TECHNICAL ACTIVITIES 1998 - NISTIR 6268

TECHNICAL HIGHLIGHTS

• Four Vector Correlations in Collisions. The formalism and experiments for a four vector correlation, collisional process have been achieved for the first time. In an atomic beam of calcium atoms, two atoms are excited with a laser, which aligns the directions of the atomic orbitals, defining two vectors. A collision in a beam defines a third vector from the relative velocity, and the fourth vector is determined by the final state alignment after a collisional energy transfer process. Up until now, there have been only brief hints of such four vector correlations in other studies, but never actual measurements of cross section values. In this study, the complete formalism to extract four-vector correlation cross sections has been developed and experiments performed to obtain a large number of the cross section values for the first time. The experiments are performed on the system of two excited Ca atoms colliding to form one higher excited state and one ground state atom. There are 27 formal cross sections for the four vector process, and relative values for 18 of these have been obtained in the measurements here. (S.R. Leone)

Surfactant-Controlled Nanodot Growth Nanoscale semiconductor devices present numerous possibilities for future studies of quantum confinement and single electron devices. The growth of germanium on silicon is a classic system in which nanodot growth occurs spontaneously because of the mismatch in lattice spacings. Typically a few layers of germanium will grow first, layer by layer, and then islands of germanium form to produce nanodots of varying sizes. In recent experiments, control of the size of the nanodots has been demonstrated by use of an arsenic "surfactant" during molecular beam epitaxy deposition of germanium on Si (100). Experiments are performed with controlled molecular beam epitaxy of germanium on silicon, with varying amounts of arsenic surfactant deposited first on the surface. Atomic force microscopy is used to investigate the island size after growth. Laser ionization mass spectrometry and reflection high-energy electron diffraction are used to correlate the desorbing fluxes of arsenic dimers and tetramers with the onset of island growth. The results show a dramatic alteration of the island size and density with application of modest amounts of arsenic surfactant in the range of 0.25 to 1.0 monolayers. (S.R. Leone)

Figure 1. Germanium islands grown on Si(100) using arsenic as a surfactant.

• Infrared Near Field Optical Microscopy. A new, infrared, near field, optical microscope has been constructed and used successfully to interrogate photoresist polymer films with wavelength tunability. The results show great promise for probing with molecular group specificity in various materials. A new method has been devised to pull small diameter fiber tips in infrared-transparent fibers. This method uses a two-step laser heating and pulling technique, so that the fiber is first pulled bluntly to an intermediate diameter and then pulled to a very small diameter tip to form the probe. The fiber is a zirconium-aluminum fluoride material, which can be successfully coated with metal to form the transparent waveguide tip. A three micron, tunable, color center laser is transmitted through the fiber with an efficiency of one part in 10000, and the improvement in spatial resolution is over a factor of six better than the diffraction limit of the light at these infrared wavelengths. (S.R. Leone)
  

  Figure 2. Fiber tips for infrared near-field microscopy. A new, two-step method, for pulling fluoride fibers leads to 200 nm tips (left) with enhanced infrared transmissivity (right).

• Optical Frequency Comb Generator -- Optical Frequency Shifter. There have been considerable efforts toward converting the Optical Frequency Comb Generator into a versatile and dependable laboratory tool. The first version had several weaknesses, mainly optical inefficiency and the sensitivity of cavity tuning and alignment to temperature. The PID temperature controllers on both the crystal modulator case and on the optical mounting plate substantially eliminated these thermal problems. The optical inefficiency arose because of the high frequency-conversion efficiency of the 10.6 Ghz phase modulator: for a coherent optical input frequency, something like 1/3 of the power is "lost" by scattering to another sideband frequency. The cure is to add another mirror in the input line, also of ~99% reflectivity, and form an auxiliary Fabry-Perot cavity with the nominal Input Mirror. With PZT control, this auxiliary cavity can be locked on a transmission fringe for the input laser frequency, leading to a resonant transmission above 50%. The resulting apparatus has the new and elegant property that an input cw laser field is "transmitted" with an efficiency of ~10%, but with a controllable frequency shift of up to about 2 THz, in ~10 GHz steps controllable with rf precision. The optical spectral purity of the output beam is good, containing only about -20 dB of the strongest non-selected (neighbor) component. (J.L. Hall).

  Figure 3. Apparent measured frequency of the a₁₀ component of R(56) at 532 nm. The data of 4/18/98 are shown with their own baseline. June measurements were made to see limits of offsets caused by pre-filter mistuning, maladjustment of tracking filter gains, etc. Corrections for rf standard frequency and 633 nm I₂-stabilized laser offsets are not yet applied.
  

• NICE OHMS Spectroscopy, and Lasers Stabilized to HCCD and Iodine. It has been interesting to compare lasers stabilized onto iodine resonances using modulation transfer method with lasers stabilized to HCCD resonances using the Noise-Immune, Cavity-Enhanced Optical Heterodyne Molecular Spectroscopy (NICE-OHMS) method. (See Fig. 4) The HCCD-stabilized system shows only 2 to 4 times more short-term noise than the I₂ system, a remarkable success considering the green I₂ transition strength is half a million times stronger than the P(5) line of the C2HD (₂ + 3 ₃) overtone band. The fact that amazing frequency stability is being achieved with such an extremely weak reference transition is a direct result of the spectrometer's ultra-high detection sensitivity. For concreteness we show the frequency stability achieved with the two approaches.

  Figure 4. (top) Stability of beat between I₂ stabilized and HCCD-stabilized lasers. The improving ultrasensitive detection of a weak overtone resonance of molecular HCCD permits progressively better results on the 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.

  The NICE-OHMS spectrometer naturally provides laser frequency discrimination information from both the cavity resonance and the molecular transition. Thus, it is an ideal system for simultaneously achieving good short- and long-term frequency stabilization. The laser frequency basically tracks the cavity resonance with a precision of a few mHz with a fast servo loop. The vibration noise and the long-term drift of the cavity can be eliminated by stabilizing to the intracavity molecular transition. (J.L. Hall)
  

Apertureless Near Field Scanning Optical Microscopy. With the ever-decreasing scale of electronic components and chip design, there is an ever-increasing need to develop efficient optical methods to measure properties of nanoscale objects with sizes well below the diffraction limit of light. There have been considerable breakthroughs in this area based on near field scanning optical microscopy (NSOM), which has conventionally been achieved by using metal coated, optical fiber tips to confine the light in rapidly tapered, optical fibers. However, this method is inherently limited by the skin depth of light in the metal cladding ( 12 nm for Al), which, even for optimum cases, yields only 20 to 30 nm resolution and, under typical operating conditions, more on the order of 100 nm resolution. An alternative, actively explored in the Nesbitt and Gallagher groups, is to develop new methods in apertureless near field scanning optical microscopy, which already have demonstrated resolution improvements down to the 2 to 3 nm length scale. This effort is based on a combination of: i) evanescent wave excitation of molecules/nanostructures on a prism surface; ii) sharp Si or Ag coated Si structures guided by atomic force microscopy (AFM) to condense the evanescent electric fields in the vicinity of the tip; and iii) resonant light scattering or fluorescence detection of the molecules/nanostructures subsequent to the excitation event. This method has been used to image Au nanospheres by resonant scattering of 543 nm light near the Au plasmon resonance, and to determine that the combination of AFM tip + particle leads to a scattering enhancement of over 4000-fold from that of the bare Au nanospheres. (D.J. Nesbitt and A. Gallagher).

Figure 5. Apertureless AFM/NSOM fluorescence image of a dye doped polystyrene nanosphere (~80 nm) with the corresponding atomic force image below.

• Fluorescence Based Near Field Imaging Method. We have been extending near field imaging methods into the fluorescence domain, where the Si and Ag coated AFM tips are used to influence the near field excitation of dye molecules doped in latex nanospheres. The resulting fluorescence is imaged with a high numerical aperture microscope and fiber coupled avalanche photodiode combination. What is observed is considerable spatial structure in the fluorescence NSOM images as a function of tip-molecule distance, indicating both enhancement as well as quenching effects due to the presence of the tip. Especially interesting in the NSOM images are the effects due to the AFM probe tip blocking the molecular fluorescence, which therefore represents the casting of a vastly sub-diffraction-limited "shadow" of a near field light source by a nanoscale object. These measurements represent a world record for the highest spatial resolution achieved to date with near field excitation and fluorescence detection. The results provide unprecedented, experimental benchmarks for comparison with near field theory, as well as a completely new method of near field optical imaging by spatially shadowing specific fluorescent sites on the nm length scale. (D.J. Nesbitt and A. Gallagher)

• State-to-State Reaction Dynamics in Crossed Molecular Beams. Chemistry is a discipline of enormous technological and economic importance and yet a detailed understanding of how the simplest of chemical reactions occur still represents a state-of-the-art area of experimental and theoretical chemical physics research. There has been major progress in this area by exploiting slit discharges as an intense source of jet cooled F atoms for state-to-state reactive scattering studies in crossed supersonic jets. As the initial target, the focus has been on the classic F + H₂   HF (v,J) + H reaction, where the nascent rovibrational product states of HF are detected via direct IR laser absorption methods. The powerful advantage of such a high resolution, laser based approach is that it offers 10^-4 cm^-1 spectral resolution on the product states, which is 5 to 6 orders of magnitude better than previous crossed beam time of flight studies. These studies have yielded for the first time, collision free, fully quantum state resolved, product state HF (v,J) distributions from the F + H₂ reaction. These can now be studied as a function of center of mass collision energy. The experimentally observed distributions are in good qualitative agreement with state-of-the art ab initio/dynamics calculations. The data also reveal significant discrepancies between theory and experiment, providing the first experimental evidence that multiple electronic potential surfaces (i.e., with both ground state, spin orbit excited F atoms) are involved in the reaction event. This evidence for the importance of non-adiabatic processes in this simplest of "benchmark" systems implies a much richer range of dynamics for chemical reaction systems than heretofore suspected. (D.J. Nesbitt).

• Second Harmonic Generation from Si(100)/SiO₂. The Si(100)/SiO₂ interface is critical to the semiconductor industry. The channel region in a MOSFET is Si(100), while the gate dielectric consists of SiO₂. The thickness of the gate dielectric must scale in proportion to the scaling of the channel length. The current generation of integrated circuits has oxide thickness of approximately 9 nm, but within 2 to 3 generations the thickness will shrink to 4 nm. The necessity for extremely thin oxides has made roughness at the Si/SiO₂ interface a topic of increasing importance in the industry.

  Optical second harmonic generation has been shown to be sensitive to roughness at this interface. In materials with bulk inversion symmetry, second harmonic generation is only dipole allowed at an interface or surface, making it a highly interface/surface selective technique. However, the underlying physics that gives rise to the roughness sensitivity is not understood. We are investigating why second harmonic generation is sensitive to roughness. Our experiments are focused on measuring the spectral dependence of the signal. There are resonant structures that can be attributed to various features in the band structure of silicon.

  The preliminary results indicate a previously unknown resonance in the second harmonic spectrum. It is manifest as a strong increase in the signal at the blue end of the spectral region that the incident laser can tune over. Most strikingly, this resonance shows different symmetry properties than adjacent spectral regimes. The symmetry properties suggest that the signal is arising from step edges at the interface. If confirmed, this would be a very interesting result because we will have identified interface features that influence the second harmonic spectrum. Efforts are underway to extend the tuning range of the incident pulses using an optical parametric oscillator. This will necessitate the use of a reference signal to allow comparison of results from the optical parametric oscillator and laser. (S.T. Cundiff)

• Isolation System. New, more sensitive seismometers with interferometric readouts have improved the performance of our low-frequency, active isolation system. The low-frequency, active isolation system comprises multiple, suspended stages with six seismometers on each to sense the diminished disturbances that penetrate the passive suspensions. Control systems cause actuators to counter the residual disturbances on each stage. The "preliminary" stage, which operates outside of vacuum, has optical imaging readouts of its seismometers. It is fully operational and meets its design performance goals of a factor 100 isolation down to 1 Hz. The two "main" stages are housed in a vacuum system that is supported by the preliminary stage and has seismometers with both low sensitivity, imaging readouts for quieting and high sensitivity, interferometric readouts for maximum disturbance reduction. The first main stage now has operating interferometers on all six seismometers. The second main stage has not been instrumented with seismometers, but has a complete suspension and dummy payload. We have demonstrated 70 dB of isolation in vertical and horizontal degrees of freedom with both the preliminary and the first main stage control loops active. This demonstrates that such stages can be stacked, that they are stable in a stacked configuration and that high gains are simultaneously achievable on both stages. The demands on seismometer sensitivity increase at each inner stage because the remaining disturbances are reduced by another outer stage. Research continues on seismometer improvements and technical challenges stemming from the intricate control systems. (J.E. Faller)

• New, Small, Absolute, and Highly Portable Gravity Instrument. Work is progressing extremely well in the development of a new, mechanical, cam-based, free-fall (i.e., dropping) system. This is to be integrated in a small, fast, portable, and simple-to-use absolute g instrument. A two-centimeter free-fall drop is created by using a rotating cam to drive the dropping chamber. This approach results in a high measurement rate of 3 drops per second. The feasibility of this approach has been demonstrated, and presently a refined and evacuable apparatus is under construction. At the same time, and in the interest of both simplicity and instrumental cost, an alternate, passive spring system is being worked on as a possible replacement for the traditionally used "super spring." During the last year a major conceptual breakthrough occurred in regard to this development. By employing a double cam with the second half of the cam driving an auxiliary mass, the center of mass of the instrument does not move. Thus its motion can be used to cancel out ground-sensed weight changes of the apparatus that could heretofore systematically bias the determination. This double cam system has been implemented and the concept successfully demonstrated. This development is being carried out as a part of an ongoing CRADA with Micro-g Solutions, the commercial supplier of our previously developed FG-5 gravimeter. (J.E. Faller).

• Gravity Measurement. The nearly 1% discrepancy between the recent PTB results and the "accepted" value of "big" 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 movable 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 now completed. Though the idea of measuring the small mass-induced g effect on top of g itself might appear almost impossibly difficult, the fact that the signal can be modulated by moving the source mass every 10 or 20 min to either increase or decrease the measured value of g makes it possible to measure well into the noise floor. Though the "expected" accuracy (given the integration time and the expected FG-5 precision) should be 5 or 6 parts in 10⁴, instrumental drifts of unknown origin have limited the accuracy obtained to about 0.14%. The final result is in reasonable agreement with the presently accepted CODATA value and does not support the PTB outlying value. However, the value is about 0.2% higher than the CODATA value which, incidentally, puts us a bit closer to the PTB result. (J.E. Faller)

• Non-linear Dynamics of Intense Femtosecond Pulse Propagation. A clear picture of the propagation of femtosecond laser pulses is fundamentally important to many scientific and technological applications. Although some applications rely primarily on the delta-function qualities of a femtosecond pulse (e.g., time-resolved gating, transmission of binary data), at a more fundamental level pulse propagation encompasses much more than the simple transport of energy. It is a basic fact of the Maxwell equations that the manner in which a field propagates is fundamentally tied (via the polarization) to the properties of the medium in which it travels. From the standpoint of femtosecond diagnostics, this implies that if the electric field can be accurately characterized after propagation through a medium of interest, then valuable information about the medium and the propagation process may be obtained. Work has focused on both measuring and modeling the non-linear dynamics of the propagation of intense femtosecond pulses in bulk media.

  Measurements of the fields of such pulses have been performed with frequency-resolved optical gating (FROG), providing a means to observe the evolution of both the temporal amplitude and phase of femtosecond pulses as they undergo rapid broadening and splitting while propagating in fused silica. These accurate measurements have initiated the development of a more complete, modified, non-linear Schrödinger equation (NLSE), which includes contributions of physical mechanisms such as Raman non-linearities, space-time coupling, nonlinear shock effects, and non-paraxiality. The improved model successfully predicts temporal asymmetries observed in the measurements. In addition, the technique of spectral interferometry has been applied to full beam measurements, permitting the measurement of the full (temporal plus spatial) electromagnetic field on a femtosecond time scale for the first time. (T.S. Clement).

• Bose-Einstein Condensation. When a gas is made sufficiently cold, it undergoes a phase transition into a Bose-condensed state: the sample becomes highly coherent, with most of the atoms participating in a macroscopic common wavefunction. The atoms in a condensate are in close analogy with photons in a laser beam. It was NIST/JILA scientists in 1995 who first observed the phenomenon of Bose-Einstein condensation in a gas. Since then the field has grown very rapidly, to the point where there are now hundreds of technical papers published every year on the topic.

  Work at JILA on BEC in the past year has concentrated in two main areas, tunable interactions and mixed condensates.

  Tunable interactions: Many of the properties of a Bose-Einstein condensate are determined by the interactions between atoms. To the extent one can deliberately modify the nature of the interactions, one has intimate control over the behavior of the condensate. JILA theorists predicted that in the presence of an ambient magnetic field of about 15.0 mT (150 gauss), two-body interactions between Rb-85 atoms should go through a resonance. Experiments at JILA recently confirmed that the rate of elastic collisions between two ultracold Rb-85 atoms in a 15.5 mT (155 gauss) field is at least a factor of 10,000 higher than it is in a 16.7 mT (167 gauss field). The 16.7 mT magnetic fields at the heart of an atomic collision are on the order of tens of Tesla (hundreds of thousands of gauss). That a mT change in the ambient field could have such a profound effect on a collision says a lot about the power of resonance effects in ultracold collisions. Efforts are now underway to create a condensate in Rb-85 in order to exploit this so-called "Feshbach resonance" in BEC studies.

  Mixed Condensates: A series of studies on the behavior of condensate mixtures has been performed, with the two components corresponding to two different hyperfine states of Rb-87. Interestingly, the same atomic physics theory that predicts the "pressure" in the condensate (the self-repulsion of the condensate atoms) also yields a prediction for the pressure-shift of the Rubidium clock transition. Thus, some of JILA's fluid-dynamical studies on mixed condensates have provided a sensitive confirmation of the accuracy of the atomic theory that predicts the ultimate limit to the accuracy of Rubidium-based atomic clocks. (E.A. Cornell).

• Cooling a Dilute Gas of Fermionic Atoms to Quantum Degeneracy Substantial progress has been made toward the goal of cooling a dilute gas of fermionic atoms to quantum degeneracy. Building on techniques used to produce BEC in a dilute atomic gas, we have implemented a double-MOT apparatus for cooling and trapping potassium. It has a fermionic isotope in addition to two that are bosons. Features of the apparatus include the use of diode lasers exclusively and the successful implementation of a novel, enriched source for the fermionic isotope ⁴⁰K. In the first stage of cooling, we have trapped ⁴⁰K atoms in a magneto-optical trap (MOT) with four orders of magnitude higher number of atoms than any previous work. In preparation for the second stage of cooling, we have transferred 70% of these atoms into a magnetic trap, and seen lifetimes that are sufficiently long for successful evaporative cooling. Since evaporative cooling depends heavily on elastic collision rates that are not well known for potassium, we are performing measurements of cold collision rates and observing the suppression of s-wave collisions in a spin-polarized sample due to Fermi-Dirac statistics. Once quantum degeneracy is reached, additional effects of the quantum statistics and the formation of a Fermi sea will be studied through measurements of thermodynamics, light scattering, and collisional properties of the dilute fermionic gas. (D.S. Jin).

• Strontium Atom Trapping. Strontium atom trapping offers some outstanding opportunities for metrology and studying the physics of cold collisions. Since Sr singlet states have no fine structure and the principle isotope has no hyperfine structure, Sr provides exceptionally clean cut tests of trap collision theories. The triplet states of Sr offer an opportunity for analysis of trap behavior, very low temperature cooling, metrology and possibly Bose condensation without evaporative cooling. We have developed a Sr trap, based on resonance-line trapping from a thermal vapor in a sapphire-window cell. Trap loss due to leakage of excited 5¹P atoms (Sr*) to the triplet manifold is prevented with lasers that recycle metastable atoms back to the ground state, resulting in high trapped-atom densities and storage times. The trap velocity distribution is nondestructively measured using intercombination line fluorescence, which allows an exacting study of the dependence of trap temperature and cloud-cooling rates on detuning and power of the trapping beams. From the transient response of resonance-line and intercombination-line fluorescence, we deduce the various radiative rates within the lowest nine states of Sr. We have also determined the Sr-Sr* collisional loss rate coefficient from trap-loading time dependence and power and density dependencies. This is dominated by excitation of the attractive, non-radiating molecular state, which gains a dipole moment at large atomic separation due to radiation retardation. This provides an exceptionally sensitive test of retardation effects in molecular spectra, as well as of trap loss theories due to the exactly known molecular potentials. Atoms trapped at mK temperatures in the resonance-line trap are also transiently loaded into an intercombination-line trap with ~50% efficiency using broad-band, red-detuned cooling. This yields high Sr densities at µK temperatures, as well as the opportunity to measure cooling rates and unique forms of Ramsey fringes. (A.C. Gallagher).

• Particle Growth in Thin-Film Deposition Discharges. Particles grow in the discharges used to deposit thin semiconducting films. Some of these particles escape to the films, causing a deterioration in their electronic properties. To understand the causes and behavior of these particles, their growth and transport to the substrate has been studied in the types of discharges used to make amorphous silicon, thin-film transistors and photovoltaics. Light scattering by the suspended particles, during and immediately after extinguishing the discharge, has been analyzed to yield particle growth rates and densities for the very small particles (<30 nm diameter) that escape to the films. This has yielded clear insights regarding the species trapped in the discharge and the reactions causing them to grow. These results have also demonstrated that particle trapping and growth is an inevitable consequence of the plasma chemistry used to produce the films. However, this does not mean that particles will inevitably escape to the growing films. The observations and related models have identified important spatial variations in plasma potential, as well as thermophoretic effects, that can be used to mitigate the escape of particles to the films. (A.C. Gallagher).

Mission  |  Organization  |  Current Directions  |  Technical Highlights  |  Future Directions

TECHNICAL ACTIVITIES 1998 - Contents

Pages designed and maintained by the Office of ECSED.
Inquiries or comments: Feedback.
Online: April 1999