Solar Composition

Long term averages. Presently available long term average composition results have mainly been based on the data from the Ion Composition Instrument (ICI) aboard the ISEE-3/ICE spacecraft (Coplan et al. 1978). The instrument is composed of a velocity filter and electrostatic energy analyzer and has produced on average 30 to 50 mass per charge spectra per day with a maximum resolution of 30 since operation began in August, 1978. Average abundance ratios over a complete solar cycle for He/He, O/He, Ne/He, Fe/He, and Si/He have been derived from the data. These have been converted to ratios with respect to oxygen and are listed in Table 1. The table also contains abundance ratios from other solar wind experiments and solar energetic particle (SEP) (energies of 1 MeV/nucleon) and galactic cosmic rays (energies of 100 to 1000 MeV/nucleon). Also included are abundance ratios for the solar photosphere. The ISEE-3/ICE long term values for the He/He and Ne/He ratios are in good agreement with ratios determined from foil measurements performed on the lunar surface during three separate Apollo missions. There is also good agreement between the ISEE-3/ICE measurements for oxygen, silicon, and iron and abundances derived from the Solar Wind Ion Composition Spectrometer (SWICS) experiment aboard the Ulysses spacecraft, SEPs, and galactic cosmic rays.

Additional insight into solar wind abundances comes from a comparison of solar wind abundance ratios and solar photospheric abundance ratios as a function of the first ionization potential, Figure 1. Also included are the data for SEPs and galactic cosmic rays (because the spectra of the inert gases are absent in the photosphere it is not possible to compare the solar wind values of ratios of He and Ne to He to those in the photosphere). There is a systematic difference between the three sets of abundance data and photospheric abundances. Elements with low first ionization potentials (FIPs) are more abundant than high FIP elements (Meyer 1985, Hovestadt 1979). The transition between low and high FIP elements occurs at about 10 eV, corresponding to Lyman-alpha radiation. It appears that atomic physics is much more important in determining abundances in the solar wind, SEPs and galactic cosmic rays than the details of the structure of the stellar atmosphere from which the particles originate and the acceleration mechanisms that are responsible for their final energies. The similarities in composition obtained by the different techniques provides us the possibility of applying knowledge gained from solar wind studies to SEPs and galactic cosmic rays.

Any model of the solar wind must take into account the conditions in the photosphere, chromosphere, and corona, for it is the photosphere that feeds material up through the chromosphere to the corona, and photospheric abundances are the baseline against which solar wind abundances are compared. The photosphere can be adequately described as a gas in thermodynamic equilibrium with a temperature of approximately 6500 K. The corona, by contrast is much less dense with a temperature of the order of 10K so that the assumption of local thermodynamic equilibrium is sometimes questionable. The mechanism of coronal heating is presently not well understood; even within the transition region there are steep temperature and density gradients. Within the corona, the temperature is a minimum at the base where the density is greatest. Farther up, temperatures increase and densities fall until mean free paths become comparable to 1 AU and the gas is essentially collisionless. The overall result is a dynamic competition as a function of height between collision frequency, charged particle density (ions and electrons), and ionization efficiency. Because the source region for the solar wind is the outer corona, solar wind composition and charge state distributions reflect the complex kinetics within the corona. Moreover, the mechanisms that produce the transition from the relatively cool, dense photosphere to the hot tenuous corona are unknown and constitute one of the important outstanding questions about the sun. While it is well established that ionization first takes place at the base of the corona, the mechanisms by which initially formed ions are preferentially swept upward and incorporated in the solar wind has yet to be fully explained. Current models hypothesize the existence of radial magnetic structures at the base of the corona that confine the ion trajectories to paths that lead into the corona. Neutral atoms, by contrast, are free to diffuse in all directions.

Fractionation in the corona has been modeled by von Steiger and Geiss (1989) who have been able to reproduce a FIP effect of approximately the observed size using fairly simple models of fractionation by ion-atom separation across magnetic field lines in the upper chromosphere. Their mechanism relies upon trapping of charged particles on magnetic field lines. In a constant UV flux, the difference in ionization times between high and low FIP atoms results in the preferential containment of low FIP atoms on magnetic field lines that conduct them up through the corona. This mechanism can reproduce the FIP effect for temperatures of 10K, densities of 10 m, and an overall height scale of 10 to 100 km. The theory also predicts a slight decrease in the abundance of He, in agreement with the data in Table 1. Burgi (1992), has shown that the flow geometry in coronal streamers can also produce strong helium fractionation, and, in absence of a verified source for the slow solar wind, such geometrical effects could probably produce the variabilities seen at low speeds.

The only isotope measurements currently available are those for the He/He ratio (see Table 1). A knowledge of this ratio constitutes a stringent test of models of solar evolution (Maeder, 1990).

Solar Wind from Well Defined Solar Structures. In an effort to learn more about the solar FIP effect and thereby gain additional knowledge about the structure and dynamics of the photosphere, chromosphere and corona, scientists have begun to analyze the composition of solar wind that originates from well defined structures on the sun. So far it has been possible to identify solar wind from coronal mass ejections, coronal holes and sector boundary crossings. These structures are associated with different magnetic field topologies that certainly play a central role in the transport and fractionation of ions and electrons from the chromosphere to the corona and solar wind. In addition to direct solar wind measurements, there are now spectroscopic studies of the abundances of a number of elements in solar flares, active regions and polar coronal holes. These spectroscopic measurements are based on data from a variety of imaging spectrographs operating from the ultra violet to x-ray region of the electromagnetic spectrum. Many of the measurements were taken during the Skylab missions, but only recently have been analyzed in detail (Feldman (1992) and references therein). Most of the work has concentrated on the difference in the abundances of low and high FIP elements under different conditions. The elements that have been analyzed are the low FIP elements, Mg, Na, Fe, Ca, and Si. These have been compared with the high FIP elements Ne and O. The solar flare results are particularly interesting because of the connection between coronal mass ejections and flares. Schmelz (1993), Schmelz and Fludra (1994) have analyzed x-ray spectroscopic observations from the Solar Maximum Mission (SMM) spacecraft and derived abundances for O, Ne, Mg, Si, S, Ca, and F for two flares. These abundances show a gradual, rather than abrupt, change as a function of the FIP, when compared to solar abundances. The magnitude of the FIP effect was approximately two, and there was a negligible change in abundances as a function of time.

Coronal mass ejections (CMEs) are some of the largest scale physical phenomena observed on the sun. They are thought to arise from closed magnetic field structures rooted in the solar corona. These structures become unstable which results in the expulsion of large volumes of coronal plasma into the interplanetary medium (Sheeley et al. 1985). Subsequent rearrangement of the magnetic field on the sun is thought to lead to flares. There have been a number of studies of the composition of coronal mass ejecta that have been incorporated into the solar wind. Identification of CME material has relied on the observation of helical magnetic field structures in the interplanetary medium (Burlaga et al. 1982) and bi-directional streaming of electrons and solar energetic particles along the field (Gosling et al. 1987). Low temperature and high velocity are other signatures. A difficulty with the solar wind studies is the uncertainty in associating the solar wind material with the mass ejection. This is due to the fact that most satellite based instruments have been located very near the earth-sun line and at this position can only intercept material from regions of the sun directly facing the earth. Optical observations of mass ejections from these regions of the sun are particularly difficult to make because of the absence of contrast between the ejection and the solar background. On the other hand, there is a great wealth of optical data on mass ejections originating from the limbs of the sun where the ejections stand out against the background of interplanetary and interstellar space. The CME project, initiated in 1991, sought to take advantage of the position of the ISEE-3/ICE spacecraft off the west limb of the sun to make the observations of CME material as it passed the spacecraft (Richardson et al., 1994). Using optical and radio data first from the SMM spacecraft and then ground stations it was possible to identify periods when CME material passed the spacecraft. Because of limitations in telemetry, the observations were incomplete, but there were sufficient data to identify differences between material from CMEs and the average solar wind. Results of CME composition observations are listed in Table 2.

Sector boundary crossings are open field structures associated with reversals in magnetic polarity on the surface of the sun. It has been established for some time that the alpha particle to proton ratio decreases dramatically at a sector boundary crossing (Borrini et al. 1981), but there is very little data on minor ion composition. Some preliminary results are listed in Table 2.

Coronal holes are another type of structure on the solar surface. These holes are easily identified on synoptic maps of the solar surface; they are associated with open magnetic field configurations, and are known to be the origin of high speed solar wind. Because of the size of the polar coronal holes the high speed solar wind often carries away most of the mass flux and is therefore of great importance. The abundance of alpha particles is remarkably constant in the high speed solar wind, but highly variable in the slow solar wind (Bame et al., 1977). The Charge Energy Mass (CHEM) instrument aboard the AMPTE spacecraft made a number of composition determinations of high speed solar wind that could be associated with coronal holes (Gloecker et al. 1986). The results relied on the fact that high speed solar wind compresses the magnetosphere so that the spacecraft which was normally inside the magnetopause was temporarily in the magnetosheath, where the abundances of multiply charged ions should not differ appreciably from those in the solar wind. From 1978 to 1982 and again after 1985 the ISEE-3/ICE spacecraft was in the solar wind in a position to observe material from equatorial coronal holes. Though between 10 and 50 equatorial coronal holes were observed each year, the Ion Composition Instrument could only acquire meaningful data when the solar wind velocity was less than 600 km/s. Only a small number of samples met this criterion, and the results are included in Table 2.

In Table 2 we compare the composition of solar wind from open and closed magnetic field structures, with the average solar wind, and the photosphere from Table 1. Also included are composition data derived from the Skylab spectroscopic observations that are also associated with well defined magnetic field topologies. There is a clear correlation between the FIP effect and the topologies. This is especially apparent in the spectroscopic data where the Ne/Mg ratio shows a strong dependence on source region magnetic field configuration. These results are in agreement with those of von Steiger et al. (1992) who find that solar wind from cool coronal sources with open magnetic fields is enriched less in C and Mg than that from hotter closed field sources.

To illustrate this further, in Figure. 2 we show a plot of the ratios O/Ne as a function of the Ne/Mg ratio. The nominal values of the ratios for the corona from SEP measurements and for the photosphere are indicated on the graph. The open circles are the spectroscopic values of Feldman and Widing. We see that, as pointed out by Widing and Feldman, the two points describing material associated with flares lie close to the typical photospheric value, while the others cluster about the typical coronal value. The solid circles are results from the ICI, AMPTE, and SWICS mass spectrometers. The ICI average (see Table 1) is identified on the figure together with its uncertainty, which represents real variability as well as measurement uncertainty. The ratio of oxygen to neon as determined with SWICS by Geiss et al. (1994) over a 100 day period at coronal temperatures of 10K and 2x10 K are also shown in Figure 2.