Observations of solar flares at high frequencies suggest that a considerable fraction of the events present flat or even increasing flux spectra at frequencies above 35 GHz. This imposes restrictions on the gyrosynchrotron emission mechanism and source parameters. We analysed a sample of 115 microwave events in order to investigate their spectra at peak flux. The analysis shows that about 50% of the sample exhibits a flat and 25% an increasing spectrum between 19 and 35 GHz. This class of events is significant and must be considered in the models. In order to better define the characteristics of this class of events it is necessary to carry out observations at frequencies well above 50 GHz with high time resolution and high sensitivity.

We describe a set of millimeter interferometric observations of solar flares carried out in conjunction with GRO experiments during the 1991 June Campaign of the Max'91 Program. We show evidence that millimeter emission probes the most energetic (MeV) electrons in solar flares; we also find that in the same flare there can be both impulsive nonthermal and gradual thermal millimeter emission. Millimeter emission usually occurs at the steep rise phase of the hard X-ray emitting electrons (25-100 KeV). There appears to exist some delay between BIMA mm-emission onset and GRO-BATSE 25-100 KeV X-ray emission. Both results have implications for the particle acceleration process.

This paper will review the use of solar eclipses in the study of the extreme solar limb at sub-millimeter and millimeter wavelengths. This approach has been used to overcome the severe limitation imposed by diffraction upon the resolution attainable by direct solar limb scans at these wavelengths. Strong absorption by water vapor in the Earth's lower atmosphere has necessitated the use of telescopes at high altitude sites or in jet aircraft. Data from several of these experiments will be reviewed, including those from the recent James Clerk Maxwell Telescope observation at a wavelength of 1.3 mm of the eclipse of 11 July 1991. In view of the success of recent measurements in improving the spatial resolution with this technique, several of the ultimate limitations placed upon it by lunar surface roughness and by diffraction at the lunar limb are outlined.

These observations have demonstrated the inadequacy of present phenomenological solar atmospheric models at sub-millimetric source heights. Newer models have been developed to fit the observed extension, brightening and detailed structure of the solar limb by attempting to include the structure of the chromospheric network and its spicular field, and their relative success in doing so will be discussed.

The 11 July 1991 total solar eclipse over Mauna Kea was a unique opportunity to study the limb profile of the 12.32 μm MgI emission line. Our observations used the NASA 3-meter Infrared Telescope Facility,1 and a new Goddard large cryogenic grating spectrometer. Spectra of the line were taken in the slitless mode at second contact. The results show that the emission peaks within ~ 300 km of the 12-μm continuum limb. This agrees with recent theoretical predictions for this line as a NLTE upper photospheric emission feature. However, the increase in optical depth for this extreme limb-viewing situation means that most of the observed emission arises from above T min, and we find that this emission is extended to altitudes well in excess of the model predictions. The line emission can be traced to altitudes as high as 2000 km above the 12-μm continuum limb, whereas theory predicts it to remain observable no higher than 500 km above the continuum limb. The substantial limb-extension observed in this line is qualitatively consistent with limb-extensions seen by other observers in the far-IR continuum, and may be indicative of departures from gravitational hydrostatic equilibrium in the upper solar atmosphere, and/or may result from temperature and density inhomogeneities. The extended altitude of formation of this line enhances its value as a Zeeman probe of magnetic fields.

We present observations of the 11 July 1991 total solar eclipse made from the Caltech Submillimeter Observatory. The 850 μm limb is extended 3380±140 km above the visible limb, and there is a 10% brightening at the extreme limb. The measured limb height agrees with previous work at shorter and longer wavelengths. The run of limb heights with wavelength is well fit by a single electron density scale height. We argue that there is no need to invoke spicule geometry to explain the observations.

We report preliminary results of the partial eclipse seen from the BIMA millimeter–wavelength interferometer in northern California. The use of an interferometer has many advantages over previous single–dish eclipse observations at 3 mm, and in particular it allows a quite different type of measurement of the height of the “3-mm limb” which is direct and precise: we find it to be 11.9″ ± 0.4″ above the optical limb.

Near IR total eclipse measurements have provided clear evidence during both 2nd and 3rd contacts for a limb extension of about 125 km for wavelengths in the range containing the CO fundamental vibration-rotation bands between 4.3 and 5.5 μm, when compared to the limb at nearby shorter wavelengths. This is interpreted as a “flash” spectrum in the CO lines, with the above extension representing the outer level of the CO emission layer. This height can be compared to the τCO = 1.0 level incorporated into recent representative atmospheric models (Ayres and Wiedemann, 1989) which is 90 km above the visible limb for a semi-empirical “hot chromosphere” model (Avrett, 1985) and 220 km for a “cool” radiative equilibrium model based upon work by Anderson (1989).

Infrared images (1.65 μm) of the eclipsed Sim were taken atop Mauna Kea, Hawaii, during the July 11, 1991 total eclipse with an Amber Engineering 128 × 128 InSb array camera. The camera, mounted on a portable solar tracker, had a 3.8-cm, f/2 objective that produced a 4.9° field of view. The primary objective of the experiment was to search for dust or rocky rings around the Sun, previously detected at about 4 R ⊙. High thin clouds, atmospheric dust and aerosols from the June 1991 explosion of Mount Pinatubo in the Philippines, and the overall brightness of the solar corona resulted in a very high infrared background. Despite this, high signal-to-noise radial infrared intensity profiles were obtained of the solar corona from the Moon's limb out to about 10 R ⊙. Preliminary analysis shows some evidence for an enhanced surface brightness between 3 to 4 R ⊙ along the east-west direction, but much fainter than seen in previous solar eclipses. The transition region between the K-corona and the F-corona clearly shows at 2.5 R ⊙, and the surface brightness of the F-corona as a function of radius (from about 2 to 10 R ⊙) can be fit by a simple power law.

The availability of relatively large format IR array detectors is incentive for reexamining the classic question of whether or not there are “dust rings” around the sun – a problem for which there are conflicting observational answers. The 1991 eclipse path included a high altitude observatory and provided a potentially ideal opportunity to study the infrared properties and dust content of the corona. Here we report results from an experiment conducted from Mauna Kea using a HgCdTe array detector sensitive to wavelengths between 1-2.5 μm. Surface brightness measurements in the H-band and polarization data in the J-band were obtained over a field-of-view of ±6 R ⊙ while K-band images further extend to 15 R ⊙ on the western side of the Sun. J-band polarization data and H and K-band surface brightness data clearly show the inhomogeneous structure in the K corona and the ecliptical flattening of the F corona. We see no evidence of a circumsolar, local dust corona (dust rings) out to 15 R ⊙.

Observations of the solar corona in a narrow band filter at wavelength 2.12 μm during the total solar eclipse of 11 July 1991 are described, and compared with results obtained by two observers during the eclipse of November 1966, and shortly thereafter. The lack of any observable signature of thermal emission in the 1991 results suggests that during 1966/67, the near-solar environs were subjected to a locally enhanced dust population, supplied by one or more sungrazer comets. Possible conditions which match the observational circumstances are discussed.

Coronal images were taken in the light of the He I 10830 Å line, the 10000 Å continuum, and the Fe XIV 5303 Å line, with the aim of studying the thermal structure of the corona. In addition, spectroscopic observations were made in the violet wavelength region (3760-4060 Å) and near-infrared (10745-10835 Å), to obtain details of physical conditions of the corona, especially of its cool part. The data obtained do not show any distinct cool structures other than ordinary prominences. Some preliminary results concerning the corona and prominence structures are given.

Preliminary results of the analysis of the white-light, emission (green and red), and far red (600-700 nm) corona during the July 11, 1991 eclipse are given. Even though the corona is of nearly-maximum type, four different principal coronal structures are seen, combined with faint, small-scale structures (loops, arches, cavities, voids or plasmoids). Scattered light is seen up to 10 R ⊙ in helmet streamers. The Ludendorff index of the corona shape turns out to be a + b = −0.02, and the estimated brightness of JK = 1.47 × 10–6B ⊙. Some aspects of multiwavelength observations are discussed.

The total solar eclipse of July 11, 1991 was observed from “La Matanza”, Baja California Sur, México, only 5 km south of the center line of totality, with several small instruments intended to obtain images of the corona during totality, and using a range of exposure times which allowed us to detect both the inner and outer corona. Relations between large and fine scale structures of the corona, the photospheric and chromospheric activity, and the presence of coronal holes are presented.

Three-dimensional simulations of solar convection are described. The simulations show that viewing convection as a hierarchy of eddies does not properly represent the large scale topology. A better picture is to view convection as a broad warm upflow with embedded cool, narrow, downdrafts. These downdrafts penetrate many scale heights through the convection zone and carry most of the net convective flux. Near the solar surface there are extremely large fluctuations in the temperature (5000-11000 K), entropy and pressure (factor of four). Radiation temperature does not provide an accurate measure of the gas temperature at a given geometric depth, because the opacity is very temperature sensitive. The emergent intensity in the infrared is smaller and has a smaller contrast than in the visible. However, in terms of radiation temperature the infrared is hotter and has a higher contrast than the visible.

High spatial resolution observations of the solar granulation up to 3 μm are possible on existing vacuum solar telescopes. They permit the analysis of the deepest photospheric layers near 1.7 μm. A high photometric accuracy is achieved using scanning techniques with a pinhole photometer; imaging methods are used for 2D-analysis. Statistically significant results on granulation, including power spectrum and histogram analysis, center-limb variations and lifetime analysis, are presented. Temperature fluctuations of periods near 5 minute are considered at 1.7 μm, as well as large-scale variations at the scale of the meso- and, especially, the super-granulation. We also discuss the “abnormal” granulation in magnetic regions and the umbral granulation in the cores of sunspots.

Movies of the solar granulation were made simultaneously at 5575 Å and 1.64 μm using the Vacuum Tower Telescope at NSO/SP. A 128 × 128 HgCdTe array was used in the infrared and an RCA 504 CCD in the visible. From the movies, we determine and compare statistical properties of the granulation and seeing conditions.

We report on a first analysis of the horizontal motions in a 45 minute (32 × 22 arcsec2 field of view) granulation time series (movie presented at the present IAU Symposium) obtained at 1.6 μm with the Vacuum Tower Telescope (VTT) of the National Solar Observatory at Sacramento Peak (NSO/SP). High signal/noise flow maps are obtained by use of the local cross correlation technique (November 1986) which incorporates efficient attenuation of seeing and 5-min oscillations. The flow pattern, showing a (≈ 30 arcsec diameter) supergranule with (≈ 8-15 arcsec) mesogranules superposed, is long lived compared to the 45 min of observations. The computed flows (velocity, divergence, vorticity) resemble the ones obtained at visible wavelengths (e.g., by Brandt et al. 1991, November 1989, November and Simon 1988, Simon et al. 1988). The high quality of the the flow maps (due to a large number of selected images (1500), and (supposedly) smaller 5-min oscillations and better seeing conditions at 1.6 μm) allows us to study time evolution (resolution ≈ 15 min) of the details of the flow (spatial resolution ≈ 3 arcsec). An interesting new finding is the short lifetime (< 45 min) of vorticity as opposed to the long lived (≫ 45 min) divergence of the flow. Our study demonstrates the possibility of using the 1.6 μm window to the opacity minimum region to study the horizontal flows at these deep layers of the photosphere.

Observations of the Sun at 2.2 and 3.75 μm have been made at the Pic-du-Midi and Crimean observatories. High resolution (~ 1″) records of the sunspot and facular contrasts at various heliocentric distances as well as those at the extreme solar limb are presented. We find substantial variations in the sunspot-core relative intensity caused by magnetic activity of the spot. The extreme-limb contrasts of faculae at 3.75 μm are strong thus support flux-tube models which result in an enhancement of the extreme-limb facular brightness.

Large amplitude solar 5-min intensity oscillations have recently been detected at 2.23 μm using broad band (650 Å FWHM) photometry (Leifsen and Maltby, 1990). Large intensity amplitudes in a broad range in the near infrared was unexpected, and several questions concerning the source of the high amplitudes were raised. In an attempt to study the nature of these oscillations, time series of spectra have been obtained with the Fourier Transform Spectrometer (FTS) of the McMath telescope at National Solar Observatory at Kitt Peak. We present preliminary results from a 10 day long run in May 1991 in support for the suggestion that the results may be useful in both helio- and asteroseismological investigations.

Near-infrared (0.7-1.0 μm) observations of solar brightness oscillations were performed in 1983-1991 at the Crimean Solar Tower with the use of two (16 × 16 and 32 × 32) photodiode arrays. In 1991 new observations of p-modes were made simultaneously in two spectral ranges, near 0.7 and 1.0 μm. The data is analysed to check for the presence of solar variation with the 160-min period. It is found that the mean relative amplitude for the 160-min solar irradiance variation at 0.73 - 1.65 μm wavelengths, ~ 2 × 10–6, is much lower than the upper limits set by the ACRIM and IPHIR space experiments.