4.3 Ultraviolet spectroscopy

Ultraviolet spectroscopy of the corona is a powerful tool for obtaining detailed empirical descriptions of solar plasma conditions (see, e.g., Kohl and Withbroe, 1982Jump To The Next Citation PointWithbroe et al., 1982). Coronal holes, being the lowest density regions of the outer solar atmosphere, exhibit a complex array of plasma parameters due to their nearly collisionless nature. As a result, every ion species tends to have its own unique temperature, its own type of departure from a Maxwellian velocity distribution, and its own outflow speed. After a brief summary of near-limb measurements made with the SUMER instrument on SOHO, this section mainly describes results at larger heights (in the more clearly collisionless extended corona) from UVCS.

The un-occulted SUMER (Solar Ultraviolet Measurements of Emitted Radiation) spectrometer has observed off-limb regions up to heights of approximately 1.3 R ⊙ in coronal holes (Wilhelm et al., 1995200020042007). In polar regions at solar minimum, ion temperatures exceed electron temperatures even at r ∼ 1.05R ⊙, where densities were presumed to be so high as to ensure rapid collisional coupling and thus equal temperatures for all species (Seely et al., 1997Tu et al., 1998Landi and Cranmer, 2009Jump To The Next Citation Point). Spectroscopic evidence is also mounting for the presence of transverse Alfvén waves propagating into the corona (Banerjee et al., 19982009Dolla and Solomon, 2008Landi and Cranmer, 2009Jump To The Next Citation Point).

Electron temperatures derived from line ratios (Habbal et al., 1993David et al., 1998Doschek et al., 2001Wilhelm, 2006Jump To The Next Citation PointLandi, 2008Jump To The Next Citation Point) exhibit relatively low values in the off-limb coronal holes (Te ∼ 800, 000 K) that are not in agreement with higher temperatures derived from “frozen-in” in situ charge states (Ko et al., 1997). It is difficult to reconcile these observations with one another in the absence of either non-Maxwellian electron velocity distributions or strong differential flows between different ion species near the Sun (Esser and Edgar, 2000Jump To The Next Citation Point). However, some models consistent with both the SUMER temperatures and the in situ charge states are being produced (e.g., Laming and Lepri, 2007).

Figure 6View Image displays a range of temperatures measured in coronal holes and high-speed wind streams, and it shows how the SUMER electron temperatures (Te) are noticeably lower than the heavy ion temperatures (e.g., shown for the O+5 ions that correspond to the bright O vi 1032, 1037 Å spectral lines) even in the low corona. The degree of agreement between the spectroscopic measurements and one-fluid models of the low corona (i.e., the semi-empirical model of Avrett and Loeser (2008Jump To The Next Citation Point) and the theoretical model of Cranmer et al. (2007Jump To The Next Citation Point)) depends on the height of the sharp transition region between chromospheric (104 K) and coronal (106 K) temperatures.

View Image

Figure 6: Radial dependence of empirical and model temperatures in polar coronal holes and fast wind streams. Mean plasma temperatures from a semi-empirical model (dashed black curve; Avrett and Loeser, 2008) and from a turbulence-driven coronal heating model (solid black curve; Cranmer et al., 2007Jump To The Next Citation Point). Te from off-limb SUMER measurements made by Wilhelm (2006) (dark blue bars) and Landi (2008) (light blue bars), Tp from UVCS measurements assembled by Cranmer (2004bJump To The Next Citation Point) (see text), and perpendicular O+5 ion temperatures from Landi and Cranmer (2009Jump To The Next Citation Point) (open green circles) and Cranmer et al. (2008Jump To The Next Citation Point) (filled green circles). In situ proton and electron temperatures in the fast wind (r > 60R ⊙) are from Cranmer et al. (2009).

The UVCS instrument on SOHO is a combination of an ultraviolet spectrometer and a linearly occulted coronagraph that observes a 2.5R ⊙ long swath of the extended corona, oriented tangentially to the radial direction at heliocentric distances ranging between about 1.5 and 10R ⊙ (Kohl et al., 19951997Jump To The Next Citation Point1998Jump To The Next Citation Point2006Jump To The Next Citation Point). In coronal holes, UVCS measurements have allowed many key details about the velocity distributions of H0, O+5, and Mg+9 to be derived. For the resonantly scattered emission lines seen at large heights with UVCS, the most straightforward plasma diagnostic is to use the Doppler-broadened line width as a sensitive probe of the overall variance of random particle motions along the line of sight. In other words, measuring the line width provides a constraint on the so-called “kinetic temperature” (i.e., a combination of microscopic stochastic motions and macroscopic [but unresolved] motions due to waves or turbulence) along the direction perpendicular to the (nearly radial) magnetic field lines.

In the ionized solar corona, a given hydrogen nucleus spends most of its time as a free proton, and only a small fraction of time as a bound H0 atom. Thus, the measured plasma properties of neutral hydrogen are considered to be valid proxies of the proton properties below about 3 R⊙ (Allen et al., 2000). Spartan 201 and UVCS observations of the H i Lyα emission line in coronal holes indicated rather large proton kinetic temperatures in the direction perpendicular to the magnetic field (Tp⊥ ∼ 3 MK) and also the possibility of a mild temperature anisotropy (with Tp⊥ > Tp∥) above heights of 2– 3R ⊙ (Kohl et al., 1997Jump To The Next Citation PointCranmer et al., 1999bJump To The Next Citation PointAntonucci et al., 2004Kohl et al., 2006Jump To The Next Citation Point).

UVCS observations indicated that the O+5 ions are much more strongly heated than protons in coronal holes, with perpendicular temperatures in excess of 200 MK (see Figure 7View Image). This exceeds the temperature at the central core of the Sun by an order of magnitude! The UVCS measurements also provided signatures of temperature anisotropies possibly greater than T ∕T ≈ 10 ⊥i ∥i (e.g., Cranmer et al., 1999bJump To The Next Citation Point2008Jump To The Next Citation Point). The measured kinetic temperatures of O+5 and Mg+9 are significantly greater than mass-proportional when compared with protons, with Ti∕Tp > mi∕mp (see also Kohl et al., 19992006Jump To The Next Citation Point). The surprisingly “extreme” properties of heavy ions in coronal holes have led theorists to develop a number of new ideas regarding the heating and acceleration of the solar wind; these are discussed further in Section 5.4.

View Image

Figure 7: Combined image of the solar corona from 17 August 1996, showing the solar disk in Fe XII 195 Å intensity from EIT (yellow inner image) and the extended corona in O vi 1032 Å intensity from UVCS (red outer image). Axisymmetric field lines are from the solar-minimum model of Banaszkiewicz et al. (1998Jump To The Next Citation Point), and O vi emission line profiles (bottom) are from SUMER (Warren et al., 1997, left) and UVCS (Kohl et al., 1997Jump To The Next Citation Point, right).

Figure 6View Image shows UVCS perpendicular temperatures for protons and O+5 ions in coronal holes. The O+5 data points were taken from the recent re-analysis of solar minimum data from 1996 – 1997 by Cranmer et al. (2008Jump To The Next Citation Point). The proton temperature data were assembled by Cranmer (2004b) from a number of individual measurements of the H I Lyα profile at solar minimum. The sources of these measurements are: Cranmer et al. (1999bJump To The Next Citation Point) (squares), Esser et al. (1999) (diamonds), Zangrilli et al. (1999) (asterisks), and Antonucci et al. (2000) (triangles). The kinetic proton temperatures are of order 2 – 3.5 MK, but in Figure 6View Image we attempted to remove the contribution of nonthermal wave broadening. The semi-empirical model of Cranmer and van Ballegooijen (2005Jump To The Next Citation Point) was used to specify the amplitude of transverse Alfvén waves as a function of height, and their contribution to the line widths was subtracted. The remaining “microscopic” T⊥p does not show as clear a signal of “preferential” proton heating as would be apparent from the larger kinetic temperature. Although one can still marginally see that Tp > Te, the existing measurements of Tp and Te do not fully overlap with one another in radius. Improved measurements are needed in order to better constrain the proton and electron heating rates in the corona.

The UVCS emission line data contain information about the Doppler motions of atoms and ions along the magnetic field (i.e., transverse to the line of sight). The so-called “Doppler dimming” diagnostic technique provides constraints on both the bulk outflow speed along the field and the parallel kinetic temperature (for more details, see Kohl and Withbroe, 1982Noci et al., 1987Kohl et al., 2006Jump To The Next Citation Point). In coronal holes, Doppler-dimmed line intensities from UVCS are consistent with the outflow velocity for O+5 being larger than the outflow velocity for protons by as much as a factor of two at large heights (Kohl et al., 1998Li et al., 1998Cranmer et al., 1999bJump To The Next Citation Point). Figure 8View Image illustrates the outflow speeds measured by UVCS in coronal holes, and compares with the theoretical model of the fast solar wind presented by Cranmer et al. (2007Jump To The Next Citation Point). Also shown for comparison are observational and theoretical data for the slow solar wind associated with equatorial helmet streamers at solar minimum.

View Image

Figure 8: Radial dependence of solar wind outflow speeds. UVCS Doppler dimming determinations for protons (red; Kohl et al., 2006Jump To The Next Citation Point) and O+5 ions (green; Cranmer et al., 2008Jump To The Next Citation Point) are shown for polar coronal holes, and are compared with theoretical models of the polar and equatorial solar wind at solar minimum (black curves; Cranmer et al., 2007Jump To The Next Citation Point) and the speeds of “blobs” measured by LASCO above equatorial streamers (open circles; Sheeley Jr et al., 1997).

In contrast to many prior analyses of UVCS data, which concluded that there must be both intense preferential heating of the O+5 ions and a strong field-aligned anisotropy, Raouafi and Solanki (2004), Raouafi and Solanki (2006), and Raouafi et al. (2007Jump To The Next Citation Point) reported that there may not be a compelling need for O+5 anisotropy depending on the assumptions made about the other plasma properties of the coronal hole (e.g., electron density). However, Cranmer et al. (2008) performed a detailed re-analysis of these observations and concluded that there remains strong evidence in favor of both preferential O+5 heating and acceleration and significant O+5 ion anisotropy (in the sense T⊥i > T∥i) above r ≈ 2.1R ⊙ in coronal holes. In determining these properties, it was found to be important to search the full range of possible ion temperatures and flow speeds, and not to make arbitrary assumptions about any given subset of the parameters.

The UVCS results discussed above are similar in character to in situ measurements made in the fast solar wind, but they imply more extreme departures from thermodynamic equilibrium in the extended corona. For example, proton velocity distributions measured in the fast solar wind between 0.3 and 1 AU have anisotropic cores with Tp ⊥ > Tp∥, and their magnetic moments increase with increasing distance; this implies net input of perpendicular energy on kinetic scales (e.g., Marsch et al., 1982Marsch, 2006Jump To The Next Citation Point). Many heavy ions appear to flow faster than the protons by about the local Alfvén speed (Hefti et al., 1998Reisenfeld et al., 2001), and in the fastest solar wind streams they are also heated preferentially in the same sense as in the corona (Collier et al., 1996).

In the years since the solar minimum of 1996 – 1997, UVCS observed a large number of other coronal holes that appeared throughout the maximum of solar cycle 23 and the new-millennium solar minimum of 2007 – 2009. UVCS tends to observe only the largest coronal holes, since when the smallest holes rotate to the solar limb their UV line profiles tend to be contaminated by emission from streamers in the foreground and background. This selection effect naturally screens out small coronal holes that have been correlated with slow solar wind streams at 1 AU (Nolte et al., 1976Neugebauer et al., 1998). In the cases where UVCS and in situ measurements were made for the same regions associated with large coronal holes, high speeds in excess of 600 km s–1 were inevitably seen in interplanetary space (Miralles et al., 20042006). However, the O+5 outflow speeds measured in the extended corona via Doppler dimming have showed a substantial range of variation. For example, Miralles et al. (2001b) found that the outflow speeds at 2– 3R ⊙ in an equatorial coronal hole were approximately three times lower than those measured in the polar coronal holes from 1996 – 1997 at similar heights (see also Poletto et al., 2002). This implies that the range of coronal heights over which the acceleration of the solar wind occurs can vary greatly, even when the wind at 1 AU ends up similarly fast.

UVCS also has measured the plasma properties in bright polar plumes. The densest concentrations of polar plumes along the line of sight are seen to exhibit narrower line widths – i.e., lower kinetic temperatures – than the lower-density interplume regions (Kohl et al., 1997Noci et al., 1997Kohl et al., 2006Jump To The Next Citation Point). Similarly, plumes are seen to have lower outflow speeds than the interplume regions (Giordano et al., 2000Teriaca et al., 2003Raouafi et al., 2007), although at low heights the data are not as clear-cut (e.g., Gabriel et al., 2003). UVCS also has put constraints on the plume/interplume density contrast and the filling factor of polar plumes in coronal holes. Cranmer et al. (1999b) used a large number of UVCS synoptic measurements to determine statistically that, at heights around r ≈ 2R ⊙, the plume/interplume density ratio is approximately 2, and polar coronal holes are comprised of about 25% plume and 75% interplume plasma (corresponding to ∼ 40 individual plumes distributed throughout the coronal hole). Earlier measurements made closer to the limb showed a higher density contrast and a smaller filling factor, so the UVCS data are generally consistent with lateral expansion of polar plumes with increasing distance.

It is still relatively unknown how much of the mass, momentum, and energy flux of the fast solar wind comes from polar plumes. Despite that uncertainty, though, there have been several reasonably successful models of polar plume formation. Wang (19941998) presented models of polar plumes as the extensions of flux tubes with concentrated bursts of added coronal heating at the base – presumably via nanoflare-like reconnection events in X-ray bright points (see also DeForest et al., 2001). In these models, the extra basal heat input is balanced by conductive losses to produce the larger plume density. The heating rate in the extended corona is not affected by the basal burst, but the larger density in the flux tube gives rise to less heating per particle at all heights, which leads to lower temperatures in the extended corona and a smaller gas pressure force for solar wind acceleration. This model is consistent with the smaller temperatures and outflow speeds measured in plumes with UV spectroscopy.

UVCS made the first spectroscopic measurements of polar jets in coronal holes (Dobrzycka et al., 20002006). The events observed by EIT, LASCO, and UVCS during the 1996–1997 solar minimum tended to be “cool jets” with higher densities, lower temperatures, and faster outflows than the surrounding coronal holes (see also Wang et al., 1998). More recently, Hinode has observed a new population of “hot” X-ray jets in coronal holes (Culhane et al., 2007Cirtain et al., 2007Shimojo et al., 2007Filippov et al., 2009). UVCS found that some of these events persist up to heights of at least 1.7R ⊙ and that the jet protons remain hotter than the surrounding coronal hole (Miralles et al., 2007). Thus, there appear to exist two distinct kinds of polar jets (cool and hot), with differences possibly related to the relative degrees of heating and adiabatic expansion of the jet parcels. Jets and plumes have roughly similar angular sizes and intensity contrasts near the solar limb, and there is growing evidence that they share a common origin (e.g., Raouafi et al., 2008). It is possible that the only substantial difference between the two phenomena is the duration of the bursts of basal heating; i.e., jets seem to be the result of short-lived bursts of heating, whereas plumes may be the product of base-heating events that last longer than several hours.

  Go to previous page Go up Go to next page