Results from ARIES-IFE Study and Research Activities on IFE Chamber and Optics at UC San Diego

Transcription

1 Results from ARIES-IFE Study and Research Activities on IFE Chamber and Optics at UC San Diego Farrokh Najmabadi and Mark Tillack UC San Diego IAEA Coordinated Research Project on Elements of Power Plant Design for Inertial Fusion Energy November 4-7, 2003 IAEA, Vienna, Austria Electronic copy: UCSD IFE Web Site:

2 Outline Results from the ARIES-IFE study, Assessment of IFE Chamber Concepts. IFE Research Activities at UC San Diego: Final Optics Experiments Chamber dynamics and clearing simulations Chamber wall simulation Experiments

3 Selected Results from ARIES-IFE Study For ARIES-IFE Publications, see:

4 ARIES Integrated IFE Chamber Analysis and Assessment Research Is An Exploration Study Objectives: Analyze & assess integrated and and self-consistent IFE IFE chamber concepts Understand trade-offs and and identify design windows for for promising concepts. The The research is is not notaimed at at developing a point point design. Approach: Six Six classes of of target were were identified. Advanced target designs from from NRL NRL (laserdriven direct drive) and and LLNL (Heavy-ion-driven indirect-drive) are are used used as as references. To To make progress, we we divided the the activity based on on three three classes of of chambers: Dry Dry wall wall chambers; Solid wall wall chambers protected with with a sacrificial zone (e.g., (e.g., liquid films); Thick liquid walls. ARIES-IFE study was completed in September ARIES-IFE study was completed in September 2003.

5 Reference Direct and Indirect Target Designs NRL Advanced Direct-Drive Targets 1 µm CH +300 Å Au LLNL/LBNL HIF Target.195 cm.169 cm.150 cm CH Foam + DT DT Fuel DT Vapor 0.3 mg/cc CH foam ρ = 20 mg/cc 5 µ CH.162 cm CH Foam + DT.144 cm DT Fuel cm laser power DT Vapor 0.3 mg/cc CH foam ρ = 75 mg/cc NRL Direct Drive Target Gain Calculations (1-D) have been corroborated by LLNL and UW time (ns)

6 Details of Target Spectra Has A Strong Impact on the Thermal Response of the Wall 1x x x10 9 1x10 8 1x10 7 Energy Deposition (J/m 2 ) in C and W Slabs (NRL 154MJ Direct Drive Target) Debris ions, C Fast ions, W Photons, W Photons, C Debris ions,w Fast ions, C 1x10 6 1x10-8 1x10-7 1x10-6 1x10-5 1x10-4 1x10-3 1x10-2 Penetration depth (m) C density = 2000 kg/m 3 W density = 19,350 kg/m 3 Heat Heat fluxes are are much lower than than predicted in in previous studies: A much smaller portion of of target yield yield is is in in X-rays. Time of of flight of of ions ions spread the the temporal profile of of energy flux flux on on the the wall wall over over several µs. µs. A cover gas gas may may not not be be necessary for for protecting the the chamber wall wall Photon and and ion ion energy deposition falls falls by by orders of of magnitude within mm mm of of surface.

7 Selected Results from ARIES-IFE Study : Dry Wall Concepts

8 Thermal Response of a W Flat Wall NRL NRL direct-drive target target in in 6.5-m 6.5-m chamber with with no no gas gas protection: 3-mm thick W Chamber Wall Energy Front Evaporation heat flux B.C at incident wall Coolant at 500 C Wall Temperature ( o C) Convection B.C. at 600 coolant wall: h= 10 kw/m 2 -K x x x x x x x x10-6 Surface 1 micron 5 microns 10 microns 100 microns 8.0x x x10-5 Time (s) Temperature variation mainly in in thin thin ( mm) mm) region. Margin for for design optimization (a (a conservative limit limit for for tungsten is is to to avoid reaching the the melting point point at at 3,410 C). Similar margin for for C slab. slab.

9 All the Action Takes Place within mm of Surface -- Use an Armor Photon and and ion ion energy deposition falls falls by by orders of of magnitude within mm mm of of surface. Beyond the the first first mm mm of of the the surface. First First wall wall experiences a much more more uniform q q and and quasi steady-state temperature (heat (heat fluxes similar to to MFE). Use an an Armor Armor optimized to to handle particle and and heat heat flux. flux. First wall wall is is optimized for for efficient heat heat removal. Depth (mm): Typical T Swing ( C): ~1000 ~300 ~10 ~1 ~ 0.2 mm Armor Structural Material Most Most of of neutrons deposited in in the the back back where blanket and and coolant temperature will will be be at at quasi steady state state due due to to thermal capacity effect Blanket design can can be be adapted from MFE blankets Significant high-energy ion ion flux flux on on the the armor. 3-5 mm Coolant

10 IFE Armor Conditions are similar to those for MFE PFCs (ELM, VDE, Disruption) ITER Type -I ELM s ITER VDE s ITER Disruptions Typi cal IFE Ope ration (direct-drive NR L targe t) Energy <1 MJ/m 2 ~ 50 MJ/m 2 ~ 10 MJ/m 2 ~ 0.1 MJ /m 2 Loca tion Surface ne ar d iv. strike poin ts surface surface bulk (~µm s) Time µs ~ 0.3 s ~ 1 ms ~ 1-3 µs Max. melting/ melting/ melting/ Temper ature sub lim ation sub lim ation sub lim ation ~ C (for dry wall) points points points Frequency Few Hz ~ 1 per 100 ~ 1 per 10 ~ 10 Hz cyc les cyc les Base C ~ 100 C ~ 100 C ~ >500 C Temper ature There is a considerable synergy between MFE plasma facing components and IFE chamber armor.

11 Target injection Design Window Naturally Leads to Certain Research Directions Direct-drive targets (initial T=18K) are are heated during their their travel in in the the chamber by: by: Friction with with the the chamber gas gas (mainly through condensation heat heat flux) flux) requiring Lower gas gas pressure Slower injection velocity Radiation heat heat flux flux from from hot hot first first wall, wall, requiring Lower equilibrium temperature Faster injection velocity Addition of of a thin thin (~70µm) foam foam improves the the thermal response considerably. Maximum condensation heat flux (W/m 2 ) 1x10 6 1x10 5 1x10 4 1x10 3 1x10 2 q'' max for D -T to reach triple point for R chamb of 6 m and injection vel. of: 400 m/s 100 m/s Max. heat flux is at the leading target surface Xe temp.,inj.vel. 1000K, 400m/s 4000K,400m/s 1000K, 100 m/s 4000K,100m/s 1x Condensation r coefficient x Pressure at RT (mtorr) (s c xp)

12 Design Windows for Direct-Drive Dry-wall Chambers M ax.e quilibrium W all Tem p. to Avoid Vap orization (C ) Graphite Chamber Radius of 6.5m Laser propagation design window(?) Experiments on NIKE Xe Density (Torr) Thermal design window Detailed target emissions Transport in the chamber including time-of-flight spreading Transient thermal analysis of chamber wall No gas is necessary Target injection design window Heating of target by radiation and friction Constraints: Limited rise in temperature Acceptable stresses in DT ice

13 Selected Results from ARIES-IFE Study : Wetted Wall Concepts

14 Aerosol Generation and Transport is the Key Issue for Thin-Liquid Wall Concepts A renewable thin-liquid protection resolve several issues: It It can can handle a much higher heat heat fluxes compared to to solid solid surfaces; It It will will eliminate damage to to the the armor/first wall wall due due to to high-energy ions. ions. A renewable thin-liquid protection, however, introduces its its own own critical issues: Fluid-dynamics aspects (establishment and and maintenance of of the the film) film) Wetted wall: Low-speed normal injection through a porous surface Forced film: High-speed tangential injection along a solid solid surface Chamber clearing (recondensationof of evaporated liquid) Source term: both both vapor and and liquid (e.g., (e.g., explosive boiling) are are ejected Super-saturated state state of of the the chamber leads leads to to aerosol generation Target injection and and laser laser beam propagation lead lead to to sever constraints on on the the acceptable amount and and size size of of aerosol in in the the chamber.

15 Two Methods for Establishment of Thin- Liquid Walls Have Been Proposed Liquid Injection ~ 5 m Wetted Film X-rays and Ions Injection Point First Wall Detachment Distance x d Forced Film

16 We Have Developed Wetted-Walls Design Widows for Establishment &Stability of the Protective Film Developed general non-dimensional charts for for film film stability over over a wide wide variety of of candidate coolants and and operating conditions. Model predictions are are closely matched with with experimental data. data. Penetration Depth [mm] Simulation Experiment Time [sec] z o ε s

17 We Have Developed Design Widows for The Forced-Wall Concepts Developed non-dimensional design widows for for longitudinal spacing of of injection/coolant/removal slots slots to to maintain attached protective film; film; δ =1 mm 1.5 mm 2 mm θ = x d [cm] Plexiglas (α LS = 70 ) Flat Curved 1 mm nozzle 8 GPM 10.1 m/s 10 inclination Re = We

18 Most of Ablated Material Would Be in The Form of Aerosol FLiBe aerosol and and vapor mass mass history in in a 6.5-m radius following a target explosion (ablated thickness of of mm) mm) Most Most of of ablated material remains in in the the chamber in in aerosol form; Only Only homogeneous nucleation and and growth from from the the vapor phase.

19 There Are Many Mechanism of Aerosol Generation in an IFE Chamber Homogeneous nucleation and and growth from from the the vapor phase Supersaturated vapor Ion Ion seeded vapor Phase decomposition from from the the liquid phase Thermally driven phase explosion Pressure driven fracture Hydrodynamic droplet formation (May be be critical in in Thick-liquid Wall Wall concepts )

20 Selected Results from ARIES-IFE Study : Thick Liquid Wall Concepts

21 Aerosol Generation and Transport is also the Key Issue for Thick-Liquid Wall Concepts Studies of of structural materials choices and and limits If If a series SS SS is is required as as a near-term base base line line for for the the design, then then Ti-modified 316SS (PCA) should be be used. used. Chamber vessel would not not be be a life-time components. However, it it was was strongly recommended to to consider alternate structural material candidates (ferritic steels and and SiC/SiC composites) offering the the possibility of of higher operating temperature & performance. In In this this case, case, chamber vessel may may be be a life-time component. Aerosol concerns (similar to to thin thin liquids) were were highlighted. Hydrodynamic droplet formation is is a key key issue. Flow Flow conditioningand and careful nozzle design are are needed to to control the the hydrodynamic source.

22 Studies of Ion Transport Modes Indicate Several Options are Feasible Chamber Concept Transport Mode Dry-wall ~6 meters to wall Wetted-wall ~ 4-5 meters to wall Ballistic Transport chamber holes ~ 5 cm radius most studied Vacuum-ballistic vacuum Not considered now: requires ~500 or more beams HIBALL (1981) Not considered: needs 0.1 mtorr Thick-liquid wall ~ 3 meters to wall Not considered: needs 0.1 mtorr Neutralized-ballistic plasma generators ARIES-IFE (2002) Possible option: but tighter constraints on vacuum and beam emittance OSIRIS-HIB (1992) ARIES-IFE (2002) Possible option: but tighter constraints on vacuum and beam emittance HYLIFE II (1992-now) ARIES-IFE (2002) Main-line approach: uses pre-formed plasma and 1 mtorr for 3 m ~ beams Pinch Transport chamber holes ~ 0.5 cm radius higher risk, higher payoff Preformed channel ( assisted pinch ) laser + z-discharge ARIES-IFE (2001) OPTION: uses 1-10 Torr 2 beams ARIES-IFE (2001) OPTION: uses 1-10 Torr 2 beams ARIES-IFE (2002) OPTION: uses 1-10 Torr 2 beams Self-pinched only gas ARIES-IFE (2001) OPTION: uses mtorr ~2-100 beams PROMETHEUS-H (1992) ARIES-IFE (2001) OPTION: uses mtorr ~2-100 beams ARIES-IFE (2002) OPTION: uses mtorr ~2-100 beams Ion-transport modes compatible with with dry- dry-and thin-liquid walls (larger chambers) are are feasible. May May require significantly smaller number of of beams simplifying the the system.

23 IFE Research at UC San Diego: Final Optics Research & Development Latest results were reported at IFSA 2003: Paper: Presentation:

24 Grazing incidence metal mirror: Design concept and key issues Key Key Issues: Shallow angle stability Damage resistance/lifetime Goal Goal = 5 J/cm J/cm 2 2,, shots shots Fabrication & optical quality Contamination resistance Radiation resistance The The reference mirror mirror concept concept consists consists of of stiff, stiff, light-weight, radiation-resistant substrates with with a a thin thin metallic metallic coating coating optimized for for high high reflectivity (Al (Al for for UV, UV, S-pol, S-pol, shallow shallow θ) θ) ~50 cm 85Þ

25 We Have Tested Several Al Fabrication Options Thin Thin films films on on superpolished substrates CVD SiC, SiC, 2-3Å 2-3Å roughness, nm nm flatness over over 3 cm cm magnetron sputtering up up to to nm nm e-beam evaporation up up to to 2 mm mm Solid polycrystalline metal Polished diamond-turned Electroplated and and turned Al Al Testing was was performed with with 25-ns, 248-nm pulses in in a controlled environment

26 Review of Progress: Accomplishments and Findings 1. No signs of a shallow angle instability under any conditions tested to date 2. Testing of diamond-turned surfaces showed high damage threshold at 532 nm, but much lower threshold at 248 nm 3. UV damage observed in air, requiring testing in vacuum

27 Review of Progress: Accomplishments and Findings 4. All evidence suggests that weakly attached contaminants are tolerable due to strong cleaning by laser 5. Thin film coatings have been difficult to produce with sufficient quality, and have a lower damage resistance than Al plates 6. Perturbations to transmitted light are measured in-situ using bright-field, dark-field and direct imaging dump profiler test specimen translation main beam probe laser

28 Review of Progress: Accomplishments and Findings 8. Our best results have been obtained with thick ( mm) electroplated coatings that are diamond-turned 100,000 shots at 3-4 J/cm2 with no discernable change to the surface 9. Scale-up and system integration (with driver and target injector) work has begun amp 1 amp 2

29 IFE Research at UC San Diego: Chamber Dynamics and Clearing Simulation

30 Response of Chamber to Target Explosion Covers Two Vastly Different Time Scales Pre-shot Chamber Environment Target injection, Laser propagation, Slow time-scale: processes Non-Equilibrium Environment SPARTAN Code Fast time-scale processes First pass of X-ray and ions through the chamber (a few µs) 1-D Codes such as Bucky

31 We Have Developed SPARTAN Chamber Dynamics and Clearing Code 2-D 2-D Transient Compressible Navier-Stokes Equations. Extension to to 3-D 3-D and and cylindrical geometry is is straight-forward. Second order Godunov method, for for capturing strong shocks. Diffusive terms (conductivity, viscosity) can can depend on on local local state state variables. Adaptive Mesh Refinement employed to to secure the the uniform accuracy throughout the the fluid fluid domain. Arbitrary boundary resolved on on a Cartesian grid grid with with Embedded Boundary method.

32 SPARTAN Sample Simulations Initial Conditions: Xenon gas gas at at mtorr, room temperature; 160MJ NRL NRL Direct drive drive target; Chamber condition from from BYCKY at at µs. µs. One One fluid fluid (Xe); (Xe); Ideal Ideal gas gas Law; Law; Diffusive terms terms depend depend on on local local state state variables through through Sutherland Law; Law; Boundary Condition: Zero Zero mass mass flux; flux; Reflective velocity; Energy: Energy: Zero Zero energy energy flux flux ( insulated wall ) wall ) Conduction to to the the wall wall (wall (wall temperature is is fixed fixed at at o C). o C).

33 Reference Case: No Viscosity, No Conductivity Pressure: t = 0.5 ms t = 5 ms t = 50 ms Temperature: t = 0.5 ms t = 5 ms t = 50 ms Convergence of of shock waves sets sets up up hot hot spots spots in in the the chamber. Flow Flow mixing is is too too slow slow to to equalize the the temperature.

34 Viscosity and Conductivity Significantly Change the Chamber Condition t = 5 ms t = 50 ms t = 100 ms No Diffusion T avg = 5.14 X 10 4o C T avg = 6.36 X 10 4o C Viscosity Only T avg = 3.79 X 10 4o C Conductivity Only

35 Chamber Conditions at 100 ms No Diffusion Viscosity only T avg = 5.14 X 10 4o C T avg = 3.79 X 10 4o C Conduction Only Full Navier Stocks T min = K T max = K T avg = 6.36 X 10 4o C T avg = 3.88 X 10 4o C

36 Evolution of Chamber Condition (Full Navier Stocks) Pressure Temperature Density

37 Chamber Evolution Occurs In Two Phases Evolution of Temperature of the Center of the Chamber Flow Flow is is dominated dominated by by shock shock bouncing bouncing Termpature (10 3 K) Large Large scale scale eddies eddies is is setup setup and and gas gas parameters parameters evolve evolve smoothly. smoothly Time (ms)

38 Turbulence Induced By Beam Channels Reduces The Chamber Temperature Significantly Ref. Case T min = 1.1 X 10 4 K T max = 14.3 X 10 4 K T min = 0.5 X 10 4 K T max = 11.3 X 10 4 K T min = 0.5 X 10 4 K T max = 8.57 X 10 4 K t = 5 ms t = 50 ms t = 100 ms T min = 1.03 X 10 4 T max = 13.5 X 10 4 T min = 0.51 X 10 4 T max = 9.01 X 10 4 T min = 0.44 X 10 4 T max = 6.28 X 10 4 K t = 5 ms t = 50 ms t = 100 ms T min = 1.0 X 10 4 K T max = 15.5 X 10 4 K T min = 0.36 X 10 4 K T max = 8.15 X 10 4 K T min = 0.32 X 10 4 K T max = 6.48 X 10 4 K t = 5 ms t = 50 ms t = 72 ms

39 Observation from Recent SPARTAN Simulations Chamber evolution occurs in in two two phases: 1) 1) Flow Flow is is dominated by by shock bouncing (first (first ~30-50 ms); ms); 2) 2) Large scale scale eddies is is setup and and gas gas parameters evolve smoothly. Diffusive processes lead lead to to a more more uniform chamber environment. Peak Peak temperatures, pressure, and and velocity are are reduced by by a factor of of two two or or more more compared to to the the case case with with no no conductivity/viscosity. Laser beam channels seed seed large large scale scale eddies in in the the chamber. The Theresulting flow flow mixing leads leads to to further reduction of of peak peak temperatures, pressure, and and velocity. It It also also enhances heat heat transfer to to the the wall. wall. Peak Peak temperature in in the the chamber is is still still too too high. high. But: But: Geometrical effects will will further reduce the the peak peak temperatures: more more beam channel, a cylindrical chamber to to shorten chock bouncing period. Effects of of background plasma (conductivity, viscosity, and and radiation)

40 SPARTAN Research Plan for Incorporate cylindrical symmetry. Completed. We We are are performing convergence studies at at present. Implement and and test test contributions of of viscosity, thermal conductivity, and and radiation from from background plasma. Models are are incorporated and and tested in in SPARTAN. Implement multi-species capability. Implement of of Equation of of State. Parametric investigation of of chamber dynamics with with different gas, gas, gas gas pressure, target yield, chamber size, size, beam ports, etc. etc. Parallelization and and extension to to 3-D 3-D geometry.

41 IFE Research at UC San Diego: Chamber Wall Response Simulation Experiments

42 Thermo-mechanical Response of the Wall Is Mainly Dictated by Wall Temperature Evolution Most Most phenomena encountered depend depend on on wall wall temperature evolution (temporal and and spatial) spatial) and and chamber environment except except sputtering and and radiation (ion (ion& neutron) damage damage Lasers Lasers can can be be used used to to simulate IFE IFE Chamber Wall Wall Response NRL Target, X-ray Only 1 J/cm 2, 10 ns Rectangular pulse Laser 0.24 J/cm 2,10 ns Gaussian pulse Wall surface 10µm depth Time (µs) Time (µs) Only Only laser laser intensity is is adjusted to to give give similar similar peak peak temperatures. Spatial Spatial temperature profile profile can can be be adjusted by by changing laser laser pulse pulse shape. shape.

43 Thermo-Mechanical Response of Chamber Wall Can Be Explored in Laser Simulation Facilities Requirements: Laser Laser pulse pulse simulates temperature evolution Capability to to simulate a variety of of wall wall temperature profiles A suite suite of of diagnostics: Real-time temperature (High-speed Optical OpticalThermometer) Per-shot Per-shot ejecta ejecta mass mass and and constituents (QMS (QMS& RGA) RGA) Rep-rated experiments to to simulate fatigue fatigue and and material material response Relevant equilibrium temperature (High-temperature sample sample holder) holder) Vacuum Chamber provides a a controlled environment Capability to to isolate isolate ejecta ejecta and and simulate a a variety variety of of chamber environments & constituents

44 High Temperature Sample Holder is Designed and is in Fabrication Specimen Sample holder is made of Mo Ceramic Insulator Copper conductor with set screw S.S. vacuum seal Air cooling inlet Thermocouple feed through Power feed through

45 Real-time Temperature Measurements Can Be Made With Fast Optical Thermometry Spectral radiance is is given by by Planck s Law Law (Wien s approximation): L(λ,T) = C /λt) 1 ε(λ,t) λ -5 exp(-c 2 /λt) Since emittance is is a strong function of of λ, λ, T, T, surface roughness, etc., etc., deduction of of temperature from from total total radiated power has has large large errors. Temperature deduction by by measuring radiance at at fixed λ One-color: Use Use tables/estimates for for ε(λ ε(λ 1,Τ) 1,Τ) Two Two colors: Assume ε(λ ε(λ 1,T) ε(λ 2,T) 1,T) = ε(λ 2,T) Three colors: Assume d 2 2 ε/dλ ε/dλ 2 2 = 0 [usually a linear interpolation of of Ln(ε) is is used] Our Our observations Two-color method achieves sufficient accuracy (~1%). Three-color method is is too too difficult.

46 Schematic of Multi-Color Fiber Optical Thermometer System is configured as three independent twocolor thermometer

47 Calibration of Thermometer Temperature is is calculated from from measurement of of radiated energy at at two two wavelengths: Thermometer is calibrated with a tungsten lamp (calibrated for 7 temperatures in the range 1,500-3,500 K) T = 1 1 c 2 λ1 λ2 5 λ 2 Lλ 2 ln λ1 Lλ1 = 1 1 c 2 λ1 λ2 5 λ 2 C2 V2 ln λ1 C1 V1. Calibration is is difficult because the the lamp lamp filament is is discontinuous image point point should be be exactly on on the the lamp lamp filament. We We developed the the protocol to to reliably calibrate the the thermometer (in (in <5 <5 minutes). Calibration is is so so accurate that that one one point point calibration is is sufficient to to ensure < 1% 1% accuracy over over T=1,500-3,500 K

48 We Have Measured Temperature Response of W Samples with ns Resolution mj, 8 ns laser pulse Noise Noise in in measured temperature is is due due to to low low signal signal level. level. Temperauter (K) PMT voltage (mv) t = 0 is the Data Acquisition trigger point Time (ns) PMT PMT pulse pulse is is much much sharper sharper than than laser laser or or temperature pulse pulse because because L ~ exp exp ( c/λt) ( c/λt) 0

49 Temperature Measurements of Thermal Diffusion Agrees with ANSYS Results Temperauter (K) Measurement ANSYS Time (ns) Temperature rise rise time time is is sharper sharper than than ANSYS ANSYS because because of of difference in in laser laser pulse pulse shape. shape. The The peak peak temperature of of the the sample sample depends depends on on thermophysicaphysical property of of the the thermo- sample sample (mainly (mainly ρc ρc p ) and p ) and laser laser pulse pulse shape. shape. Thermal Diffusion time time constant, k/ρc k/ρc p, of the p, of the sample, sample, however, should should be be close close to to pure pure W. W. Temperature measurements are are compared with with ANSYS ANSYS calculations with with similar similar peak peak sample sample temperature and and good good agreement has has been been found. found.

50 Melting of Sample Surface is Captured by the Fast Optical Thermometer K: K: Melting Melting temperature of of W Sample surface has melted Sample surface has melted Temperauter (K) mj mj Time (ns)

51 Initial Scan of Sample Peak Surface Temperature with Optical Thermometer K Surface Temperature (K) Laser Energy (mj)

52 Thermometer Data Acquisition System We We have have developed the the software for for downloading, post post processing, and and plotting of of the the thermometer data. data. Same interface is is used used for for both both calibration and and data data acquisition Graphical User Interface Temperature PMT2 signal

53 Thermometer Reliability Test 1: 1: Successive calibration: the the basis basis for for developing calibration protocol.. Test 2: 2: Chamber installation test: test: thermometer is is removed from from calibration stand, mounted in in the the chamber, returned to to calibration stand. Calibration held held in in repeated tries tries Test 3: 3: Long-term reliability, i.e., i.e., how how long long the the calibration is is holding. Successive calibration was was performed repeated in in a days days period. Calibration held held within 1.5%.

54 Sample Exposure Experiments Will Begin in December 2003 We are in the process of moving our laboratory into a much larger area. We are in the process of moving our laboratory into a much larger area. September December 30 October