In October 2005, an intensive three-year Laser Triggered Gas Switch (LTGS) development program was initiated to investigate and solve observed performance and reliability issues with the LTGS for ZR. The approach taken has been one of mission-focused research: to revisit and reassess the design, to establish a fundamental understanding of LTGS operation and failure modes, and to test evolving operational hypotheses. This effort is aimed toward deploying an initial switch for ZR in 2007, on supporting rolling upgrades to ZR as the technology can be developed, and to prepare with scientific understanding for the even higher voltage switches anticipated needed for future high-yield accelerators. The ZR LTGS was identified as a potential area of concern quite early, but since initial assessments performed on a simplified Switch Test Bed (STB) at 5 MV showed 300-shot lifetimes on multiple switch builds, this component was judged acceptable. When the Z{sub 20} engineering module was brought online in October 2003 frequent flashovers of the plastic switch envelope were observed at the increased stresses required to compensate for the programmatically increased ZR load inductance. As of October 2006, there have been 1423 Z{sub 20} shots assessing a variety of LTGS designs. Numerous incremental and fundamental switch design modifications have been investigated. As we continue to investigate the LTGS, the basic science of plastic surface tracking, laser triggering, cascade breakdown, and optics degradation remain high-priority mission-focused research topics. Significant progress has been made and, while the switch does not yet achieve design requirements, we are on the path to develop successively better switches for rolling upgrade improvements to ZR. This report summarizes the work performed in FY 2006 by the large team. A high-level summary is followed by detailed individual topical reports.
Focused Beams from high-power lasers have been used to command trigger gas switches in pulse power accelerators for more than two decades. This Laboratory-Directed Research and Development project was aimed at determining whether high power lasers could also command trigger water switches on high-power accelerators. In initial work, we determined that focused light from three harmonics of a small pulsed Nd:YAG laser at 1064 nm, 532 nm, and 355 nm could be used to form breakdown arcs in water, with the lowest breakdown thresholds of 110 J/cm{sup 2} or 14 GW/cm{sup 2} at 532 nm in the green. In laboratory-scale laser triggering experiments with a 170-kV pulse-charged water switch with a 3-mm anode-cathode gap, we demonstrated that {approx}90 mJ of green laser energy could trigger the gap with a 1-{sigma} jitter of less than 2ns, a factor of 10 improvement over the jitter of the switch in its self breaking mode. In the laboratory-scale experiments we developed optical techniques utilizing polarization rotation of a probe laser beam to measure current in switch channels and electric field enhancements near streamer heads. In the final year of the project, we constructed a pulse-power facility to allow us to test laser triggering of water switches from 0.6- MV to 2.0 MV. Triggering experiments on this facility using an axicon lens for focusing the laser and a switch with a 740 kV self-break voltage produced consistent laser triggering with a {+-} 16-ns 1-{sigma} jitter, a significant improvement over the {+-} 24-ns jitter in the self-breaking mode.
We report on the successful attempts to trigger high voltage pressurized gas switches by utilizing beam transport through 1 MO-cm deionized water. The wavelength of the laser radiation was 532 nm. We have investigated Nd: YAG laser triggering of a 6 MV, SF6 insulated gas switch for a range of laser and switch parameters. Laser wavelength of 532 nm with nominal pulse lengths of 10 ns full width half maximum (FWHM) were used to trigger the switch. The laser beam was transported through 67 cm-long cell of 1 MO-cm deionized water constructed with anti reflection UV grade fused silica windows. The laser beam was then focused to form a breakdown arc in the gas between switch electrodes. Less than 10 ns jitter in the operation of the switch was obtained for laser pulse energies of between 80-110 mJ. Breakdown arcs more than 35 mm-long were produced by using a 70 cm focusing optic.
Electrical breakdown simulations are carried out for liquids in response to a sub-microsecond ({approx}100-200 ns) voltage pulse. This model builds on our previous analysis and focuses particularly on the polarity effect seen experimentally in point-plane geometries. The flux-corrected transport approach is used for the numerical implementation. Our model adequately explains experimental observations of pre-breakdown current fluctuations, streamer propagation and branching as well as disparities in hold-off voltage and breakdown initiation times between the anode and cathode polarities. It is demonstrated that polarity effects basically arise from the large mobility difference between electrons and ions. The higher electron mobility leads to greater charge smearing and diffusion that impacts the local electric field distributions. Non-linear couplings between the number density, electric field and charge generation rates then collectively affect the formation of ionized channels and their temporal dynamics.
This report documents measurements in inductively driven plasmas containing SF{sub 6}/Argon gas mixtures. The data in this report is presented in a series of appendices with a minimum of interpretation. During the course of this work we investigated: the electron and negative ion density using microwave interferometry and laser photodetachment; the optical emission; plasma species using mass spectrometry, and the ion energy distributions at the surface of the rf biased electrode in several configurations. The goal of this work was to assemble a consistent set of data to understand the important chemical mechanisms in SF{sub 6} based processing of materials and to validate models of the gas and surface processes.
In this paper, the authors report the absolute intensities of ultraviolet light between 4.9 eV and 24 eV ( 250 nm to 50 mn ) striking a silicon wafer in a number of oxide-etch processing discharges. The emphasis is on photons with energies greater than 8.8 eV, which have enough energy to damage SiO{sub 2}. These discharges were in an inductively-driven Gaseous Electronics Conference reference cell which had been modified to more closely resemble commercial etching tools. Comparisons of measurements made through a side port in the cell and through a hole in the wafer indicate that the VUV light in these discharges is strongly trapped. For the pure halocarbon gases examined in these experiments (C{sub 2}F{sub 6}, CHF{sub 3}, C{sub 4}F{sub 8}), the fluxes of VUV photons to the wafer varied from 1 x 10{sup 15} to 3 x 10{sup 15} photons/cm{sup 2} sec or equivalently from 1.5 to 5 mW/cm{sup 2}. These measurements imply that 0.1% to 0.3% of the rf source power to these discharges ends up hitting the wafer as VUV photons for the typical 20 mT, 200 W rf discharges. For typical ashing discharges containing pure oxygen, the VUV intensities are slightly higher--about 8 mW/cm{sup 2} . As argon or hydrogen diluents are added to the fluorocarbon gases, the VUV intensities increase dramatically, with a 10/10/10 mixture of Ar/C{sub 2}F{sub 6}/H{sub 2} yielding VUV fluxes on the wafer 26 mW/cm{sup 2} and pure argon discharges yielding 52 mW/cm{sup 2} . Adding an rf bias to the wafer had only a small effect on the VUV observed through a side-port of the GEC cell.
In-situ optical diagnostics and ion beam diagnostics for plasma-etch and reactive-ion-beam etch (RIBE) tools have been developed and implemented on etch tools in the Compound Semiconductor Research Laboratory (CSRL). The optical diagnostics provide real-time end-point detection during plasma etching of complex thin-film layered structures that require precision etching to stop on a particular layer in the structure. The Monoetch real-time display and analysis program developed with this LDRD displays raw and filtered reflectance signals that enable an etch system operator to stop an etch at the desired depth within the desired layer. The ion beam diagnostics developed with this LDRD will permit routine analysis of critical ion-beam profile characteristics that determine etch uniformity and reproducibility on the RIBE tool.
Rectangular trench profiles are modeled with analytic etch rates determined from measured ion distribution functions. The pattern transfer step for this plasma etch is for trilayer lithography. Argon and chlorine angular ion energy distribution functions measured by a spherical collector ring analyzer are fit to a sum of drifting Maxwellian velocity distribution functions with anisotropic temperatures. The fit of the model ion distribution functions by a simulated annealing optimization procedure converges adequately for only two drifting Maxwellians. The etch rates are proportional to analytic expressions for the ion energy flux. Numerical computation of the etch profiles by integration of the characteristic equations for profile points and connection of the profiles points is efficient.
In this paper we report absolute intensities of vacuum ultraviolet and near ultraviolet emission lines (4.8 eV to 18 eV ) for aluminum etching discharges in an inductively coupled plasma reactor. We report line intensities as a function of wafer type, pressure, gas mixture and rf excitation level. IrI a standard aluminum etching mixture containing C12 and BC13 almost all the light emitted at energies exceeding 8.8 eV was due to neutral atomic chlorine. Optical trapping of the WV radiation in the discharge complicates calculations of VUV fluxes to the wafer. However, we see total photon fluxes to the wailer at energies above 8.8 eV on the order of 4 x 1014 photons/cm2sec with anon- reactive wafer and 0.7 x 10 `4 photons/cm2sec with a reactive wtier. The maj ority of the radiation observed was between 8.9 and 9.3 eV. At these energies, the photons have enough energy to create electron-hole pairs in Si02, but may penetrate up to a micron into the Si02 before being absorbed. Relevance of these measurements to vacuum-W photon-induced darnage of Si02 during etching is discussed.
Plasma discharges involving mixtures of chlorine and boron trichloride are widely used to etch metals in the production of very-large-scale-integrated circuits. Energetic ions play a critical role in this process, influencing the etch rates, etch profiles, and selectivity to different materials. The authors are using a gridded energy analyzer to measure positive ion energy distributions and fluxes at the grounded electrode of high-density inductively-coupled rf discharges. In this paper, they present details of ion energies and fluxes in discharges containing mixtures of chlorine and boron trichloride.
We report measurements of the energies and angular distributions of positive ions in an inductively coupled argon plasma in a GEC reference cell. Use of two separate ion detectors allowed measurement of ion energies and fluxes as a function of position as well as ion angular distributions on the discharge centerline. The inductive drive on our system produced high plasma densities (up to 10{sup 12}/cm{sup 3} electron densities) and relatively stable plasma potentials. As a result, ion energy distributions typically consisted of a single feature well separated from zero energy. Mean ion energy was independent of rf power and varied inversely with pressure, decreasing from 29 eV to 12 eV as pressure increased form 2.4 m Torr to 50 mTorr. Half-widths of the ion angular distributions in these experiments varied from 5 degrees to 12.5 degrees, or equivalently, transverse temperatures varied form 0.2 to 0.5 eV with the distributions broadening as either pressure or RF power were increased.
We report the results of to separate long-term tests of batteries and charge controllers in small stand-alone PV systems. In these experiments, seven complete systems were tested for two years at each of two locations: Sandia National Laboratories in Albuquerque and the Florida Solar Energy Center in Cape Canaveral, Florida. Each system contained a PV array, flooded-lead-acid battery, a charge controller and a resistive load. Performance of the systems was strongly influenced by the difference in solar irradiance at the two sites, with some batteries at Sandia exceeding manufacturer`s predictions for cycle life. System performance was strongly correlated with regulation reconnect voltage (R{sup 2} correlation coefficient = 0.95) but only weakly correlated with regulation voltage. We will also discuss details of system performance, battery lifetime and battery water consumption.
Sandia has developed a third-generation set of specifications for performance and reliability testing of photovoltaic concentrator modules. Several new requirements have been defined. The primary purpose of the tests is to screen new concentrator designs and new production runs for susceptibility to known failure mechanisms. Ultraviolet radiation testing of materials precedes receiver section and module performance and environmental tests. The specifications include the purpose, procedure, and requirements for each test. Recommendations for future improvements are presented.