Diffraction, Energetics, and Dynamics at High Pressures 2007 COMPRES Annual Meeting

1. Diffraction, Energetics, and Dynamics at High Pressures 2007 COMPRES Annual Meeting UNLV: L. Borowski, C. Chen, H. Giefers, Y. Ke, E. Kim, R. Kumar, B. Lavina, K. Lipinski-Kalita, M. Nicol, M. Pravica, L. Tanis, O. Tschauner, B. Yulga, P. Weck HPCAT: P. Chow, M. Hu, H. Liu, Y. Meng, W. Yang LANL: L. Daeman, M. Hartl, S. Liu, S. Rudin, S. Vogel, J. Zhang, Y. Zhao APS : A. Alatas, E. Alp, M. Lerche, W. Sturhahn, J. Zhao New Mexico State U: B. Kiefer SRNL: A. Stowe GSECARS: M. Rivers NSLS (now at Yonsei U): Y. Lee Wilber Wright College: W. Pravica Geophysical Laboratory: P. Dera U Arizona: B. Denton, R. Downs ESRF: A.I. Chumakov, U. Ponkratz, R. Rueffer, C. Strohm

2. Dynamics, Diffraction, and Energetics at High Pressures 2007 COMPRES Annual Meeting UNLV: L. Borowski, C. Chen, H. Giefers, Y. Ke, E. Kim, R. Kumar, B. Lavina, K. Lipinski-Kalita, M. Nicol, M. Pravica, L. Tanis, O. Tschauner, B. Yulga, P. Weck HPCAT: P. Chow, M. Hu, H. Liu, Y. Meng, W. Yang LANL: L. Daeman, M. Hartl, S. Liu, S. Rudin, S. Vogel, J. Zhang, Y. Zhao APS : A. Alatas, E. Alp, M. Lerche, W. Sturhahn, J. Zhao New Mexico State U: B. Kiefer SRNL: A. Stowe GSECARS: M. Rivers NSLS (now at Yonsei U): Y. Lee Wilber Wright College: W. Pravica Geophysical Laboratory: P. Dera U Arizona: B. Denton, R. Downs ESRF: A.I. Chumakov, U. Ponkratz, R. Rueffer, C. Strohm

3. ——————— HIPSEC Outline Topics: Brief Overview of HIPSEC Structures, phases, and phonons of Fe-Sn materials by XRD and nuclear resonant inelastic scattering New approaches for diffraction from microcrystals Shear-induced structural transition in PETN Sponsors: DOE-NNSA Cooperative Agreement DE-FC88-06NA27684 US Army RDECOM ACQ CTR Contract W9011NF-05-1-0266 NSF DMR 0521179 HPCAT by DOE [BES,NNSA] The APS and NSLS are supported by DOE-BES Contracts Nos. W-31-1009-Eng-88 and DE-AC02-98CH10886.

4. ——————— HIPSEC • UNLV’s HIPSEC is a university-based teaching and research center supported by DOE-NNSA with the broad goals of: • … advancing weapons materials science at pressures, • temperatures, and strain rates needed to interpret • non-nuclear tests at NTS and to verify design codes, • … preparing scientists and engineers needed by DOE to • assure the effectiveness and safety of the stockpile • without requiring nuclear tests well into the 21st century, • … involving UNLV science and engineering faculty in research • related to critical SSP, DOE lab, and NTS interests, • … helping coordinated growth of materials science and • engineering at UNLV

5. ——————— HIPSEC The HIPSEC Program for 2007-08involves: 11 Regular, Research, and Adjunct Faculty from Physics, Geosciences, and Mechanical Engineering plus 2 other professionals and more than 40 collaborators from other institutions. Seven faculty are on UNLV hard-money appointments. (Two more faculty are being recruited.) 6 Postdoctoral Scholars (1 being recruited) 12 Graduate Students 10 Undergraduate Students 15 Staff at HPCAT Budget – about $3.5 million

6. ——————— HIPSEC • HIPSEC Projects and Topics for this Presentation • High-Pressure Science • We study structures, chemistries, thermal and mechanical properties of materials to pressures more than 106 atmosphere at temperatures from absolute zero to > 8000 oC. FeSn2, NaBH4 • Chemistry of Energetic Materials at High Pressures • We try to understand the chemistries of explosions by determining EOS of explosives and by recovering/identifying decomposition products at high pressures. PETN phases and new Laue methods • Mechanical Properties of Fresh and Aged Foams [ending] • We analyze the kinetics and mechanisms of aging of composite polyurethane foams from the atomic to the bulk scale. • Theory and Simulation • We develop and test models to simulate these experiments Sn • Geoscience • We do experimental and theoretical mineral physics and geophysics High-pressure silicas: stishovite and post-perovskite,

7. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS One of the largest uncertainties in high pressure science is: What is the pressure of any experiment above ~ 4 GPa (40 kbar)? Why is this so? Two measurements give absolute P above 4 GPa: Sound velocities by Brillouin spectra (to 50 GPa at ~1% in one case). Shockwave P-V-E EOS of simple metals to >> 100 GPa. HOWEVER, SHOCK COMPRESSION INVOLVES SIGNIFICANT HEATING; THE FINAL T IS UNKNOWN. CORRECTIONS TO ISOTHERMS INVOLVE MODELS WITH LARGE UNCERTAINTIES ABOUT THERMAL CORRECTIONS. In the original ruby scale paper for >20 GPa, John Shaner estimated that, reductions from the Hugoniots caused systematic uncertainties in P at 300 K of the order of 10%! This has been reduced but not eliminated.

8. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS Things have improved. Silvera, Nellis et al. estimate that deviations of available thermally-corrected shock EOS data from a new pressure scale proposed by them in 2005 for many materials are < 3% at 80 GPa. Work by other authors show similar discrepancies. The largest contributions to the uncertainties are the thermal corrections and shear strength of the shocked material. For Cu at 100 GPa, the correction is from the fluid phase at T > 4000 K! Better understanding of the thermal properties of dense materials, especially their vibrations, are needed to determine how to attribute differences between experiment and models.

9. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS Another approach, which works for powders including high-pressure phases, measures vibrational spectra (densities of states) at high P and variable T by Mössbauer excitation spectra. The basis of this method is that some fraction of γ emission from (or absorbed by) an atom embedded in a solid involves no recoil! In other cases, recoil momentum is conserved by creating or annihilating phonons. The sidebands to the recoilless transitions measure phonon densities of states. We detect absorption (emission) of one or several phonons simultaneous with nuclear excitations by measuring the total nuclear fluorescence as a function of excitation energy. This requires high resolution. For 119Sn in Sn metal at ambient conditions, the γ is ~ 25 KeV, and the phonons are ~5-10 meV! For 57Fe, the γ is ~ 14 KeV High resolution synchrotron beam lines are essential for this work.

10. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS The experiment scans at ~ 1meV resolution and senses with Detector #1 excitation of a resonant nucleus to show a central (no-phonon) peak and side bands involving phonon-assisted excitation or annihilation transitions. A challenge is to deconvolute a one-phonon density of states from the spectra. Source: A.I. Chumakov and W. Sturhahn, Hyperfine Inter. 123/124 (1999) 781.

11. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS Our variant of the Paderborn panaromic cell with a specially machined, low-rise Re gasket used for many of these experiments. By removing much of the excess Re from around the diamonds, most inelastically scattered X-rays reach the APD detectors placed in the large openings.

12. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS Young’s phase diagram of tin shows 2 phases along 300 K isotherm to 10 GPa. A fourth (bcc) phase occurs at higher pressures. We have data for all three phases but discuss today only two phases, each at one pressure. Source: D. Young, Phase Diagrams of the Elements (UC Press, Berkeley, 1991)

13. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS Spectra: Central Peak (0 meV) is no-phonon nuclear transition at ~25 KeV; sidebands involve phonons 1 meV ~ 11 K or 10 meV ~ 110 K Many multi-phonon processes are active under these conditions.

14. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS Spectra: Central Peak (0 meV) is no-phonon nuclear transition at ~25 KeV; sidebands involve phonons 1 meV ~ 11 K or 10 meV ~ 110 K Phonon frequencies are higher at this P. The no-phonon intensity increases, and there are fewer multi-phonon processes. Many multi-phonon processes are active under these conditions.

15. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS From this spectrum, we can evaluate the total single-phonon density of states and compare it with an ab-initio DFT-LDA 27-cell computation by S. Rudin (LANL): The agreement is excellent in view of approximations in the theory and neglect of instrumental broadening.

16. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS A DFT-GGA 54-cell model by Ke and Chen gives nearly the same dispersion relations (left) and DOS (right) as Rudin’s DFT-LDA model! Rudin extended the LGA calculation to 54 cells and came closer to the GGA results. As expected, GGA frequencies are still slightly higher at a given P.

17. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS H. Gieffers et al., Phys. Rev. Lett. 98 (2007)245502.

18. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS

19. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS

20. ——————— HIPSEC Mössbauer Spectra, Phonons, and HP EOS At 1 atm., many multi-phonon processes make it difficult to extract the single-phonon DOS. So Ke and Chen took the reverse approach. They use their model DOS to calculate a spectrum. The agreement except the intensity of the no-phonon peak is excellent.

21. ——————— HIPSEC Next Steps Fe:Sn Compounds and Alloys: XRD and NRIXS Both 57Fe and 119Sn are Mössbauer isotopes that can yield partial DOS for the five Fe:Sn intermetallic compounds at ambient T, P: FeSn2FeSn Fe3Sn2Fe5Sn3 Fe3Sn antiferromagnetic, stableferromagnetic, unstable at ambient T & P and for a high-pressure phase of Fe3Sn2 that Hubertus Giefers discovered. Today I describe some results for Fe3Sn and FeSn2 under pressure.

22. ——————— HIPSEC Fe:Sn Compounds and Alloys: XRD and NRIXS Fe57 NRIS show several trends. Patterns for Fe3Sn, FeSn, and FeSn2 are simpler because each has only on Fe site; Fe5Sn3 has multiple sites. Fe-weighted phonon DOS moves to higher energy with higher Sn content. Alloying expands α-Fe to “negative” pressures.

23. ——————— HIPSEC Fe:Sn Compounds and Alloys: XRD and NRIXS For both compounds compression shifts the 57Fe-weighted DOS to higher energies. Experiments on the 119Sn resonance have begun and will continue as beam time is available.

24. ——————— HIPSEC UNLV (HPCAT)-LANL-CalTech-NMSU: Silicas at High Pressures Tschauner (Liu, Somayazulu)-Luo-Ahrens, Azimow-Kiefer Recovery of stishovite at ambient T from shock-generated amorphous SiO2 Structural polymorphism in Al-bearing magnesium silicate post-perovskite Stishovite and coesite at Canyon Diablo and other sites are taken as evidence of shock impact; but most of the silica recovered in laboratory shock experiments is amorphous. Can these observations be reconciled? What are the implications for laboratory experiments? An XRDstudy of post-perovskite (PPV) Mg0.96Al0.03Fe0.005SiO3 finds yet another PPV structure composed of kinked SiO2 and MgO layers.

25. ——————— HIPSEC UNLV (HPCAT)-LANL-CalTech-NMSU: Silicas at High Pressures Recovery of stishovite at ambient T from shock-generated amorphous SiO2 • Sample history: • Shocked by ring-up to 57 GPa and 2100 K – then recovered to ambient P,T as a mixture of amorphous SiO2 and a new silica phase intermediate between 4- and 6-fold Si-coordinated [Liu, Tschauner, et al. (2004)]. • 2. Six months later, XRD patterns of the sample show no Bragg peaks implying that the silica is completely amorphous! • 3. A clearly amorphous section of the sample was loaded in a diamond-anvil cell with argon as the pressure medium, the pressure was raised in modest increments at ~5 min intervals, and XRD patterns were collected. To 1 0 GPa, no crystalline diffraction features were found.

26. ——————— HIPSEC UNLV (HPCAT)-LANL-CalTech-NMSU: Silicas at High Pressures Recovery of stishovite at ambient T from shock-generated amorphous SiO2 At 13 GPa, crystalline patterns of stishovite (S) and the intermediate (T) phase appear in the patterns: Implication, with time, amorphous material will recrystalize at high P (and, presumably, high T) to yield the materials observed at meteor sites.

27. ——————— HIPSEC UNLV (HPCAT)-LANL-CalTech-NMSU: Silicas at High Pressures Structural polymorphism in Al-bearing magnesium silicate post-perovskite The perovskite structure is observed in many ABO3 systems including Mg metasilicate deep in the Earth’s mantle. Many recent studies report observations of post-perovskite (PPV) phases, but the structure(s) are uncertain. The CaIrO3 structure is a common model. Calculations predict various structures. XRD patterns from Tschauner and Liu’s recent study of PPV Mg0.96Al0.03Fe0.005SiO3 and Kiefer’s calculations find a structure composed of kinked SiO2 and MgO layers. Minor element chemistry and the actual stress regime are plausible reasons for the occurrence of this different structure.

28. ——————— HIPSEC UNLV (HPCAT)-LANL-CalTech-NMSU: Silicas at High Pressures Structural polymorphism in Al-bearing magnesium silicate post-perovskite When viewed along <001> of the relaxed structures, two types of layers are seen as indicated by light and dark shading of the SiO6 – octahedra in each layer. Mg atoms are shown as black circles and unit cells are indicated by thin solid lines. Perovskite (1x1) 2x1 3x1 4x1 5x1 2x2 3x3 CaIrO3-type x0

29. ——————— HIPSEC UNLV (HPCAT)-LANL-CalTech-NMSU: Silicas at High Pressures Structural polymorphism in Al-bearing magnesium silicate post-perovskite Fit to 4x1 structure (left) is significantly better than fit (right) to CaIrO3-type. 5x1 and 6x1 structures also provide better fits than CaIrO3-type.

30. ——————— HIPSEC PETN (Pentanitroerythratol) PETN, C-(CH2ONO2)4, is a moderately sensitive high explosive with several DOE uses. Shock wave studies by J.J. Dick and others imply that a second solid phase of PETN occurs above ~ 5 GPa. The structure of the high pressure phase is not known (WSU comments notwithstanding).

31. ——————— HIPSEC PETN (Pentanitroerythratol) Raman spectra and angularly dispersed X-ray powder patterns forneat PETNat high pressures show an anomaly above ~5 GPa which seems to correlate with a discontinuity in the shock Hugoniot for PETN at similar pressures. However, Raman spectra for PETN under N2 (or Ar) do not show this anomaly to > 11 GPa.

32. ——————— HIPSEC PETN What might be the structure of PETN at high P? Large unit cells challenge state-of-the-art diffraction beam lines for powders in DAC’s at high pressures: Sampling of reciprocal space is limited. Strain and pressure gradients are a problem. Powder statistics are insufficient, and patterns are low resolution (high flux versus high resolution) As a result, (powder) diffraction data from DAC’s constrain structures only partially A combined experimental and computational approach can provide structural solutions

33. ——————— HIPSEC Progress and New Results - PETN At ambient P, the PETN primitive cell contains 4 molecules, each with 29 atoms. ADX powder data with tightly focused synchrotron beams at APS-16-ID-B lack the resolution needed to index the pattern for pure PETN. High-resolution ADX powder data from the unfocused NSLS-X7A allow unambiguous indexing of the high pressure phase, but are too weak to determine atomic coordinates. Ab initio calculations coupled to diffraction studies provide a solution. APS-ID-16-B expt, PETN NSLS-X7A expt, KBr, PETN

34. ——————— HIPSEC PETN Fit of 6.3 GPA XRD pattern from Comparison of XRD pattern NSLS to PETN-III and KBr from NSLS to PETN-I adjusted to 6.3 GPa

35. ——————— HIPSEC PETN Experiments by Profs. Tschauner and Lee and modeling by Prof. Kiefer show that: (1) an orthorhombic P21212, phase becomes stable relative to the tetragonal structure under anisotropicstress above 5 GPa. It is not stable under isotropic stress. (2) volume, atomic positions, and molecular conformation are essentially conserved in the orthorhombic phase, a×a×c  transf = E + (0,0,¼) a×b×c -11 = 22 = (borth – btet)/btet

36. ——————— HIPSEC PETN This figure maps the structure of PETN-III onto PETN-I viewed along (110). Analysis of the structures of PETN-I and –III imply that the transition involves softening of a phonon of B2 symmetry at the -point, making the transition ferroic and sensitive to anisotropic stress in tetragonal a x a plane. This leads easily to twinning defects which might be sources of hot spots The spontaneous strain transforms with the same Irep as the order parameter implies that the transition is ferroelastic.

37. ——————— HIPSEC PETN at higher pressure ADX diffraction patterns of PETN changed again between 27.,5 and 30 GPa. After several hours at 30 GPa, large crystals werefound. These crystals persist. Several months later the sample looked like:

38. ——————— HIPSEC PETN at higher pressures The proof of that we had grown crystals was that we collected their white-Laue pattern + Cross is at the approximate center of the X-ray beam for the white-Laue experiment.

39. ————————— HIPSEC Other Challenge for Micro-diffraction at High Pressures Synchrotrons are state-of-the-art instruments for characterizing structures of by single-crystal X-ray diffraction (SXD); however, classical SXD approaches intrinsically limit the experiments. They require either many sample rotations that are inconvenient when microstructures of larger samples are of interest or the proper crystal structure model is not known. While modern synchrotron optics provide the sub-micrometer spatial resolution needed to determine structures of individual microcrystallites in a powder, traditional instruments do not allow us to take full advantage of this high resolution.

40. ————————— HIPSEC New methods for solving crystal structures from micrograin samples. We have improved on “white” radiation methods by measuring peak energies with an area detector alone by streaking the patterns by a slight ω rotation, and measuring the relative transmissivity of diffracted beams through absorption foils in the beam path. This approach has been developed and tested at HPCAT.

41. ————————— HIPSEC New methods for solving crystal structures from micrograin samples. The slight rotation of the small sample smears the Laue peaks into oscillation Laue analysis (OLA) curves. Each point along a curve has a different energy. The absorption foil causes a sudden drop of intensity along the OLA curve with the energy at the drop point equaling the energy of an absorption edge for the foil material. With multiple foils, energies along OLA curves can be determined ± about 10 eV. This figure compares OLA curves through air with those for a tin absorber.

42. ————————— HIPSEC New methods for solving crystal structures from micrograin samples. This picture compares OLA curves through air with those through tin, niobium, and silver foils. Once the energy of the principal component of the diffracted beam is constrained, higher harmonic contributions can be easily detected and deconvoluted to obtain very good estimates of the structure factor amplitudes needed for structure solutions.

43. ————————— HIPSEC New methods for solving crystal structures from micrograin samples. OLA curves record peak intensities over a large and continuous range of energies. Monochromatic intensity data sets can be extracted by sampling the OLA curves at appropriate points, that are calculated from the orientation matrix. By extracting monochromatic intensity sets at energies close to an absorption edge of the atoms present in the studied sample one can take advantage of MAD phasing and resonant contrast enhancement.

44. ————————— HIPSEC New structural methods for micrograins. The OLA idea can be inverted to yield high-resolution X-ray absorption spectra (XAS) with an absorption foil (not shown) in the beam path before the sample, a micro-monochromator crystal behind the sample, and a slit to mask diffraction from the sample. Diffraction from the oscillating micro-monochromator crystal is recorded on a CCD detector.

45. ——————— HIPSEC NaBH4 at moderate pressures – experiment and theory We have investigated pressure-induced structural transitions in a promising hydrogen-storage material, NaBH4, through DFT calculations and X-ray and neutron diffraction. Calculations confirm XRD observations that cubic α-NaBH4 is stable up to about 6 GPa and orthorhombic γ-NaBH4 occurs above 9 GPa. Between these pressures, calculations, XRD, and neutron diffraction show that a tetragonal β-NaBH4 occurs. Splitting of the (220) line of α-NaBH4 into (200) and (112) lines of β-NaBH4 at 6.3GPa signals the α-β transition, and splitting of the (110) line of the β-NaBH4 into the (002), (210), and (102) lines of γ-NaBH4 at 8.9GPa provides clear evidence of the β-γ transition. However, XRD alone does not provide a clear understanding of these changes because of the delicate energetics and insensitivity to proton positions. Neutron diffraction and calculations are needed to solve the structures.

46. ——————— HIPSEC NaBH4 at moderate pressures – experiment and theory The combination of XRD, DFT and neutron diffraction show that the structures of these phases differ significantly: α-NaBH4β-NaBH4γ-NaBH4

47. ——————— HIPSEC NaBH4 at moderate pressures – experiment and theory Above: Neutron diffraction pattern of β-NaBD4 at 6.3 GPa. Upper right: Ambient compression curve: filled features are calculation; open are experiment. Lower right: Lattice parameters of three phases of NaBH4.

48. Thank you for your attention.