgado Amburo of the Centro Nacional de Metrologia in Mexico, E.S. Ferre-Pikal, H. Ascarrunz, and F.L. Walls surveyed ten different commercial amplifiers to assess their level of baseband (1/f) phase noise. Based on their earlier work on the generation of 1/f noise in amplifiers, they expected to find a dependence of the noise on the bias current. Their measurements indeed confirm that in general those amplifiers drawing more channel current had lower 1/f noise. But they also found that amplifiers using heterojunction bipolar transistors (HBT) had lower 1/f noise than amplifiers of the similar configuration using traditional silicon bipolar transistors. Furthermore, they showed that an amplifier that incorporates a linearizing technique to reduce intermodulation products has 1/f PM noise at least 20 dB below that of amplifiers of conventional design.

  This work provides additional confirmation of earlier NIST studies on 1/f noise generation processes in rf amplifiers. It should provide useful guidance to both the designers of rf amplifiers and those selecting amplifiers for low-phase-noise applications. (F.L. Walls)

• Calcium Optical Frequency Standard Using All Diode Lasers. A laser-cooled, calcium optical frequency standard has been recently developed and tested in a preliminary way by C. Oates, F. Bondu, and L. Hollberg of the Time and Frequency Division. This work supports efforts by the Precision Engineering Division of the Manufacturing Engineering Laboratory (MEL) to develop improved realizations of the meter. The calcium resonance, being of very narrow linewidth (400 Hz), provides less uncertainty in the determination of optical frequency than the more conventional systems based on transitions in iodine. The calcium transition is also very insensitive to electric and magnetic field. The frequency of this transition has been measured at PTB and is now recognized as providing the highest-accuracy reference for realization of the meter.

  The calcium system uses a frequency-doubled diode-laser system at 423 nm to laser cool and trap calcium atoms in a magneto-optic trap (MOT). The trap captures about 10⁷ cold calcium atoms in about 20 ms. The cold atoms are then probed on the narrow "clock" transition at 657 nm with a diode laser using the method of time-domain, optical Ramsey interrogation. Ramsey fringes as narrow as 6 kHz (corresponding to a resonance Q of ~10¹¹) have been observed with a good signal-to-noise ratio.

  Since the diode-laser systems are small, such a calcium standard can also be quite portable providing for practical intercomparison of optical frequency. In fact, NIST is already participating in the first international intercomparison of calcium frequency references. The results of comparison should be available soon. (L. Hollberg)

• A Microwave Optoelectronic Oscillator. J. Kitching, E.S. Ferre-Pikal, L. Hollberg, and F.L. Walls of the Time and Frequency Division have demonstrated low-noise oscillation at 1 GHz and 8 GHz in a new class of optoelectronic oscillator (OEO). While these oscillators are still in an early phase of development, their performance (see Fig. 7) is already superior to that of all conventional, competing technologies. Furthermore, these can be scaled to operate at still higher frequency, and their noise performance should improve with continuing development. OEO's have potential application wherever high-performance microwave oscillators are used, including frequency synthesizers, atomic frequency standards, and radars.

  The optoelectronic oscillator is made up of a long length of optical fiber (about 1 km) on a spool, a diode laser that launches a modulated signal into the fiber and a photodetector that receives the signal. The received signal is amplified, filtered, and sent back to modulate the diode laser, thus providing gain and closing the loop. A key element of the design is the filter that suppresses all but the desired mode in the fiber. For the 8 GHz oscillator, the modes are separated by 90 kHz, and a filter in the optical path reduces the nearest side modes to 50 dB below the carrier. (L. Hollberg and F.L. Walls)

Figure 7. Noise performance of the optoelectronic oscillator. The OEO is compared with both a yttrium-iron garnet (YIG) oscillator and a dielectric-resonator oscillator (DRO). The apparent flattening of performance beyond 10 kHz is simply a reflection of the noise floor of the measurement system. The actual performance certainly lies below this line.

• Simplified Optical-Frequency Tripler. A simplified optical-frequency tripler has been demonstrated in a collaboration between staff of the Time and Frequency Division and scientists at the University of Colorado, Rice University, and Lightwave Electronics Corporation. NIST staff involved in this effort included J. Wells, L. Hollberg, L.R. Zink, and D. Van Baak. This success is built on a new nonlinear optical crystal, periodically poled lithium niobate (PPLN), where a number of special phase-matching conditions were discovered. Their experimental demonstration involved a simultaneously phase match of second-harmonic generation (SHG) and a sum-frequency mixing process (SHG plus fundamental) to produce the third harmonic. They point out that there are 11 other wavelengths between 1 µm and 4 µm where similar matching conditions exist. In their experiments they tripled the frequency of a CO laser at 3.56 µm to obtain useful output signal levels at 1.19 µm. With an input power of 195 mW, they observed an output of 7.3 nW using only a single pass through a 2 cm long PPLN crystal. Figure 8 shows the third-harmonic power as a function of wavelength for several CO overtone frequencies.

  Figure 8. Third-harmonic-generation (THG) power for several CO overtone frequencies. The width of the signals is the resolution of the optical spectrum analyzer, fixed by the diameter of an optical fiber used to couple the output to the analyzer.

  This appears to offer an especially important means for constructing certain stages of an optical-frequency-synthesis chain for coherently linking the cesium-atomic-clock frequency in the microwave region to the frequency of future optical clocks (such as Hg⁺) in the visible. In their paper on this topic the group also notes 10 special phase match points where second harmonic generation is cascaded twice to yield fourth-harmonic generation. Such processes could also contribute to a synthesis chain. (L. Hollberg)
• Optical Coherences in Dense Atomic Media. L. Hollberg and H. Robinson have collaborated with scientists from Texas A&M University, the University of Munich, the Max-Planck Institute, and the Lebedev Institute in the study of optical-coherence effects in a dense rubidium vapor. They discovered narrow resonances that result from interferences in stimulated Raman transitions in rubidium vapor. These effects drastically alter the observed lineshape and make it more difficult to find the true center of the atomic line. This has detrimental implications for high-accuracy measurements in certain types of optical detection in rubidium atomic clocks. However, because the coherence-induced interferences are spectrally sharp, they are potentially useful for higher resolution spectroscopy and possibly higher detection sensitivity for other applications. The results of these studies were recently published in the Physical Review Letters. (L. Hollberg)
• New Far-Infrared Lasers Lines. In collaborations with scientists from the Universidade Estadual de Campinas (Brazil), Meiji University (Japan), and the University of New Brunswick (Canada), K.M. Evenson and L.R. Zink have discovered and measured approximately 300 far-infrared lasing lines in hydrazine, difluoromethane, and methanol. Many of these will extend the range of usefulness of laser-magnetic-resonance (LMR) spectroscopy since they span LMR spectral regions where there is a scarcity of laser lines. These new laser lines are reported in recent publications.

  More than 20 new lines were reported for ¹³CH₃OH methanol pumped by a continuous-wave CO₂ laser. For ¹²CH₂F₂, 13 new far-infrared lines were reported, and 7 new lines were reported for ¹³CH₂F₂, both molecules again being pumped by a CO₂ laser. For N₂H₄, 14 new far-infrared lines were discovered when pumping with an N₂O laser. In pumping this gas with a CO₂ laser, only 30 lines had been previously reported, and these were all near and above 200 µm in wavelength. The work described here adds 133 new lines with 83 % being shorter than 200 µm. The shortest of these was 49.2 µm. Major improvements in the CO₂ and N₂O lasers have been the key to these new observations. (K.M. Evenson)

• New Spectral Observations Using LMR. K.M. Evenson and L.R. Zink of the Division, along with J. Brown and F. Tomassia of Oxford University, have directly measured the frequencies of fine structure lines in Al, F⁺, Fe⁺, and FO using laser-magnetic-resonance (LMR) spectroscopy. The most significant aspect of this work involves the atomic lines, the frequencies of which are needed by radio astronomers for spectral searches and identification. These measurements improve the accuracy of the atomic-line frequencies by more than 1000 times and provide heretofore unavailable information on hyperfine parameters of these atoms. The FO measurements are significant in that they involve magnetic-dipole transitions that are very difficult to observe with other methods, but which appear with great sensitivity and resolution using LMR. This provides a strong indication that the method will prove successful in observing magnetic-dipole transitions in a number of other, important molecules. (K.M. Evenson)
• Far-Infrared, Frequency-Synthesis Spectroscopy to 9 THz. L.R. Zink and K.M. Evenson of the Division, in a collaboration with H. Odashima of Toyama University, have succeeded in extending the range of far-infrared, frequency-synthesis spectroscopy from 6 THz to 9 THz. This region of the spectrum is a particularly difficult one in which to work, and yet it covers absorptions by some especially important molecules, including ClO, which plays a significant role in the chemistry of the upper atmosphere. In this spectroscopic method, the probe radiation is synthesized as the difference frequency between radiations from the ammonia laser and the carbon-dioxide laser. (K.M. Evenson)
• Sub-Doppler Stabilization of a Far-Infrared Laser. Sub-Doppler stabilization of a far-infrared laser to its own gain curve has been achieved for the first time in collaborative work between K.M. Evenson and L.R. Zink of the Time and Frequency Division and E. Telles of the Universidade Estadual de Campinas in Brazil. Application of this method will lead to improved stability of FIR lasers by about a factor of 50, thus improving the signal-to-noise ratio and accuracy of FIR spectroscopy, which is an important tool for laboratory measurements on molecules of importance in space and the upper atmosphere. It will also improve the stability of far-infrared lasers used as local oscillators in radio astronomy. (K.M. Evenson)

Most Recent Technical Activities   |   Archive of Technical Activities