Quantum Physics Division

Technical Highlights

• Investigating the Properties of Bose-Einstein Condensates. More than seven decades ago Albert Einstein came to the startling conclusion that if a gas of atoms could be cooled to a certain extremely low temperature, the gas would undergo a condensation process in which a large percentage of the atoms in the sample would settle into the lowest energy state available to them. This process, called Bose-Einstein condensation (BEC), was first observed in a gas in the Spring of 1995 by NIST and CU scientists at JILA. While still maintaining many of the properties of a dilute vapor, the atoms nonetheless take on a coherent nature, a high degree of organization, which is very similar conceptually to the coherent, monochromatic, collimated nature of photons in a laser beam. There are now many research groups around the world capable of producing BEC, and internationally there has been considerable experimental and theoretical work done towards understanding the nature of this novel material. Experimental BEC work accomplished recently at JILA is in four categories:
  • Basic thermodynamics and energetics. A number of basic, quantitative measurements have been made including determinations of the critical temperature, measurements of the specific heat, and measurements of the interaction energy of the condensate. These measurements yielded no surprises; all results were in quantitative agreement with theoretical predictions. The results can be summed up by two basic observations: first, that as small as the condensate samples are, their properties are already indistinguishable from the large-sample limit, and second, that as dilute as the samples are, their properties are already profoundly changed by inter-atomic interactions.
  • Excitations (Sound Waves). Standing-wave density fluctuations have been produced in condensates. The frequencies of these excitations at near-zero temperature agree well with theoretical predictions. As a sample warms to nearer the transition temperature, the frequencies and damping rates both show very strong temperature dependence. These results are not at all understood theoretically and have stimulated considerable investigation.
  • Coherence Properties. A fascinating aspect of BEC is the close analogy which can be made between BEC and the much more familiar technology of lasers. A quantitative measurement has been made of the third-order coherence in BEC, which makes the analogy still more compelling.
  • Mixed-species condensates. It is possible to make two different varieties of condensate, two distinguishable spin-states, and confine them together in the same trap at the same time. Under certain conditions these two gases are immiscible, like oil and water, and under other conditions they freely intermingle. A type of quantum beat-note has been observed between the two condensate species, and the significance of this phenomenon for precision metrology and standards is being investigated.

  Two years after BEC was first realized, its ultimate technological significance remains a challenge. However, the level of scientific understanding is greatly improved, and the potential for still further advances is magnificent.
• Atom Guiding in Hollow Fibers. Previously, the atom guiding group demonstrated guiding room-temperature rubidium atoms through hollow-core optical fibers from 3 cm to 12 cm in length. Hollow-core optical fibers with inner radii of 10 µm to 100 µm connect two ultrahigh vacuum systems. Light is injected into the annular glass region surrounding the hollow core. When the light is detuned to the blue of an atomic resonance transition, the evanescent field, which extends into the hollow-core region from the annular core, exerts a repulsive force on the atoms. This repulsive force reflects the atoms from the glass surface, in analogy to total internal reflection for laser light in a step index glass fiber. These laser-light-doped hollow fibers may then be used for a variety of atom optics experiments, including atom interferometry, atom lithography, and atom transport.

  In the last year a new atom source was built to load atoms into the optical fiber. The Low Velocity Intense Source (LVIS), uses a standard magneto-optical trap (MOT) to collect approximately 10⁹ atoms at temperatures less than 1 mK. The MOT uses three orthogonal laser beams which are retro-reflected to provide cooling and trapping in all 6 directions. To create LVIS, a small hole is drilled in one of the retro-optics which is inside the vacuum chamber. This creates a shadow, and atoms that enter into this shadow feel no balancing force and exit the hole in the optic. An atomic beam produced by LVIS has been shown to supply 10⁹ ultracold atoms per second. One end of the fiber is placed into the LVIS beam to direct the ultracold atomic beam into the fiber. The other end of the fiber is sent into a detection chamber where the guided atoms are ionized on a moveable hot-wire detector and the ions are collected by a channeltron.

  Recently laser cooled atoms have been successfully guided for the first time. An example of the velocity profile of the guided atomic beam is given in Fig. 1. Atoms have been guided through fibers up to 25 cm in length, the longest achieved so far. This allows the study of some of the technical issues necessary to develop atom fiber technology: coherence, evacuation of the hollow-core region, laser-speckle interference in the inside interface of the fiber, and air-to-vacuum chamber coupling of solid and hollow-core optical fibers. Studying the latter issue has led to the development of novel Teflon-based vacuum couplers for optical fibers. The next series of studies will involve studying the coherence properties of the atom guide and building an atom interferometer inside the fiber.

  Figure 1. Spatial profile of a laser cooled atom beam guided through 25 cm of a hollow-core optical fiber.

• Nonlinear Light Interactions with Matter. Recently developed techniques to measure the full electric field (both amplitude and phase) of a single ultrashort laser pulse provide a new way to study, with femtosecond time resolution, various aspects of the interaction of light with materials. These techniques have recently been implemented to study the nonlinear propagation of femtosecond pulses in bulk media. The goal of this work is to make accurate measurements of the propagation of intense femtosecond pulses in a variety of materials, and to use these measurements to guide the development of theoretical models of propagation. The propagation of intense femtosecond pulses involves a wide range of complicated linear and nonlinear phenomena, and these experiments are important for a complete understanding of light-material interactions, as well as for applications in areas such as communications, atmospheric propagation, and new wavelength generation.

  Initial experiments have examined pulse propagation in the normal dispersion regime in fused silica, and future experiments will be performed in a variety of dielectric and semiconductor materials. Fused silica is one of the most widely used transmissive materials in femtosecond laser systems operating in the visible and infrared because it has low dispersion and is therefore presumed to contribute little to the temporal broadening of femtosecond pulses. However, at high intensities, the combined effects of the nonlinear contribution to the index of refraction and the linear dispersion can lead to surprisingly large changes in the pulse width over relatively small propagation lengths. For approximately plane-wave beam propagation through 2.54 cm of fused silica, the full electric field of a femtosecond pulse has been measured, demonstrating a large increase in temporal pulse broadening from 92 fs to 170 fs. For comparison, temporal broadening due to group velocity dispersion alone would be just 5 fs. This temporal broadening can be explained by the combined effects of normal dispersion and self-phase modulation due to a positive nonlinear index of refraction, which continuously separate the long and short wavelength components of the pulse.

  Temporal Pulse Splitting. The transverse spatial separation during the propagation of intense femtosecond pulses in glass results in interesting and unexpected phenomena. The inclusion of diffraction and self-focusing can result in a situation where the pulse splits temporally. The top part of Fig. 2 shows the measured electric field intensity and phase of a high power (peak power equals 5 MW) pulse after propagating through 2.54 cm of fused silica. The pulse undergoes temporal pulse splitting due to the combined effects of dispersion, material nonlinearity, and diffraction. In addition, the split pulses are nearly two times shorter than the original pulse width. At even higher powers, the pulse is observed to split into multiple sub-pulses. The bottom part of the figure shows the result of a numerical simulation that has been performed by solving a three-dimensional nonlinear Schrödinger equation model that accounts for instantaneous Kerr nonlinearity, dispersion and diffraction. The model predicts the pulse splitting observed experimentally, but cannot reproduce the asymmetries seen in both the intensity and phase of the experimental data, implying the need for additional terms in the model. These pulse-propagation experiments are one example of using diagnostics that measure the full electric field evolution of femtosecond pulses for areas such as materials characterization and ultrafast dynamic spectroscopy. In addition, basic properties of electronic and optoelectronic materials can be determined by measurement of the change of the phase of an optical pulse after the pulse traverses a given material. This technique will also be used to determine, with a high degree of accuracy, linear and nonlinear optical properties of materials such as dispersion, nonlinear refraction, two-photon absorption, as well as other third-order nonlinear coefficients and higher-order nonlinear properties. Group velocity dispersion and higher-order dispersion will also be measured for a variety of materials with white-light interferometry.

  Figure 2. (top) Measured electric field intensity (solid line) and phase (dots) of a high-power (peak power = 5 MW) pulse after propagating through 2.54 cm of fused silica. (bottom) Result of a numerical simulation that has been performed by solving a three-dimensional nonlinear Schrödinger equation model.

• Temporal Contrast Near-Field Optical Microscopy. As the spatial dimensions of semiconductor devices continue to shrink, new methods to probe line dimensions and the characteristics of individual features or defects become important. The technique of near-field optical microscopy provides the potential for measurements with 100 nm resolution or less, while at the same time utilizing light in the visible or near infrared for excitation or probing. Thus far, limited applications of time-resolved methods, such as pump-probe measurements, have been achieved in combination with near-field optical microscopy. A novel pump-probe method with ultrafast laser pulses has been implemented using a conventional fiber optic, near-field scanning optical microscope.

  The first experiments investigate the time response of semiconductor defects on GaAsP. A 100 fs optical pulse from a Ti:sapphire ultrafast laser at 780 nm is split into two, with the pump pulse being a factor of ten more energetic than the probe pulse. These pulses illuminate a GaAsP photodiode through the near-field optical tip. The protective coating has been removed from the diode so that impurity defects in the surface layer are exposed and the near-field microscope can interrogate the surface in close proximity. The weak probe pulse is modulated, and the photocurrent produced by the near-field optical microscope is detected only at this modulation frequency; thus the effect of the strong pump pulse on the weak probe pulse at the sample is obtained. The wavelength of light is selected so that, in the pure GaAsP material, a two-photon absorption is required to obtain a photocurrent. At the impurity defect, the Fermi level is shifted in energy, and a one-photon photoconductive signal is observed at 780 nm. If the strong pulse is followed immediately by the weak one, the absorption of the weak pulse is observed to be enhanced, thus producing a measurable time contrast mechanism as the photogenerated carriers relax. Figure 3 shows the effect of the strong pulse on the time-delayed weak pulse at several time delays, spatially resolved in the vicinity of a large defect. Over the defect-free regions of the semiconductor, the photocurrent is governed by the two-photon autocorrelation effect, which shows little change in amplitude. In the images the pulse durations are 300 fs. However, the signal at the defect shows a recovery time using reverse-chirped pulses of 100 fs or less. Work is ongoing to interpret the physical basis for the temporal contrast mechanism. The results of this work will permit new methods of temporal and spatial contrast imaging for the investigation of semiconductor devices and defects.

  Figure 3. Sequence of spatial and time-resolved photocurrent signals acquired at a GaAsP semiconductor defect using femtosecond pump/probe methods and near-field scanning optical microscopy. The first time delay is when the probe comes before the pump by 400 fs.

• Nonlinear Optical Interactions. High-power, near-resonant focused radiation propagating through a dense atomic vapor can form a constant-size, self-focused filament. Furthermore, the strong saturation that occurs within this filament generates new optical frequencies, and these exit the vapor in a range of directions. Although this phenomenon has been known for 20 years and is closely connected to fiber-communication losses, the first measurements of this phenomenon have been completed that characterize stable, self-focused filaments. The results are quite surprising; they contradict previous models based on four-wave-mixing and suggest a major importance of the transient aspects of the light pulse.
• Nanoscale Imaging with Combined Atomic Force Microscopy (AFM) and Near-field Scanning Optical Microscopy. There is a keen interest in fabrication and characterization of nanoscale objects, particularly below the diffraction limits associated with optical wavelengths. Novel methods for atomic force and electric field imaging of nanoscale size particles on surfaces have recently been developed, and a scanned-probe optical microscope with very exceptional spatial resolution has been devised. Specifically, a unique combination of non-contact AFM and evanescent laser excitation has been implemented to image nanoscale (10 nm to 30 nm diameter) Au spheres on quartz substrates, whereby the ultra sharp Si AFM tip acts as a spatially sensitive AC light field scattering probe for the electric fields surrounding the particles. These studies reveal strong enhancements of the evanescent electric fields due to optical excitation of plasmon resonances in the Au nanospheres, which is closely related to the enhancement of Raman signals observed for molecules near roughened Ag surfaces. This instrument is now being developed for studying individual molecules on the surface. The Si AFM tips are being coated with Ag in order to permit a spatially controllable, electric field enhancement of evanescent waves for single-molecule Raman spectroscopy on surfaces.
• Radical Product Detection in Etching. Etching is the process of material removal and is exceedingly important in patterning semiconductor devices. Detection of product species during the semiconductor etching process can be a valuable adjunct for determining the end point of the process. A laser ionization method has been developed to detect the silicon chloride radical products during etching. The method provides additional selectivity to determine the product species under various etching conditions.

  The ninth harmonic of a Nd:YAG laser at 118 nm is produced by a nonlinear optical process in a Xe/Ar gas mixture. These photons have 10.5 eV of energy, which is sufficient to ionize radical species during etching, such as SiCl and SiCl₂, but not sufficient to ionize and dissociate these species. Thus, the simplest radical species produced in the etching process can be detected directly and selectively after ionization by using a time-of-flight mass spectrometer. In a thermal etching experiment of silicon by molecular chlorine, both SiCl and SiCl₂ are detected as a function of the silicon wafer temperature. The SiCl is produced preferentially at the highest wafer temperatures and SiCl₂ at lower temperatures. This is the first direct detection of SiCl formation during the etching process. The kinetics of the etch-product formation are obtained, indicating that the formations of both species take place by complex kinetic mechanisms.

  In new studies, kinetic-energy-enhanced neutral and ion bombardment will be used together with the laser detection to measure the etch product species under conditions similar to plasma processing. This will provide a detailed understanding of the mechanisms of the etching process and establish parameters for which the etching process can be carried out with high efficiency and minimum damage by the kinetic-energy-enhanced species.

• Particulate Growth in Silicon Deposition Discharges. Silicon particulates form in the plasma of rf discharges used to deposit silicon thin-film display devices and photovoltaics. Under some conditions these particles subsequently deposit into the growing films, influencing performance. The particle growth and transport have been studied with sensitive light-scattering techniques, and these observations have identified the primary plasma forces and chemistry involved. The particles are negatively charged and trapped in the plasma most of the time, but when large particle densities occur, a fraction becomes neutralized and can escape to the growing film. Mitigation techniques are being developed, based on this developing understanding.
• Limits of Laser Stabilization. Potential highly accurate optical frequency standards are available on transitions at 532 nm in molecular iodine. Stable infrared lasers are being used based on Nd:YAG, efficiently pumped by a laser diode, and the infrared output is frequency-doubled in an external buildup cavity. The I₂ transition linewidth is 250 kHz and a theoretically possible stability is ~5 × 10^-14 at 1 s. At 100 s a 10-fold improvement to 5 × 10^-15 is expected, and this unprecedented level is in fact achieved (Fig. 4). At still longer times, changes in technically-generated offsets cause deviations from ideality. Nevertheless the results are the best achieved world-wide, by a large margin.



Figure 4. Representative frequency stability of Iodine-stabilized Nd:YAG laser. Calculated short-term performance is achieved, but variations of systematic offsets from the electro-optic modulator limit performance for times ~30 s and beyond.

Comparison of I2 and HCCD as Optical Frequency Standards. Recent work compares the frequency standard transitions in I2 with those at 1064 nm in HCCD molecules using the newly-developed method of Noise-Immune Cavity-Enhanced Optical Heterodyne Molecular Spectroscopy (NICE-OHMS), where unprecedented stability has also been achieved: 2 × 10-13 at 1 s, and below 5 × 10-15 at 1000 s (Fig. 5). This stabilization performance is perhaps 4- or 5-fold worse than achieved with the Iodine resonances, while the integrated absorption of the HCCD line is a million-fold weaker than that of the Iodine. This result shows two things. Stability improvements would be possible with iodine, using an appropriate variant of the NICE-OHMS method. Secondly, it shows that a range of potential secondary reference standards are available using other overtone transitions in HCCD, HCCH, NH3, HBr and the like. Interestingly, the HCCH lines at 795 nm are ~30-fold stronger than the HCCD lines used at 1064 nm. The lines at 1030 nm may be 100-fold stronger. At present the strategy is to use many different transitions in such molecules to cover by heterodyne means a grid from the RB standard at 778 nm to beyond 800 nm. These reference frequencies will be accurately determined using the Optical Comb Generator technology described later.

Figure 5. (top) Stability of beat between I2 stabilized and HCCD-stabilized lasers. The improved ultrasensitive detection of a weak overtone resonance of molecular HCCD permits progressively better results for laser stabilization. (bottom) The heterodyne reference laser is stabilized on an I2 transition at 532 nm using modulation transfer spectroscopy. This reference laser has a stability ~5 × 10–14 at 1 s, from beating experiments with two I2-stabilized systems.

• State-to-State-Reaction Dynamics via IR Laser Absorption Spectroscopy in Crossed Supersonic Jets. Most of chemistry in the gas phase occurs via isolated binary collisions of highly reactive species. The detailed dynamics of such reactive collisions is often hidden in the "nascent" quantum-state distributions (i.e., vibrational, rotational) of the product molecules, which have been quite challenging to obtain under conditions where subsequent collisional relaxation can be neglected. A new and very general probe of such reaction dynamics has been developed, based on direct IR laser absorption of the products of radical-neutral reactions in crossed supersonic jets. As one specific example, recent work has taken advantage of high fluorine radical densities available in novel discharge sources to investigate reactions of F + H₂ ↔ HF(v,J) + H under single collision conditions. The distributions of the HF(v,J) product are probed with full quantum-state selectivity by absorption of a high-resolution color center laser in the crossing region, which is sufficiently sensitive to see down to <10⁸ molecules/cm³. By slowing down the F and H₂ sources, the threshold energy dependence of this chemical reaction has been investigated. These measurements, in conjunction with quantum close-coupling calculations, provide stringent tests of available theoretical potential energy surfaces. Of particular interest, these high-resolution threshold studies have provided a clear indication of non-Born-Oppenheimer chemical reactions, i.e., that the spin-orbit excited state, F^*(²P_1/2), is in fact significantly reactive with H₂(v=0). Though this has long been discussed theoretically, these experiments provide the first empirical evidence for non-adiabatic behavior in such a fundamental and well studied chemical reaction system.
• Supersonic Slit Discharges: an Intense New Source of Jet-Cooled Radicals, Molecular Ions and Radical Clusters. The vast majority of chemical reactions taking place in the atmosphere, combustion, flames, plasmas, chemical vapor deposition, and semiconductor etching occur via open-shell "radical" species and/or molecular ions. The reactivity of these open-shell radicals is so much greater than the corresponding closed-shell species that they dominate the reaction kinetics, even though the radicals are typically present in extremely low concentrations. It is this high reactivity that makes them challenging species to generate in sufficient density to characterize spectroscopically under controlled gas-phase laboratory conditions.

  A new method has been developed for generating intense sources of radicals and molecular ions, based on striking a pulsed discharge in the stagnation region behind a slit supersonic jet. This has several important advantages over more common discharge sources. First, the species are formed and then supersonically cooled down to the lowest few quantum states. Secondly, the molecules have their velocities collimated perpendicular to the slit direction, which is both ideal for long-path, direct-absorption spectroscopy with tunable lasers, and generates "sub-Doppler" absorption profiles that are as much as 10-fold narrower than under non-supersonic discharge conditions. Thirdly, the closed-shell precursor species are moving supersonically in the discharge for only a few microseconds, which permits reactive species to be formed faster than they can be removed via secondary chemical reactions. Thus far, these laser methods have permitted spectroscopic study of hydrocarbon radicals, CH₃ (methyl), CH₃CH₂ (ethyl), and CH₂-CH-CH₂ (allyl), cooled to T_rot = 10 K to 20 K. The low jet temperatures offer two important advantages; i) spectral congestion at these temperatures is virtually eliminated and ii) detection sensitivity is enhanced by collapsing radical population into a much smaller number of initial quantum states. These methods greatly extend the current ability to characterize highly reactive radicals with high-resolution lasers, and should permit detection of even much larger species such as polycyclic aromatic hydrocarbons.

• Femtosecond Laser Control of Wave Packet Dynamics. Preparation of a superposition of excited states in an atom or a molecule with a broadband femtosecond laser pulse produces a "wave packet" that evolves in time. Such coherent preparation of states provides a powerful means of manipulating information, which can be encoded into the time evolution of a system. Both the amplitudes and phases of the states involved in the wave packet can be manipulated to contain information or to create a molecular system with novel properties. Such work is related to the quantum encoding of information.

  The experiments are carried out on the molecular system of lithium dimers. An ultrafast laser is used to prepare the wave packets from a single launch state, which is formed by a narrow-band cw laser. The results and corresponding theoretical interpretations demonstrate both precision assembly of wave packets and new aspects of amplitude and phase control of wave-packet motion.

  Modifications of individual wave function amplitudes in the wave packet are achieved with ultrafast laser pulse shaping techniques that vary the amplitude of the laser electric field at specific frequencies. The spectral phases of the femtosecond pulses are also controlled selectively to obtain precise phase manipulation of the wave-packet states. The results demonstrate that different wave packets can be formed out of single set of states by altering the amplitudes and phases of individual states, resulting in dramatically different time dependencies.
• Higher Vector Correlations in Atomic Energy Pooling. The manipulation of atomic systems has evolved from measuring individual cross sections, or rates, to aspects that stress the vector nature, or anisotropic properties, of individual systems. Thus far, two- and three-vector correlations have only recently been studied and worked out in detail. Three- and four-vector correlations involve important aspects of phase information, which are conveyed in the collision process. The potential exists to understand how to control collision probabilities and to utilize selective aspects of these effects to manipulate systems.

  In recent work, three- and four-vector correlations have been observed in collisions of two electronically excited calcium atoms. The product states are one at higher energy and one at lower energy. In the case of a four-vector correlation, two of the vectors are the initial alignment of each excited atom, the relative velocity vector, and the final alignment of the excited product state. These experiments are performed in a single beam apparatus with a new magnetic precession method to sample all the different alignments of the atomic states during a single laser excitation process. The final state alignment is probed by polarized fluorescence.

  The formalisms for three- and four-vector processes are derived to analyze collision systems for the individual conventional and coherence cross sections involved in the energy-transfer process. In the case of the coherence cross sections, both real and imaginary values play a role in the quantum interferences or phase information contained in the collision process. Such details play a remarkably significant role in the magnitude of the collision rates observed.
• Gravity Measurement. The nearly 1 % discrepancy between the recent PTB (Physikalisch Techniche Budesanstalt) results and the "accepted" value of G, the Newtonian constant of gravitation, is of considerable interest to the standards community. The fact that the PTB measurement appears to have been competently and thoughtfully carried out makes the discrepancy even more intriguing.

  NIST personnel along with collaborators from the National Geodetic Survey, Micro-g solutions, and JILA, have used an FG-5 absolute gravimeter together with a moveable 500 kg tungsten mass which surrounds the dropping chamber to measure G through the effect this large mass has on the measured value of g in the free-fall region. The measurement has been carried out and the data analysis is nearing completion. Though the notion of measuring the small mass-induced &Delta";g effect on top of g itself might appear almost impossibly difficult, the fact that the g-signal can be modulated (by moving the source mass every 20 min to either increase or decrease the measured value of g) makes it possible to measure well into the traditional little-g noise floor which is comprised of effects such as tide model uncertainties and atmospheric direct attractions as well as loading –effects whose time signatures range from ˝ day and longer. It appears that the "expected" accuracy (given the integration time and available FG-5 precision) of 5 or 6 parts in 10⁴ will be achieved. The analysis, while still in process, gives a preliminary number which lies within one sigma of the original CODATA value.

• Isolation System. Excellent progress has been made in developing an active low-frequency isolation system. Two stages are now operational and provide more isolation at 1 Hz than any other research or commercial isolation system. The "preliminary" stage which operates outside of vacuum has optical imaging readouts of the seismometers used for vibration sensing. It is fully operational. The two "main" stages are housed in a vacuum system which is carried on the preliminary stage and have seismometers with high sensitivity interferometric readouts, as well as lower sensitivity optical imaging readouts. The first main stage has been fully instrumented with seismometers and is fully operational using the low-sensitivity readouts. The first main stage is also partially operational using the high-sensitivity readouts. The second main stage has not yet been instrumented with seismometers, but it does carry a dummy payload. At this point, we have clearly demonstrated stable, low-frequency, 6° of freedom control of a platform with 40 dB of isolation. We have also demonstrated that such stages can be stacked and remain stable. With each additional stage of isolation, the requirements on seismometer sensitivity increases. Special seismometers are being developed to meet these progressively lower noise requirements. In addition to fundamental challenges, a large number of technical difficulties have been recognized and overcome.

Most Recent Technical Activities   |   Archive of Technical Activities