Figure 1: The scheme of Dopplerometer configuration adopted in the experiment (left). Bi-sector plane of the crystal pair #1 and #4 (dotted line) points momentarily towards the source (flare). Crystals are fixed mechanically at the exact angle of 2
. By rotating fixed crystal section back and forward, the Bragg-reflected spectra in the vicinity of strong lines are measured by a pair of double proportional counters. In the ideal case of perfect co-alignment and for the source plasma being at rest, the line centres are recorded in channels #1 and #4 at the very same time during the scans performed in the opposite wavelength sense (upper right panels). Any radial motions of the hot flaring plasma will displace (in time) the relative position of lines (lower panel). The displacement is proportional to the line Doppler shift and is independent of the position of the source (flare) on the disc and/or the relative pointing, provided they are
stable.
Figure 2: An example of measured spectra recorded nearly simultaneously in channels #1 and #4 during the maximum phase of the X5.3 flare. Indicated are the times, when the centre of the strongest resonance line is
measured.
Based on design experience gained from early experiments, we developed and flight tested novel type of spectrometers and deterrs (Sylwester, 2001). One of them is so-called X-ray Dopplerometer. A scheme of the Dopplerometer section (left panel) and the corresponding expected spectral recordings (right panel) are shown in Figure 1. The section consists of the two crystals fixed mechanically on the common shaft at the exact (to within few arcsec) angle 2
where
is the corrected Bragg angle corresponding to the rest wavelength of a strong emission line.
According to Bragg condition:
(Q = Bragg angle, k diffraction order, d = crystal lattice spacing).
By rotating the crystal section, the X-ray spectra are recorded using a pair of appropriate proportional counters. The scanning takes place in the opposite directions for each of the crystals in the couple. Such arrangement of the crystals allows for precise (difference) measurements of the possible Doppler shifts of the measured lines.
For the source plasma being at rest, the line maxima are to be recorded at the same wavelength (i.e. same instant during the scan). Any radial motions of hot flaring plasma will cause the maxima of the lines to be recorded slightly shifted in time. This time shift is proportional to the value of radial component of source plasma velocity. The time difference effect is
independent of the position of the source (flare) on the disc and/or the relative pointing of the spectrometer, provided the pointing does not change fast. The accuracy of the velocity determination is expected to be very good, down to few km/s as a result of adopted concept of difference measurement.
The scanning (uniformly changing the angle of incidence of X-rays) enables to measure spectral features in the chosen X-ray range using crystals with appropriate spacing. Two of the crystals used in Diogeness are identical Quartz
(
) with 2d =
6.6859 Å which enable us to measure the spectra in the vicinity of Ca
XIX w line.
Bragg condition for the chosen line is fulfilled approximately at the same time for both crystals in the section. Provided the lines are Doppler-shifted, the maxima of lines will not coincide in time and precise determination of plasma radial velocities can be made.
Diogeness has been placed aboard the CORONAS-F satellite launched 31 July 2001 from Plesetsk (Russia). During first month it collected more than 2000 spectra for a number of flares, including a strong X5.3 flare on 25 August 2001. For details concerning the Diogeness spectrometer construction see the paper by Siarkowski et al., (2002). In the present contribution, we concentrate on determination of the line Doppler shifts for this X5.3 flare.
Figure 3: Relative drift of the Diogeness reference plane (satellite pointing/flare transversal motions), as projected on the dispersion plane of the spectrograph. This residual motion has been detected with accuracy of few arcsec by the Dopplerometer. The drift shown does not influence the accuracy of wavelength scale determination.
On 25 Aug. 2001, during two satellite orbits 15 left and right scans have been recorded by the Dopplerometer for a very intense long duration flare. Most of the spectra recorded correspond to the maximum and decay phase of this flare when its intensity was above the M1 X-ray level. In Figure 2 an example of measured spectra recorded nearly simultaneously in Channels #1 and #4 (forming the Dopplerometer) during the maximum phase of X5.3 flare on 25 Aug. 2001 is presented. On the horizontal axis the motor step Nos. are plotted. The motor scanned (forward and backward) the X-ray spectra within a narrow angular range of ~ 140 arcmin including the Ca XIX triplet lines. The data gather interval (DGI) of 0.25 s has been used (corresponding to 21 arcsec change of crystal rotation). The most prominent spectral features seen on both spectra are the resonance w lines corresponding to the 1 s2 1S0 - 1 s2p 1P1 transition in the He-like Ca XIX ion. The scanning in both channels is executed in the opposite wavelength sense as seen by proportional detectors #1 and #4. Thus the intercombination (x, y) and forbidden z lines comprising the Ca XIX triplet are seen on the opposite sides of the presented motor step range, being recorded approximately 20 s apart in time for z lines. It is of primary importance to have the precise determination of wavelength scale in order to determine the small Doppler shifts. Determination of ''absolute'' experimental wavelength scale (i.e. assignment of wavelengths to time through appropriate calibration of the motor steps) has been achieved in the several steps. In the first step, the background and the continuum emission have been removed from all the spectra in the sequence, and the spectra have been normalized (respective to w line maximum) in order to correct for a slightly different spectrometer sensitivities in both Dopplerometer sections. In order to find the relative time offset Dt between the instants corresponding to line centres as recorded by the two crystals in a given scan, the line centre motor step positions have been determined. In this respect the fitting (c2 minimization) of the Gaussian profiles has been used. Appropriate uncertainties of the fits have been determined based on the count statistics. The time tc (located central between recording of the w resonance lines in both crystals sections) defines the orientation of the bi-sector plane of Dopplerometer (we assume the scanning being uniform, the source and the satellite pointing being stable within this short ~ 3 s time interval). In order to check for the pointing stability of the CORONAS-F spacecraft, we analysed the crystal addresses (motor steps) corresponding to these mid-times. In case there is no S/C pointing changes (or flare transversal motions), these addresses should keep constant. We observe however small changes of these crystal addresses during the period of ~ 2 hours when the X5.3 flare has been observed. In Figure 3, we plot this detected slow drift of the spacecraft pointing (or transverse source motion) with time. For clarity, we express the drift/motion in arcsec. Slightly different shades of the points plotted corresponds to forward/backward scans respectively. The observed amplitude of the drift is well within the limits of ~ 10 arcmin pointing accuracy of the z-axis of CORONAS-F and the maximum angular velocity of the drift ( ~ 0.0005 arcsec/s) is less than respective velocity allowed in the S/C technical specification. However, it comes from the prompt inspection of the Yohkoh SXT Be119 deter, that the flaring plasma changed its location slightly, so the interpretation of the drift observed is under study. The amplitude of the effect indicates that the assumption of negligible role of S/C slews (plasma transversal motions) for differential Doppler velocity determinations is justified. In the next step leading to assignment of wavelength scale, we based on laboratory calibrations of the crystals used in the spectrometer. In Table 1 we present results of laboratory measurements of the crystal properties important for the present study. In the next step, we checked for possible effects of thermal dilatation, by looking into readouts of the temperature sensors placed on the crystal support. No substantial (within less than 2oC) changes of the temperature have been found during two hours of measurements, but we found temperatures of each of crystals slightly different (by ~ 3oC). Appropriate corrections have been taken into account in the further analysis. The fact, that the temperature of the crystals do not change substantially is in agreement with the S/C orbital information indicating that CORONAS-F was on a constantly Sun-lit portion of its semi-sun-synchronous orbit during the measurements. In respect with, one can safely assume that the orbital plane was approximately perpendicular to the direction towards the Sun, so the projected radial motions of the satellite along this line are negligible.
Quartz crystal properties as
determined from laboratory measurements
a - from laboratory measurements (T = 22 oC) done using collimator with single crystal monochromator tuned to characteristic K-line of Cr.
b - extrapolated for the Bragg angle corresponding to the Ca XIX w resonance line from measurements done using double crystal monochromator tuned to K-line of Ca. The extrapolation uncertainties are included in the measurement error.
c - from laboratory measurements done using double crystal monochromator (1,-1), tuned to characteristic K-line of Ca.
Figure 4: Velocities determined from X-ray line shifts observed by Diogeness. In the upper panel determinations for Ca
"w" line from both sections of Dopplerometer are superimposed. In the lower two rows, plots of the velocities found from S
XV and Si XIII are displayed. A high blue-shift observed early in flare are of similar value in all three lines measured. The uncertainties shown represent combined relative error of wavelength scale determination. For Ca
"w" line the uncertainties around the flare maximum are few km/s only.
Figure 5: The comparison of normalized Ca
xIx spectra observed during the rise (left) and decay (right) phase of the flare by Diogeness and Yohkoh BCS. A substantial difference in the widths of w line is seen between the rise and decay phases. The
BCS spectra overplotted (black line) have been observed during collection of Diogeness spectral scan (rise phase) or 20 s afterwards (decay phase).
Diogeness was not a unique Bragg spectrometer observing the 25 August 2001 X-flare. By a fortunate coincidence the Bragg, bent Crystal Spectrometer (BCS, Culhane et al., 1991) aboard Yohkoh recorded X-ray spectra in the wavelength ranges matching two of these observed by our instrument. BCS records spectra instantly in the entire spectral band covered with its crystal being bent and the detector position sensitive. It is also much more sensitive than Diogeness, which actually is a disadvantage when observing strong flares, since it becomes saturated for events above M2 X-ray class. This leaves small room for cross-comparison of the two spectrometers for the period during the X-flare when its intensity was above Diogeness sensitivity threshold and below BCS saturation limit. Few spectra in common have been identified, and direct comparison of the spectral records is shown in Figure 5 for the initial rise and late decay phase.
The soft X-ray Dopplerometer concept has been successfully flight tested within the Diogeness spectrometer flown aboard CORONAS-F solar observatory. It has been proven that the use of this new concept of Bragg spectrometer construction allows for very precise determinations of X-ray line shifts due to line-of-sight velocity component. Further analysis of the recorded spectra collected by this spectrometer will allow for determination of the other physical characteristics (temperature, emission measure, elemental abundances) of the investigated X-class flare. The result presented indicate that the Dopplerometer crystal configuration would be very useful concept to be incorporated in the future Bragg spectroscopy missions.
This work has been supported by the Grant 2.P03D.002.22 of Polish Committee for Scientific Research.