VIEWS: 2 PAGES: 30 POSTED ON: 3/15/2011
Solar Elemental Abundances LPL/NSO Summer School 2008 A. A. Norton Composition of the Sun and meteorites serves as a standard reference for all other solar system, galactic and cosmic chemical/abundance research. Outline of Talk: I. Review: Solar Chemical Composition II. Solar Oxygen Abundance Revision --Summary of Research until 2007 --Changes in 2008 D. Dooling NSO Outreach materials Historical Determination of Solar Abundances In 1800’s, Fraunhofer discovered dark lines in solar spectrum. Kirchoff & Bunsen in the 1850’s found that when chemicals were heated, they emitted at wavelengths coinciding with the solar spectrum. Henry Russell quantified solar abundances of 56 elements using eye estimates of line intensities, 1929 (using the Saha equation and the curve of growth). In 1929, Cecilia Payne showed that almost all middle aged stars have the same composition as the Sun. Nucleosynthesis: processes that create the elements and determine their abundance. Big Bang: The observed abundances of H, He and Li are consistent with the photon to baryon ratio assumed in the Big Bang. % by mass: 75 H, 24 He, 0.01 Li Stellar Nucleosynthesis: The burning of the lighter elements - H, He, C, Ne, O, Si, and the CNO cycle. Explosive Nucleosynthesis - Supernova: produces the elements heavier than Fe. (This image is Kepler’s SN 1604 - composite image from Chandra, Spitzer and Hubble.) Cosmic Ray Spallation: Produces light elements 3He, Li, Be and B as a result of cosmic rays impacting the interstellar medium. Elemental abundances: Solar Photosphere vs Meteorites (CI*) ** QuickTime™ and a TIFF (Un compresse d) decompressor are needed to se e this picture. Meteorites are the oldest solar system objects studied in the lab. Can use 87Rb with a 4.8 x 1010 year half life and it’s daughter 87Sr to determine age. * C1 denotes carbonaceous chondrites that have undergone no or little heating. ** Values from Asplund, Grevesse, & Sauval, 2005 Astrophysics log scale with reference to H, units explained later. Units • Parts per billion by weight (mg of Element/1000 kG) • Mass fractions often quoted X - hydrogen, Y - helium, Z - heavy elements. Z/X or Y/X values often cited. • Parts per billion by atoms (# atoms of Element/billion atoms) • Solar uses the logarithmic astronomical scale - the # of Hydrogen atoms is assumed to be 1012. A(H) = log n(H) = 12 dex A(El) = log (El) = log [n(El)/n(H) ] + 12 • Cosmochemical scale normalizes using the number of Silicon atoms to be 106. http://www.webelements.com Universe Sun Elemental abundances in the Universe, the Sun and the Earth’s Crust. Earth’s Crust Elemental abundances: Solar Photosphere vs Meteorites (CI*) Plot from Holweger 1996 review paper. Explanations (with caveats): Li and Be are fragile nuclei that are depleted in the solar convection zone. C, N and O may have only partially condensed in the solar nebula (they form volatile gases). Difference between solar atmospheric conditions and the laboratory: strong temperature and pressure gradients, plasma is in a strong, anisotropic radiation field and is turbulent, just to name a few. Elemental abundances: Solar Photosphere vs Meteorites (CI*) Plot from Holweger 1996 review paper. Explanations (with caveats): Li and Be are fragile nuclei that are depleted in the solar convection zone. C, N and O may have only partiallyAbundances are inferred, condensed in the solar nebula (they form not measured! volatile gases). Difference between solar atmospheric conditions and the laboratory: strong temperature and pressure gradients, plasma is in a strong, anisotropic radiation field and is turbulent, just to name a few. What are the current dilemmas in solar abundance research? Helium: Not present in photospheric spectrum and is largely lost by meteorites. Values must be inferred from the corona or the solar wind, but these have large uncertainties (lines formed in non-LTE). Best to use models to get He abundance. Lithium, Beryllium & Boron: Can all be burned by nuclear processes. Li at ~2.5 x 106 K. Be at 3 x 106 K. Li is depleted by 160 whereas Be and B are not depleted. Evidence of the depth of the convection zone! It appears the the solar convection cell has reached deep enough to burn Li, but not Be and B. Neon, Argon: Not present in photospheric spectrum and lost by meteorites so there is uncertainty in the values. Carbon, Nitrogen, Oxygen: These elements are lost by meteorites but are found in the photosphere. Their abundances are dependent upon the treatment of the atmospheric conditions - LTE or non-LTE. Oxygen is also a reference line for Ne and Ar, so if its abundance is changed then the abundances of Ne and Ar also scale up or down. Another reason why abundances are important -- The Standard Solar Model (SSM) Assumptions: 1 solar mass, zero age, initial homogenous chemical composition. Equations: Laws of mass, momentum and energy conservation + energy transport and nuclear reactions. Run time: Model is allowed to evolve to current solar age. Crucial: Results need to match observed solar luminosity, radius and mass. The model should reproduce the observed surface composition.* Abundances: Observed surface values assumed to be the initial solar chemical composition. Excepting - H, Li, Be & B -- affected by nuclear burning and diffusion He which is a free parameter and is not observed in photosphere. Astrophysics Importance: Stellar evolutionary calculations are calibrated with respect to the SSM. Uncertainties/Areas for Improvement Uncertainties: Opacities add 10-20% of the uncertainty in the solar model. The other uncertainties include nuclear reaction cross sections values & the elemental abundances. Areas for Improvement: Add convection, rotation, magnetic fields, account for element diffusion. (Note: Solar Model becomes Non-Standard when these are added.) Helioseismology Results: The Standard Solar Model was able to reproduce the radius, mass and luminosity to within 0.1% fairly easily which didn’t motivate additional research. Then helioseismology research was able to measure oscillation frequencies to within hundreds of a %. The solar model now had a more stringent set of observations to satisfy. The Dilemmas continued… Gravitational Settling: It is expected that heavier elements should settle to the base of the convection zone and increase the opacity there. This plot shows the relative difference of the squared sound speed as measured by helioseismology and as predicted with a solar model. Is this due to gravitational settling? If yes, then why do Be and B not show a deficiency in the photosphere? (Graph from first two years of MDI data.) Dark Energy Dark Matter Warning: this talk deals with the Cosmic Web bottom-most rung of Drake’s Ladder, where, sadly, sexiness is Dark Holes low, but on positive side, knowledge content was thought to ExoPlanets be high; even so, a few surprises ---------- still were to be found… Solar Abundances in Perspective: The Visible and the Baryonic Bias A NASA pie chart indicating the proportional composition of different energy-density components of the universe, according to CDM model fits. Roughly 95 % is in the exotic forms of dark matter and dark energy. 70 % or more of the universe consists of dark energy, about which we know next to nothing. “The Solar Oxygen Crisis” • Oxygen is the 3rd most abundant element in the universe. • Oxygen was not created in the Big Bang. • In stars with M ≥ 4Msun, O is created in Carbon burning process. • If M ≥ 8Msun, O is created in Neon burning process. • CNO-II cycle in massive stars creates O. Solar Values of log O (dex) 8.93 Anders & Grevesse 1989 Traditional value - spatially unresolved intensity - 1D semi-empirical model 8.83 Grevesse & Sauval 1998 Revised traditional value 8.66 Asplund et al. 2004 New value - spatially unresolved intensity - 3D theoretical model 8.85 Ayres et al. 2006 Semiemperical 1D model Stellar Values 8.54/8.65 Sofia & Meyer 2001 B stars/F & G dwarfs *A change from 8.93 to 8.63 is a factor of 2 in number densities: ~800 to 400 ppm. Implications of a Revised Oxygen Abundance Good Implications - Lower solar value fits better within galactic environment. Bad Implications - Lower value ruins agreement between predicted (solar interior models) and measured (helioseismology) sound speed. The Solar Oxygen Crisis: Probably not the last word Socas-Navarro & Norton, 2007, ApJ, 660, L153 • Example of Data: Spectro- Polarimeter for Infrared and Optical Regions (SPINOR) • Allows simultaneous observations of multiple lines anywhere in the wavelength range 0.4 to 1.6 microns. • Utilizes the adaptive optics system. • Data from 2004: Rows are Stokes I,Q,U,V, and columns are (left to right): Ca II 849.8 nm, Ca II 854.2 nm, and He I 1083.0 nm. Inversion Codes: Often use a least-squares fitting based on the ME solution of the Unno- Rachkovsky equations of a plane-parallel magnetized radiative transfer of the Stokes line profiles (Skumanich and Lites, 1987). Inputs describing the atomic transition of the spectral line are needed. The inversion code fits nine free parameters: line center, Doppler width, damping coefficient, the line to continuum opacity ratio, the slope of the line source function with optical depth, fill fraction, the magnitude of the magnetic field, the inclination and the azimuth. A New Approach to Measure the Oxygen Abundance • (2007) Spatially resolved (0.7”) spectro-polarimetric observations taken 2006 Fe I lines at 6302 A as well as the O I triplet at 7774 A. • Use inversion codes to get vertical stratification of temperature, density, line of sight velocity and magnetic field for each pixel in field of view. • Produce 3D semiempirical model from Fe I lines. pore Temperature (left) Magnetic Flux Density (right) Use O I observations to determine abundance at each pixel. Synthetic O I profiles were computed at levels of 0.1 dex and the 2 vs log O curve was interpolated to find the minimum with ~0.01 dex accuracy. Results: 8.93 dex - LTE 8.63 dex - NLTE (+/- 0.08 dex) Comparisons: Traditional value (1D) = 8.93 dex (Anders & Grevesse 1989) 3D theoretical NLTE simulation = 8.66 dex (Aplund et al. 2004) Use O I observations to determine abundance at each pixel. Synthetic O I profiles were computed at levels of 0.1 dex and the 2 vs log O curve was interpolated to find the minimum with ~0.01 dex accuracy. Systematic errors become visible as spurious spatial fluctuations! Results: 8.93 dex - LTE 8.63 dex - NLTE (+/- 0.08 dex) Comparisons: Traditional value (1D) = 8.93 dex (Anders & Grevesse 1989) 3D theoretical NLTE simulation = 8.66 dex (Aplund et al. 2004) HELIOSEISMIC CONSEQUENCES •Oxygen provides a lot of opacity around the base of the convection zone. •Lowering the opacity at the base of the CV will increase the region where radiative transport is efficient, thus making the convection zone shallower. •Agreement between solar model and NLTE abundance measurements could be restored if opacity is increased ~10-20% at base of CV. (how? Thermal diffusion? Gravitational settling?) •Basu et al (2007) studied the fine-structure spacings of low degree p modes that probe the solar core. They find the lower abundance values simply aren’t supported. Atmospheric modelers Interior modelers Thanks www.edwebproject.org/ Last year, Tom Ayres showed this slide…. Seismology Constraints Left: noted British helioseismologist frets over low-O, ruining previous excellent agreement with solar interior models. Oxygen accounts for over half of the heavy metal mass fraction Z, and is crucial in the interior opacity. Helioseismology prefers O in the narrow range 640-680 ppm. Tom Ayres also showed this slide…. Evolution of the Solar Oxygen Abundance Over the past decade, or so, solar oxygen abundance has fallen precipitously; Sun is in danger of becoming oxygen free circa 2015 . . . However, the missing Oxygen was found in 2008…. New Observations: Spectropolarimetric measurements in a *sunspot* of O I with a Ni I blend at 6300 Angstroms using SPINOR put the O abundance right back into the traditional range. -- Centeno, Socas-Navarro, 2008, ApJ, still in press! Ni I abundance is well-known due to having more well-suited lines. 1st version of this paper had the O abundance lower than any previous findings, but the referee caught an oversight and the final paper presents an almost traditional value. Or was it? In a paper still in publication, Tom Ayres says “I believe that the issue of the solar metallicity should be considered open, and fair game for further study.” IN CONCLUSION Solar abundances are important. As our closest star, we should be able to agree upon abundances! Every spectral line has its own personality. It is subject to different non-LTE effects, may be affected by convection, etc. Line formation physics! Disagreement on the proper model atmospheres is largely responsible for revisions in the O abundance. Different treatment of spectral lines in differing model atmospheres result in variations in the inferred abundances. It’s not clear how the Oxygen abundance dilemma will be resolved, but it presents us with an opportunity to improve our models and generate discussion between modelers and observers. 2007 2008 EZ Program - Evolve ZAMS Stars Program is Peter Eggleton’s Stellar Evolution Model Modified by Bill Paxton QuickTime™ and a MPEG-4 Video decompressor are need ed to see this picture.
"HAO Science Objectives for HMI"