1. Spectral Analysis of Archival SMM Gamma-Ray Flare Data Gerald Share 1,2, Ronald Murphy 2, Benz Kozlovsky 3 1 UMD, 2 NRL, 3 Tel Aviv Univ. Supported under NASA Grants NNX07AH81G, NNX07AO74G, and NNG06GG14G.

2. We have re-evaluated the SMM gamma-ray instrument response using a Monte Carlo calculation. We have also completely updated the nuclear gamma-ray line production code using both recent experimental cross sections and the TALYS nuclear reaction code that also provides the ‘continuum’ arising from thousands of weak lines (Murphy et al. 2009). We study the 89 October 19 solar flare as an example of the progress made. These same analysis procedures will be applied to RHESSI flare data when the new DRM has been installed in SSW.

3. Example of the spectra produced by the modified gamma-ray line production code for the interaction by protons and alpha particles following a power-law spectrum with index 4.

4. We fit flare spectra with pl + pl*exp (bremsstrahlung), 0.5, 2.2 MeV,  – 4 He lines, a solar Compton-scattered 2.2 MeV continuum, individual line templates for p,  interactions on C, N, O, Ne, Mg, Si, and Fe, individual line templates for accelerated C, N, O, 3 He, and a combined template (coronal composition normalized to Fe) for Ne, Mg, Si, and Fe on H and He. We have included a 5% systematic in the data >1 MeV. These nuclear line templates were calculated for the heliocentric angle of the flare and for various values of  /p and ion spectral index (assuming the same index for all accelerated particles; no significant difference found when allowing the alphas and protons to have different indices in some test cases.)

5. Fits to the 1989 October 19 flare showing the total fit, bremsstrahlung (dashed/dotted blue), direct interaction (i.e. accelerated p and  on ambient elements; dotted red , p and a on neon (dot/dash red), inverse interaction (i.e. heavy ions on ambient elements; dashed red),  -4He, and 0.511/2.223 MeV line components (dashed green) and Compton-scattered 2.2 MeV line and positronium continuum (dotted green).

6. 1989 October 19 chi2 contour plots for spectral index and  /p ratio. From Monte Carlo studies a   2 = 3.8 has a 68% probability. Upper plot: Map for fit using Asplund ambient abundances and coronal accelerated particles abundances. Lower plot: Map for fit where all the elemental abundances are free to vary (combined template for accelerated Ne, Mg, Si, Fe). Red (blue) curves show the indices and  /p ratios consistent with photospheric He/O for Asplund 2004 (Grevesse 1998) abundances based on fits to the flux measured in the  -4He fusion lines (dotted curves show the 1 sigma statistical uncertainties). Green curves show the locus of points where the Ne/O ratio is 0.15 (top) and 0.20 (bottom).

7. The ambient abundances relative to O for the 89 October 19 flare compared with 2 photospheric models and the SEP derived coronal abundance. The C/O ratio is consistent with all. Mg and Fe are in better agreement with photospheric abundances while Si is consistent with coronal abundances (caveat the Si line is near the strong 2.2 MeV backscatter peak).

8. Accelerated abundances relative to O for the 89 October 19 flare compared with 2 photospheric models, the SEP-derived coronal and heavy-enhanced impulsive SEP abundances. The C/O ratio is consistent with all. Ne, Mg, Si, and Fe (heavy) have been added together and normalized to Fe. The heavies are consistent with coronal or photospheric abundances.

9. Best fit to the spectrum of gamma-ray observed from the Earth’s atmosphere on 89 October 20 from impact of ESP particles associated with a shock reaching the Earth. We used the proton/alpha gamma-ray production code. Most of the lines are in the data. The fit will improve signifiantly when we include gamma-ray lines from secondary neutrons produced in the atmosphere.

10. Fit to an “electron rich” interval in the 89 March 06 flare showing that the best fitting PL + PL*exp bremsstrahlung from 300 to 8500 keV extrapolated to higher energies is in reasonable agreement with SMM high-energy data (Chupp and Dunphy, 2009)

11. Summary of Results Fit 89 October 19 flare spectrum with bremsstrahlung, 0.511/2.2 MeV lines, and nuclear templates developed from empirical measurements and TALYS using a new instrument response. Obtained excellent fits. Ambient Mg and Fe abundance ratio consistent with photospheric abundances while ambient Si abundance appears to be coronal. Best combined fit is for a photospheric ambient abundance. The average Ne/O abundance in the lower chromosphere has been measured directly for the first time with a value of ~0.15-0.20. Accelerated heavy ion (Ne, Mg, Si, and Fe)/O abundance ratio consistent with corona and photosphere but not impulsive SEPs. Accelerated alpha/proton ratio (~0.2 +- 0.1 ) is about twice that of the photospheric H/He ratio and several times that found in the corona. Accelerated heavy ion/proton ratio (50% uncertainty) is about twice that of the (Z > 2)/H ratio in the photosphere. Variability in accelerated and ambient composition from flare to flare

12. Measured Compton Scattered 2.2 MeV Line at the Sun from 1989 October 19 Flare Scattered Flux (>300 keV)/2.223 MeV Line Flux = 0.7 +- 0.2 Representive calculated ratio for the continuum >200 keV ~1.2 This is the first measurement of the 2.223 MeV Compton scattered flux.

13. TALYS is most reliable for particle production reactions involving heavy nuclei; less so for gamma- ray production reactions involving light nuclei. We have found that the best approach is to use the energy dependence of the cross section supplied by TALYS but to normalize the cross section whenever possible to measurements. Not all light-element branching ratios are correct: there are some excited states that decay primarily via particles but TALYS treats as a gamma-ray transition. We have corrected the appropriate TALYS library files.

14. n + 1 H → d +  2.223 The gamma-ray line from neutron capture by 1 H to form deuterium is very narrow (measurements show it is much narrower than the instrumental width of even high-resolution Ge detectors), and it is centered precisely at the laboratory line energy of 2.223 MeV. This implies that the neutrons are captured on 1 H at very low energies.

15. This is because essentially all of the neutrons are thermalized before being captured on 1 H since (1)the elastic-scattering cross section is much larger than the capture cross section, and (2)the scattering particles have the same mass, resulting in optimal energy loss per scattering. Results of Monte Carlo calculations show that, on average, neutrons are thermalized in ~30 scatterings, but ~300 scatterings are required for capture.

16. Because the cross section for neutron capture on 3 He is never larger than that for capture on 1 H, essentially all captures on 3 He will also occur after thermalization. The flux ratio of the two capture lines is then proportional to the two capture cross sections at thermal energies, regardless of the production kinetic-energy spectrum of the neutrons. At thermal energies, this ratio is  ( 3 He)/  ( 1 H) ≅ 10 –4. This, combined with the abundance ratio [ 3 He]/[ 1 H] ≅ 3 × 10 –5, implies that the flux ratio is expected to be less than 10 –8.