Refractive-Index Measurements of LiF Ramp Compressed to 8 GPa Pressure (GPa) 1 4 68 1.6 1.55 Refractive index D. E. Fratanduono Lawrence Livermore National Laboratory 1.5 1.45 1.4 1.35 Weighted mean Wise Lalone 3 4 5 6 7 Density (g/cc) 53rd Annual Meeting of the American Physical Society Division of Plasma Physics Salt Lake City, UT 14 18 November 11
Summary The refractive index of ramp-compressed LiF depends linearly on density up to 8 GPa Knowledge of LiF s compressed index of refraction is important for high-pressure equation-of-state (EOS) measurements The refractive index of shock-compressed LiF has previously been measured to 115 GPa* Ramp-compressed LiF is measured up to 8 GPa LiF is observed to remain transparent over this range A single-oscillator model suggests that the band gap will close at pressures above 4 GPa these are the highest ever refractive-index measurements of an insulator E19145b J. L. Wise and L. C. Chhabildas, Sandia National Laboratories, Report SAND-85-31C, NTIS (1985).
Outline I. Motivation II. Theory/technique III. Experiments A. Method of characteristics B. VISAR measurements C. Analysis IV. Results A. Refractive index B. Effective oscillator mode C. Metallization E387
Collaborators T. R. Boehly and D. D. Meyerhofer University of Rochester Laboratory for Laser Energetics M. A. Barrios, J. H. Eggert, D. G. Hicks, R. Smith, D. Braun, P. M. Celliers, and G. W. Collins Lawrence Livermore National Laboratory
Motivation There is significant technical utility in ramp-compressed optical windows for material studies Optical velocimetry is a powerful tool that probes the response of materials under dynamic compression Drive Sample Velocimetry Transparent optical windows maintain compression within a sample enabling in-situ particle-velocity measurements for EOS studies Accurate optical velocimetry measurements require that the n(t) of the window is known Drive Sample Transparent window Velocimetry E388
Theory The refractive index of optical windows affects VISAR velocity measurements VISAR detects Doppler shifts from moving surfaces U true n = 1 VISAR Optical windows influence the Doppler shift Changes to the window alter the optical path length of the probe beam U app n 1 VISAR The refractive index is determined if the apparent velocity (U app ) and the true velocity (U true ) are known n shock n U app U shock VISAR E19483a
The transparency of shocks in LiF windows makes it possible for VISAR to probe the material interface U p LiF window U c LiF VISAR n c n Reflecting surface Single shocks up to 16 GPa are transparent in LiF* multishocks up to 5 GPa are transparent VISAR probes through the compressed material; this alters its sensitivity VISAR measures the rate of change of the optical path length corrections must be made for the compressed refractive index (n c ) E1846d *J. L. Wise and L. C. Chhabildas, Shock Waves in Condensed Matter 1985, DEAC4-DP789
Shock experiments are limited by Hugoniot melt and shock entropy Transparent insulators transform into conducting fluids pressure-induced reduction of the band gap and thermal promotion of electrons across that gap Entropy produced by high-pressure shock waves transforms materials into conductive matter highly reflective at the shock front Temperature (1 K) 8 7 6 5 4 3 1 LiF* Boehler melt (1997) SESAME 771 SESAME 771 MD simulations 5 Melt line 1 Isentrope 15 Pressure (GPa) Hugoniot 5 3 E19149 *R. Boehler, M. Ross, and D. B Boercker, Phys. Rev. Lett. 78, 4589 (1997).
Technique A two-section target enables simultaneous measurements of the apparent and true particle velocity to be made Laser ablation Diamond LiF Drive beam U app U fs VISAR VISAR Compressed region Hayes* shows that for ramp compression duapp dn = n t dutrue d t Two measurable parameters are required to determine n(t) U true is determined using specially designed targets and the method of characteristics E1891e *D. Hayes, J. Appl. Phys. 89, 6484 (1).
Method of characteristics and impedance matching determines the true particle velocity Measure U fs (t) Pressure (GPa) 4 3 1 Shot 57575: Applied pressure 1 3 4 Time (ns) Drive beam = P(t)= Diamond U true U fs Compressed region Backward characteristic algorithm to drive surface Calculate P(t) Forward characteristic algorithm to diamond/lif interface Calculate U true (t) E389a Diamond ramp isentrope from D. K. Bradley et al., Phys. Rev. Lett. 1, 7553 (9).
OMEGA Experiments LiF refractive-index experiments were performed using laser-drive ramp compression on OMEGA Laser ablation Diamond Drive beam U app U fs Compressed region VISAR LiF VISAR Image relay from target to interferometer Target Vacuum chamber Probe laser (53 nm) delivered through multimode fiber VISAR* beam splitter Self-emission from target SOP Streak camera Delay element Velocity interferometer Streak camera Experiments used laser energies between 7 to 77 J delivered in 3.7- or 7.-ns ramp profiles VISAR* has a time resolution of <3 ps E18346c *Velocity interferometer system for any reflector
The applied diamond-pressure scaling law is determined from VISAR and laser-power measurements* Pressure (GPa) 1 4 1 3 1 1 Experiments Weighted mean Fit 1 1 1 1 1 Intensity (TW/cm ) 1 P^GPah = 4^! 3h 8I_ TW cmib. 71^!. 1h E19616 *D. E. Fratanduono et al., J. Appl. Phys. 19, 1351 (11).
Diamond free-surface and apparent interface velocity are measured simultaneously with VISAR on OMEGA OMEGA Diamond LiF LiF Diamond Shot 57575 Shot 57575 14 1 1 3 4 5 6 7 Time (ns) Velocity (nm/ns) 1 1 8 6 4 LiF Diamond 1 1 3 4 5 6 7 Time (ns) E1956c
Applied pressure is determined from diamond freesurface velocity using the backward characteristics 7 6 Shot 57575: Free-surface characteristics Stress (GPa) 4 35 4 Shot 57575: Applied pressure Time (ns) 5 4 3 1 3 5 15 1 5 Pressure (GPa) 3 1 1 3 4 Lagrangian depth (nm) 1 3 4 Time (ns) E1957
The applied pressure is forward propagated through the LiF window to infer the true particle velocity Time (ns) 8 7 6 5 4 3 1 Shot 57575: Interface characteristics Diamond Gap LiF 1 3 4 5 6 7 Lagrangian depth (nm) Stress (GPa) 4 35 3 5 15 1 5 Velocity (nm/ns) 14 1 1 8 6 4 Shot 57575: Interface velocity Free-surface velocity Apparent velocity True velocity 1 3 4 5 6 Time (ns) E19169
The Monte Carlo technique is used to estimate errors associated with the calculated true velocity Monte Carlo errors Experimental diamond isentrope standard deviations provided* LiF EOS (SESAME 771) assumed 1% error in pressure Gap thickness between diamond anvil and LiF window Timing errors (of the order of the etalon delay) were incorporated following the Monte Carlo routine Velocity (nm/ns) 8 7 6 5 4 3 1 3.5 True velocity: Shot 57575 4. 4.5 Time (ns) E19155 *D. K. Bradley et al., Phys. Rev. Lett. 1, 7553 (9).
The apparent velocity depends linearly on the true particle velocitiy for 17 experiments Apparent velocity (nm/ns) 1 1 8 6 4 Pressure (GPa) 1 4 6 14 Vacuum gap (17 shots) 4 6 8 1 1 True velocity (nm/ns) E19156d
LASNEX simulations were performed to validate the method of characteristics and address concerns regarding shock formation in the LiF window LASNEX simulations tested the effects of shock formation and gap thickness on the calculated interface velocity It was found that the gap thickness and shock formation do not significantly affect the interface velocity Excellent agreement was observed between LASNEX simulations and the method of characteristics Time (ns) 6 5 4 3 1 Shot 56113 Peak compression Diamond Shock formation Gap LiF 4 6 8 Lagrangian depth (nm) Stress (GPa) 8 7 6 5 4 3 1 E19154
LASNEX simulations show that shock formation in the LiF window does not perturb interface measurements Velocity (nm/ns) 15 1 5 Free-surface velocity Characteristics LASNEX Calculated diamond/lifinterface velocity Characteristics LASNEX 3 Shot 56113 Shot 56113 4 5 3 4 5 Time (ns) Time (ns) Agreement between LASNEX simulations and the method of characteristics validates this technique and uncertainties caused by the vacuum gap. E1961
A second-order, orthogonal-polynominal regression determines the relation of the apparent and true particle velocities Apparent velocity (nm/ns) 1 4 68 1 4 68 18 1.4 Weighted mean Weighted data 16 Wise Wise data 14 Jensen 1.36 Lalone data Orthogonal fit Orthogonal fit 1 1.3 1 8 6 4 Pressure (GPa) 4 6 8 1 1 14 True velocity (nm/ns) Apparent/True 1.8 1.4 Pressure (GPa) 1. 4 6 8 1 1 14 True particle velocity (nm/ns) E193
Results A linear dependence on refractive index is observed Refractive index is determined from the Hayes formula* du du app true = n For linear U app (U true )** n = a + bt dn t d t n = 1.751 (±.8) +.44 (±.) t Refractive index 1 4 68 1.6 1.55 1.5 1.45 1.4 1.35 Pressure (GPa) Weighted mean Wise Lalone 3 4 5 6 7 Density (g/cc) E1916c * D. Hayes, J. Appl. Phys. 89, 6481 (1). ** D. E. Fratanduono et al., J. Appl. Phys. 19, 1351 (11).
The effective-oscillator model is used to interpret the linear dependence of the refractive index and density Only electronic excitations are considered ~ is assumed to lie above the vibrational modes Refractive index 1 8 6 4 Vibrational modes Electronic modes Changes in the refractive index caused by increases in density are due to a shift in the electronic resonance to lower frequency E197 Extinction coefficient, k 1 1 4 1 6 1 8 1 1 1 14 1 16 Frequency (Hz) Frequency of this study E. D. Palik, Handbook of Optical Constants of Solids (Academic Press, Orlando, 1985).
The single-oscillator model accurately describes the dispersion of the refractive index Over 1 solid and liquid insulators obey this model n 1 = E E d E & ~ E d is the dispersion energy E is the single-oscillator energy - for the alkali-halides, E is related to the excitonic energy (E T ) by E - 1.36 E T Previous studies on compressed H and H O to 1 s of GPa have shown that E d is insensitive to changes in density 1/(n 1) 1.9 1.8 1.7 1.6 1.5 1.4 1.3 LiF Dispersion data Linear fit 1. 4 6 8 1 1 14 & ~ (ev ) E19484
An effective-oscillator model suggests LiF will remain transparent to pressures above 4 GPa E t (ev) = E (ev)/1.36 14 1 1 8 6 4 Pressure (GPa) 5 1 3 Weighted mean Exciton energy Hertzfeld closure 4 6 8 1 1 Density (g/cc) E19485 Extrapolation of these results suggests metallization will occur at ~4 GPa.
The effective-oscillator model suggests LiF will remain transparent to pressures above the Goldhammer Herzfeld metallization E t (ev) = E (ev)/1.36 14 1 1 8 6 4 Pressure (GPa) 5 1 3 Weighted mean Exciton energy Hertzfeld closure 4 6 8 1 1 Density (g/cc) P GH (GPa) P OA (GPa) P BSC (GPa) LiF 8 Rbl 17 85 1 Kl 15 115 155 Csl 15 111 1 Xe 15 13 He 11, Ne 158, GH: Goldhammer Hertzfeld criterion OA: Optical absorption BSC: Band structure calculation E19163a High metallization pressure of LiF suggests that it will be a valuable window for high-pressure ramp-compression experiments.
Summary/Conclusions The refractive index of ramp-compressed LiF depends linearly on density up to 8 GPa Knowledge of LiF s compressed index of refraction is important for high-pressure equation-of-state (EOS) measurements The refractive index of shock-compressed LiF has previously been measured to 115 GPa* Ramp-compressed LiF is measured up to 8 GPa LiF is observed to remain transparent over this range A single-oscillator model suggests that the band gap will close at pressures above 4 GPa these are the highest ever refractive-index measurements of an insulator E19145b J. L. Wise and L. C. Chhabildas, Sandia National Laboratories, Report SAND-85-31C, NTIS (1985).