Conclusions. 1. Plasma parameters in the arcade: Ne ~10 10 cm -3 , Te ~6 MK.

1. Fig. 2. The overlay of various images of the arcade. Background: composite of SOHO/EIT 195 Å (blue) and Yohkoh/SXT (red-yellow) images.White contour: NoRH 17 GHz (104 K), green contours: SPIRIT 8.42 Å. Green ellipse shows the NoRH beam. Fig. 1.The posteruptive arcade observed on October 22, 2001 in various emissions. Fig. 5. Difference SXT frames showing dark features moving downward onto the arcade. Fig. 4. Same as in Fig. 2 overlaid by contours of the degree of polarization computed from NoRH images averaged within 06:15-06:27 UT (solid positive). Fig. 3.Extrapolated in the potential approximation coronal magnetic fields overlaid with red contours of the NoRH total intensity & the degree of polarization (white contours positive). 08:21-08:58 UT 06:29 UT Simulation from soft X-rays NoRH Fig. 7. NoRH image at 06:30 UT (right) and a ‘radio image’ simulated using EM and Te inferred from Yohkoh/SXT data two hours later (left). Differences are due to 1) continuing activity within the two hours, 2) limited dynamic range of measurements with Yohkoh/SXT. Fig. 6.Temperature distribution over the arcade region (top; blue shows unreliable regions) and dark features moving downward (bottom, three short dark lanes). The arrows indicate the direction of their motion. Observations of a posteruptive arcade on October 22, 2001 with CORONAS-F, other spaceborne telescopes, and in microwaves V.V.Grechnev1, V.N.Borovik2, O.I.Bugaenko3, S.A.Bogachev3, I.Y.Grigorieva2, S.V.Kuzin4, S.V.Lesovoi1, M.A.Livshits5, A.A.Pertsov4, G.V.Rudenko1, V.A.Slemzin4, A.I.Stepanov5, K.Shibasaki6, A.M.Uralov1, V.G.Zandanov1, I.A.Zhitnik4 1Main Astronomical Observatory, St. Petersburg (Pulkovo), Russia; 2Institute of Solar Terrestrial Physics, Irkutsk, Russia; 3Sternberg Astronomical Institute, Moscow, Russia; 4P.N.Lebedev Physical Institute, Moscow, Russia; 5IZMIRAN, Troitsk, Russia; 6Nobeyama Radio Observatory, Japan Abstract. The CORONAS-F/SPIRIT sometimes observes large-scale hot features existing for many hours high in the corona. We identify such a feature observed on October 22, 2001 at ~ 100 Mm with a posteruptive arcade. Using multi-spectral data, we estimate plasma parameters in the arcade and reveal the coronal large-scale magnetic field configuration. Several hours after the eruption, T = 6 MK, Ne ~ 1010 cm-3. We state a problem of long-term existence of the hot arcade high in the corona: either the magnetic field surrounding the arcade well exceeds 20 G at that height, or  > 1 both inside and outside the arcade. A downflow observed in soft X-rays can also contribute to the equilibrium conditions. Instrumentation. The CORONAS-F space solar observatory (Oraevsky & Sobelman 2002, Astron. Letters, 28(6)) was launched in August 2001. The SPectrographIc X-Ray Imaging Telescope (SPIRIT; Zhitnik et al.. 2002, ESA SP-506) aboard the CORONAS-F contains several EUV and soft X-ray telescopes and spectroheliographs, which provide full-disk solar images in several bands: 175 Å (FeIX-XI, Te  1.5 MK), 284, 304, and 8.42 Å (MgXII,Te  5–15 MK). Observations.The SPIRIT sometimes observes in a high-temperature MgXII 8.42 Å line large-scale features existing many hours high in the corona (Zhitnik et al. 2003, MNRAS, 338). The sensitivity range in this line has a maximum of 8–10 MK. Those features look like a ‘ball’ or spider with legs directed downwards. Such a feature was observed on October 22, 2001 at 10:13–10:22 UT above the SW limb at 115 Mm (Fig. 1). It adjoins to the posteruptive arcade observed in cooler EUV lines from above and is cospatial with an arcade observed in soft X-rays with Yohkoh/SXT (Fig. 2). Hence, it is really an arcade formed in the course of CMEs occurred that day at 00:50 (SF/M1.0 subflare), 03:26, and 09:2l at that position angle, likely in the AR 9658 near or behind the west limb. The diverge-temperature soft X-ray and EUV images closely adjoin and even interlace each other. This shows that temperature gradients are high, and the temperature increases with height in the arcade (cf. Ichimoto et al. 1993, Proc. 2nd Japan-China seminar on Sol. Phys.) Dissimilar appearance of the arcade in different emissions is due to different observation times, temperature ranges, and instrumental functions. The Nobeyama Radioheliograph (NoRH) shows the arcade at 17 GHz and the Siberian Solar Radiotelescope (SSRT) at 5.7 GHz. The RATAN-600 radio telescope shows a weak off-limb source cospatial with the arcade in high-sensitivity one-dimensional scans at many frequencies from 2.8 to 15.6 GHz around 09 UT. This source is distinctly polarized which promises the possibility to reveal the large-scale magnetic configuration in the corona.To detect very weak polarization in NoRH images, we produced them with a long integration time of 120 s (normally 1 s) and averaged several images. This results in a sensitivity of ~ 20 K. The degree of polarization shows a short-term appearance of a positively polarized region southward of the arcade (Fig. 3), whereas the dominant negative polarization northward of the arcade persists (Fig. 4). Within the arcade it is close to zero.The polarization measurements with NoRH and RATAN agree. Dark features moving downward are seen in Yohkoh/SXT difference images (1 and 2 in Fig. 5; cf. McKenzie & Hudson 1999, ApJ, 519L). Estimations and Simulations. Temperature and emission measure in the arcade 8 hours after the first eruption were estimated from Yohkoh/SXT data. The temperature distribution is shown in Fig. 6 (top). The main part of the arcade has a temperature of 6 MK, and its upper part 8 MK. Column emission measure within a 2.46  2.46 region is 2.51045 cm-3 for the brightest region and for resolved loop legs ~ 51044 cm-3. Plasma density within the brightest part of the arcade is 5109 cm-3 assuming its depth equal to the width. However, high brightness of the arcade center can be also due to the integration of column emission measure along the line of sight. If so, we estimate density using a loop leg of 9109 cm-3. Optically thin free-free radio emission calculated from Yohkoh/SXT data fits microwave observations both visually (Fig. 7) and quantitatively (mainly, 17 << 5.7 < 0.025) to confirm the estimates of plasma parameters. Fig. 6 (bottom) shows the dark features moving downward with the background image being the minimum over difference images overlaid with contours of temperature. The temperature of the inflow regions seems to be slightly less than in their vicinity. SOHO/MDI, Yohkoh/SXT, and SOHO/EIT 284 Å data several days before the event show the large-scale magnetic configuration almost unchanged. It was mainly bipolar with the N-polarity at the eastern part and S-polarity at its western part, with a weaker opposite magnetic field in the SE part of the activity complex. With the latitudinal orientation of long coronal loops this corresponds to the configuration observed on the limb.The line-of-sight magnetic field (extrapolated from the SOHO/MDI magnetogram of October 19, 2001 into the corona in the potential approximation) and the degree of polarization (NoRH, 06:11–06:17) are shown in Fig. 3. The overall configuration corresponds to the microwave polarization, with predominant LCP and short-term appearance of RCP southward of the arcade. The magnetic field is 7G at 100 Mm. Discussion. The arcade environment is negatively polarized both in NoRH and RATAN data at 17 GHz, thus delineating the magnetic configuration (x-mode, Fig. 3). The short-term appearance of the positive polarization in NoRH images southward of the arcade shows the presence there of the large-scale N-polarity magnetic field. The absence of the polarization in the brightest part of the arcade is due to the proximity of the opposite magnetic fields. If the microwave emission mechanism would be optically thin bremsstrahlung, the polarization measured of (10–12)% resulted in incomprehensible magnetic fields of ~300 G at ~100 Mm. However, the total intensity spectrum of the microwave source measured with the RATAN-600 shows its nonthermal origin. Indeed, an addition of 40%-polarized gyrosynchrotron emission with TB~ 250 K can contribute the whole polarization observed. The extrapolation of the magnetic fields confirms our conclusions. The arcade exists more than 7 hours keeping its configuration without conspicuous changes. The existence of long-living giant hot structures high in the corona produces a fundamental problem. Two statements about the solar corona are generally accepted: a) The magnetic pressure well exceeds the plasma pressure,  << 1, and b) the magnetic field rapidly falls off with height. The two statements forbid long steady existence of hot structures high in the corona (cf. Getman & Livshits 2000, Astron. Rep., 44; Shibasaki 2001, ApJ, 557). The cooling time of the arcade is ~3 hours for both the radiative and conductive ways. McTiernan et al. (1993, ApJL, 416) and Livshits et al. (2002, Astron. Rep., 46(7)) proposed that the reason is a long-term heating. A possible candidate can be a slow reconnection process. It can also supply extra plasma density into the arcade as well as plasma flows that can affect the equilibrium conditions in the arcade and its environment. The dark features moving downward onto the arcade visualize the magnetic configuration above it. Fig. 6 (bottom) hints the inflow regions to be slightly cooler than their vicinity. Those features can be due to cool remnants of a filament pre-ejected (cf. Hyder 1967, Sol. Phys., 2 and Shibasaki 2002, ApJL, 567). This downflow can contribute to: 1) mass supply to the arcade, 2) equilibrium conditions to balance the gas pressure, 3) long-term heating. Conclusions.1. Plasma parameters in the arcade: Ne ~1010 cm-3, Te ~6 MK. 2. High-sensitivity NoRH and RATAN-600 polarization data reveal the large-scale magnetic configuration surrounding the arcade high in the corona, in accord with the expected one from the magnetograms few days before. The potential-approximation extrapolation of the photospheric magnetic fields into the corona confirms the configuration. 3. The analysis of the posteruptive arcade gives rise to a problem of the existence of long-living giant hot structures high in the corona: either the magnetic field surrounding the arcade well exceeds 20 G at that height (the magnetic pressure outside the arcade is equal to the gas pressure inside it), or  > 1 both inside and outside the hot structure high in the corona. 4. Dark features moving downward onto the arcade can be important for this problem. They likely show the downflow due to the falling down remnants of the filament pre-ejected. This downflow can contribute to the mass supply into the arcade and to the equilibrium conditions.