Following several years of design studies of UV imaging interferometers for Solar Physics Space Missions (SIMURIS/SUN-MUST-SOLARNET), we decided, in 1995, for demonstration purpose, to realize a complete test of the cophasing feasibility and performance directly on the Sun. Accordingly, our laboratory breadboard of a 2-Telescope cophased interferometer (on which demonstration of the cophasing method were performed from 1992 to May 1995) was moved to the "Grand Sidérostat de Foucault" at Meudon Observatory. During summer 1995, and up to March 1997, the feasibility and performances of the cophasing of two telescopes on extended objects like the Sun, the Moon and planets (Mars, Saturn, Jupiter) were demonstrated. These results really open the possibility to use and discover from solar interferometers, not only in Space but also on ground. With a 1 meter baseline or so, a ground imaging interferometer (e.g. THEMIS with a phased pupil) will reach a PERMANENT spatial resolution of 0.1" on a coherent field-of-view of 40", allowing to untangle the confining and dissipating mechanisms and processes in magnetic flux tubes, prominences, flares, etc. This factor 10 in resolution is a factor 100 in flux concentration, i.e. a major gain for MAGNETIC FIELDS and POLARIZATION measurements too. We describe progress in designing Solar Interferometers in Space and on ground and a possible adaptation of the interferometry techniques to THEMIS.
Since the Solar Ultraviolet Network (SUN) proposed to ESA
in 1989 (Damé et al., 1989, Coradini et
al., 1991), we have developed an instantaneous direct imaging
interferometry concept based on a compact interferometer
configuration (several telescopes, 3 to 5, on a circular baseline,
but large enough compared to the baseline so that the spatial
frequencies plane is filled). It was proposed to ESA as a satellite
mission in 93 (Multi-mirror Ultraviolet Solar Telescope,
MUST/SIMURIS proposal, Damé et al., 1993a) and,
more recently, April 15, 1997, as an Externally Mounted Payload on
the International Space Station (Solar and Planetary High Resolution
UV Imaging by Interferometry, SOLARNET, Damé et al.,
1997). In support to ESA studies, CNES engaged in 92 Research and
Technologies (R & T) funds for the realization of a 2-Telescope
Breadboard to demonstrate the heart of the system: the measurement of
an absolute phase and the cophasing control of the
interferometer.
The breadboard was completed in spring 1994 and by September 1994
a complete laboratory demonstration of the cophasing of
two-telescopes on extended objects was achieved with remarkable
performances. During summer 1995 the breadboard was installed at
Meudon Observatory at the "Grand Sidérostat de Foucault" and
the first direct cophasing on the Sun was made with a phase control
of l/140. These cophasing experiments were
repeated in 1996 and better performances were achieved (l/240). Following these successes we can now
sustain that we have the recipe and the cooking skills for the
realization of a major instrument for the advance of Solar Physics at
the beginning of next century: the Solar Interferometer, either on
ground (MACS: Multi-Aperture Cophased System) or in Space
(SOLARNET).
In the following, we briefly recall the objectives and concepts of a ground Solar Interferometer, explain the constraints imposed by the measurement of a phase over extended objects, present the laboratory and sky results obtained with the first solar interferometric experiment of cophasing on extended objects (Sun and Planets), and indicate the next step of the demonstration programme on the breadboard: a 3-Telescope cophased (active cophasing) and pointed (active pointing) Imaging Interferometer for the Sun and planets. We conclude on the possible adaptation of these techniques - MACS Techniques - to THEMIS.
The relevant minimum observable scale in the solar atmosphere may
be of the order of 10-30 km since smaller scales will probably be
smeared out by plasma micro-instabilities (such as drift waves). This
scale range is comparable to - though slightly smaller - the photon
mean free path in the chromosphere (~ 60-70 km). Slightly larger
scales can be expected in the corona (though gradient across coronal
loops may also be a few km). Altogether this situation is rather
fortunate because we have access to higher resolutions in the far UV
than in the visible and X-rays (multilayer telescopes are limited to
resolutions of 1 arcsec or so). In the UV, the emission lines are
generally thin, i.e. not affected by the optically thick transfer
conditions which prevail in the visible and near UV lines accessible
from ground, and we can expect to see structures with scales 10 to 30
km. In the visible, thick transfer in the atmosphere blurs the
signature of structures and nothing smaller than 70-100 km should be
observed. This is, indeed, well delineating the maximum useful
scientific spatial resolution. This means that with a single
instrument of meter class diameter we have the appropriate,
scientifically justified, spatial resolution for both the UV
(20 km at Lyman Alpha 121.6 nm) and the visible (60 km in the Ca II K
line 396.3 nm). More details on the scientific objectives in the UV
and FUV can be found Damé et al. (1993b). For the
visible, most of the THEMIS literature could be cited since the 0.1"
is the observing objective of all programs.
The breakthrough in high spatial resolution observations (10 times more resolution, 100 times more flux in fine structures if they are of 0.1" in size) will allow to study the coupling between turbulent convective eddies and magnetic fields in the photosphere, fibrils and their chromospheric connections, network bright points and their coupling to waves and magnetic fields, etc. Also of prime interest are the plasma heating processes and thermal inputs of flares and microflares and their fine magnetic field structures.
If the need for high spatial and spectral resolutions is commonly
agreed, the question left is why an interferometer and not a
single-dish large telescope? The answer is that the required
measurement needs exceed conventional instrumentation limitations. A
1 m telescope diffraction-limited is difficult to construct even in
the visible (THEMIS does not escape this situation). And, even
assuming that such a perfect 1 m telescope could be built, it would
be extremely costly to control its stability because of the primary
figure evolution but also of the large secondary and critical
primary-secondary alignments and distance control.
The Michelson interferometric approach represents significant
advantages over direct diffraction-limited large telescope imaging.
Only small telescopes (or subpupils of a main pupil) are necessary
and active pointing of the telescopes can be made directly on the
secondary of each of the small telescopes (or on a small active
mirror afterwards). Small telescopes or segments of a large telescope
like THEMIS can possess a perfect - or near perfect - figure and
reach their diffraction limit. Interferometry requires to control the
residual optical path delays between telescopes but this,
consequently, guarantees a perfect output wavefront suitable for
diffraction-limited imaging.
Adaptive optics is not an alternative to obtain the correct figure
precision of large mirrors or to control the resulting errors,
because of thermal cycling but also (on ground) of the mechanical
deformations due to their weight. Note that aligning a segmented
mirror requires 6 degrees of freedom and a control of the distance
between the primary and secondary mirrors. This very complex control
loop is not required with an interferometer made of small telescopes
or when working on a limited set of pupil's segments (beside fine
pointing needs, only one degree of freedom is required: the phase
control).
Altogether, the modest baseline required to obtain major scientific results and the simplified control of an imaging interferometer (which doesn't need an absolute metrology like astrometric programs) result in a very reasonable optical complexity and cost which open solar interferometry programs to the medium size satellites programs, to the accommodation possibilities of the International Space Station, but also to large ground telescopes like THEMIS.
To study the ultimate fine structure of the Sun, a solar
interferometer needs to image an extended field-of-view (FOV) covered
with complex structures. And, since many structures of interest are
evolving rapidly (in a few seconds or even less), this imaging cannot
be achieved by classical long-baseline interferometry techniques
where fringes' visibilities are measured sequentially.
These constraints (FOV and time resolution) prompt to design an
interferometer with instantaneous imaging capability i.e., first, to
the choice of a compact array. By compact is meant that the spatial
frequency coverage of the array is comparable to a single dish
telescope in one fundamental aspect: complete coverage of spatial
frequencies, i.e. there are no zeroes in the modulation transfer
function of the array (cf. Fig. 1). Image restoration is, in this
case, based on a direct deconvolution. A central issue for
interferometric imaging is, therefore, a proper (i.e. compact)
configuration of the array.
The other important requirement is to control the residual optical
path delays between the different telescopes to a fraction of a
wavelength, i.e. to cophase the interferometer. This allows
all the recorded fringes to be used instantaneously, since not
affected by a significant phase problem (thus allowing a robust image
reconstruction approach). The consequence of importance brought by
this cophased approach is that, permanently, we have the insurance of
a near perfect wavefront (stable transfer function), the telescopes
of the array being controlled to their optimum phase position.
Fig. 1 - SOLARNET configuration (3 x Ø35 cm telescopes on a 95 cm baseline) and MTF.
This complete spatial frequencies coverage is the basic difference
between classical two-telescope interferometers and compact
multi-telescopes interferometers. In the two-telescopes case, the
fringes do contain high resolution information even though most of
the spatial frequencies are lacking between the low frequencies - due
to the area of the small telescopes' primaries - and a high
resolution peak due to the pair of telescopes. The data analysis then
relies, first, on a measurement of the fringes' visibilities at these
specific spatial frequencies, second, on a sampling of the other
spatial frequencies (e.g. by moving the telescopes) and, third, on an
image reconstruction from these data.
In a compact multi-telescopes case, all the spatial frequencies
are included in the image formed, but their relative weighting is not
as smooth as in the single dish case (cf. Fig. 1). However, the
absence of zeroes enables stable image restoration with direct
deconvolution algorithms which, in practice, simply re-weight
properly the spatial frequencies.
In this approach, images (i.e. extended FOV's) are recorded and restored without the intermediate steps of fringes' visibility measurement, u,v plane sampling and painful image reconstruction, necessary when diluted interferometric arrays are considered. Further, since the number of phase relations is very large, the FOV is important (³ 40 arcsec) while a diluted array of a few telescopes would have a FOV of 5 x 5 pixels or so... Then, a multi-telescopes compact interferometer has equivalent imaging capabilities than a classical telescope.
The major conceptual choice of our interferometric approach is to
cophase the array. By cophase we mean real-time control of the phase
differences between the telescopes (or pupil's segments), i.e.
constant monitoring of the equality of the optical path-lengths
traveled by the different beams. By this mean, the transfer function,
although still rather poor (cf. Fig. 1), is stable. Cophasing has a
sound justification since only cophased arrays can integrate light,
i.e. benefit from long exposures and, thus, from better signal to
noise ratios. The result is a significant increase in the complexity
and dynamic range with which images can be reconstructed. Numerous
simulations have been performed (Damé, 1994, Damé and
Martic, 1992) to demonstrate that such cophased arrays can properly
observe complex and extended objects. It was shown that, at a given
wavelength, l/10 of phase stability
guarantees that near perfect image reconstruction can be achieved.
This capability requires a specific cophasing control - using
reference interferometers - which we have been studying for several
years (see, e.g., Damé, 1992, 1993). The measurement of an
absolute phase on an extended object is not straightforward and we
will first recall some basic notions of coherence.
5.1 Reminders
"Cophasing" two or more telescopes supposes a minimum of knowledge
of our method of phase control. In order to be as much as possible
understandable in what we are doing and the tests that were carried
we present some reminders, first on coherence and, then, on the
Double Synchronous Detection Method.
When the light of a source is not monochromatic, from the knowledge its spectrum we can calculate the mutual coherence function G:
where n is the frequency and t the retardation time. The mutual coherence function has the dimensions of intensity. From it can be defined the dimensionless degree of coherence g:
It is the direct observable quantity of optics, being measured by
simple interference, e.g. in the Mach-Zehnder set up which we used
for our demonstrations of cophasing. We have both the addition
interferogram and the subtracted interferogram, with a phase
difference of ¹ between the signals; the interferogram can be
subtracted from the background, and therefore we can separate
interferogram and background in low visibility cases.
When the source is extended and when the light is not monochromatic, temporal effects are combined with spatial effects and what is to be considered is the result integrated over an extended source z of spectral irradiance E(x,n). The coherence function in this more general case is given by (for an explanation of the notations see Fig. 3):
[1]
Fig. 2 - Visibility
as a function of the optical path delay (in µm) for an ideal
white light source (point source with the spectrum shown left). Note
that we have "only" 5 fringes on ± 2 µm
Fig. 3 - Notation for the mutual coherence function.
With the usual approximation for path difference (i.e. that
variations in path differences are caused only by either u or
x changing from zero, which implies that the imaging system is
perfect, converting a spherical wave at the entrance pupil to a
spherical wave of different curvature at its exit pupil), this
becomes a three-dimensional Fourier transform. In other words, the
coherence volume which has for base the coherence area and height
ct, is connected by an uncertainty
relation with the volume in reciprocal space, defined by the product
of the solid angle subtended by the source and its bandwidth.
We have formally evaluated the integral [1] for a white light source (whose spectrum is shown on Fig. 2: ~ 300 nm FWHM centered at 850 nm) of different source sizes (diameter) for a reference interferometer which interbaseline (distance between two telecopes center to center) is 1.5 time the telescope diameter:
a) S = 0.3" c) S = 0.6" e) S = 1"
b) S = 0.5" d) S = 0.8"
This is assuming that the source is uniform (and, in this case,
that the telescope has a 20 cm diameter; results are indeed the same
for smaller sources and larger telescopes - inverselly proportional -
as long as the relation interbaseline to diameter is the same, i.e. a
1.5 time relation).
Fig. 4 - Fringes visibility as a function of the Optical Path Delay for a circular configuration in which the interbaselines (distance center to center of the telescopes) are 1.5 time the telescope diameter. Illustration for sources of diameter 0.3", 0.5", 0.6", 0.8" and 1". Curves go from light gray, 0.3", to black, 1" and telescopes of 20 cm (same for sources half size and telescopes 40 cm for example). Interesting to note is that the contrast inversion of the central fringe happens near the interbaseline resolution (as one could expect since the source is resolved).
However, in more realistic cases, the extended source can be arbitrary and, in particular, it can be a double source (e.g. a double star). We have illustrated this on Fig. 5.
Fig. 5 - Explanation
of the uniform and complex source definitions. The source (left) is
extended and uniform. The source (right) is also extended but complex
(e.g. a double star with one star brighter than the other).
What can be seen from calculations (cf. Fig. 6) is that for very
small uniform but extended sources, e.g. 0.12 arcsec (case A),
the contrast of the central fringe is, indeed, slightly reduced (>
80%) but isn't affected by any phase shift. Larger sources (e.g.
0.36") reduce even more the contrast (~ 40%) but yet without phase
shift. By controlling the position of the central fringe, the
cophasing is ensured. Though, if the source is highly structured
(double source case), an important shift of the central fringe may
happen on "large" reference sources (case B). Problem with
sources 0.36" with 30 cm interbaseline will occur at 0.18" if the
interbaseline is 60 cm...
In the case of "real" interferometers (looking for "real reference sources" for cophasing) there are, therefore, less problems to be anticipated with small (and equal) interbaselines even if the reference sources are extended or complex. Because our configurations are compact, we are always in the most favorable case where the interbaselines are the same and small (only 1.5 x the telescope's diameter). For THEMIS (cf. #8, Fig. 15) the situation would be even more favorable since the interbaseline distances are the same as the width of the pupil's segments.
Fig. 6 - "Plain" and
"Complex" sources (as defined in Fig. 5) visibility functions for
different source extensions from 0.12 to 0.36" (again for a 1.5 ratio
interbaseline to diameter of the telescope and a nominal 20 cm
telescope). Note that problems occur only when the source is highly
complex (such high contrast on structures 0.03" in size are not
expected on the Sun, at least in the visible and on a large spectrum)
or when the reference field of view transmitted to the cophasing is
somehow larger than half the Airy disk of a telescope (possible error
larger than l/100).
Service d'Aéronomie has a long-standing expertise in
white-fringe acquisition and stabilization (cophasing). The method,
based on Zero-OPD detection by servo-control on the central
white-fringe through synchronous demodulation, has been studied in
details and its performances assessed both in laboratory experiments
(Damé, 1994, Damé et al., 1994, 1995b), and on
the sky, on stars, e.g. on Altair with l/225 phase stability, and on the Sun, with l/140 measured phase stability (Damé et
al., 1995a, 1995b). The method was further refined in 96 during a new
observational campaign at Meudon Observatory where the 2-Telescope
breadboard was cophased on a limited selection of stars, Jupiter,
Mars, the Moon and on different solar structures. It is worth
recalling that the method is an absolute OPD method which
provides calibrated fringes amplitude by the use - while the
fringe position is servo-controlled by the error signal of a first
synchronous demodulation - of a second synchronous demodulation at
twice the reference frequency.
Fig. 7 - Schematic
of the laboratory and sky setup used to validate the fringes
acquisition and stabilization on the central fringe.
In the Scientific Interferometer (cf. Fig. 7), the delay line "B"
is moved up to the region where a signal is detected by a first
synchronous detection (a modulation is introduced in the Reference
Interferometer by another delay line: "C"). Then, a second
synchronous detection, at twice the frequency, measures the amplitude
of the signal while the first synchronous detection is active. By
this mean, and since the first synchonous detection is active (we are
stabilized on the submit of the fringe then), the amplitude of the
fringe is directly measured.
Fig. 8 explains the principle and Fig. 9 illustrates a practical
acquisition and cophasing control on the central fringe made in May
1996 (laboratory measurement). On Fig 9, which displays the
synchronous detection signal (i.e. the fringe signal multiplied by a
sinus), one can clearly see that the observed signal is very near the
theoretical one (cf. Fig. 8). Crosses indicate the measurements made
by the synchronous detection at double the frequency while the
servo-control is active stabilizing the optical path between the
telescopes using whatever fringe providing a signal (even less than
1% contrast: see for example Table 1, Damé, 1996). Jumping
from fringe to fringe we are, then, able to determine the largest
one, i.e. the central one.
Fig. 8 - Fringes,
Synchronous Detection and Synchronous Detection at twice the
reference frequency.
Fig. 9 - Synchronous
Detection signature of the fringes as observed during an
acquisition.
Fig. 10 - Current
Delay Line implementation on the 2-Telescope Solar Interferometer
Breadboard at the Observatory of Meudon (March 1997). The Delay Line
consists in a retroreflector (3 assembled mirrors), a step motor
(Micro-Contrôle, course 20 mm, elementary steps of 0.1 µm)
and a piezoelectric ceramic (± 7 µm).
Table 1. Comparison of theoretical and observed contrasts (extremum - positive or negative - not necessarily corresponding to the central fringe value), and the corresponding cophasing stabilities that were achieved.
|
(as a fraction of telescope Airy disk) |
|
|
|
|
(lref. = 550 nm) |
|
1/3 |
5 |
70.4 |
74.6 |
3 |
|
|
1/2 |
10 |
-19.5 |
-18.9 |
5 |
|
|
1 |
15 |
6.7 |
5.2 |
7 |
|
|
1 1/4 |
20 |
2.7 |
3.5 |
- |
|
|
1 1/2 |
25 |
-3.1 |
-3.6 |
- |
|
|
2 |
30 |
3.1 |
3.2 |
7 |
|
|
3 |
50 |
-1.3 |
-0.8 |
9 |
|
The acquisition process is such that we first pass over the fringes when we have detected them and record the intensities measured (see the "Journal: Etude courante" on Fig. 9 which reflects that). Then we back up to the first detectable fringe and put the servo on (i.e. the piezoelectric ceramic on the delay line retroreflector is fed - cf. illustration, Fig. 7 - by the synchronous detection error signal). At this time we are on the submit of a fringe and we measure, via the second synchronous detection, the amplitude of that fringe. When this is done, we cut the servo and "jump" to the next fringe using the first stage of the delay line (the step motor). Because of the very large spectral bandpass that we use (300 nm), we always have more or less 5 fringes on ± 2 µm, fringes at 0.8 µm one from each other, i.e. 4 steps (at 2 x 0.1 µm) of the delay line. When the "jump" is made, the servo is put back on which leads us to the submit of the next fringe. And, when we are there, and stabilized, we measure the amplitude by the synchronous detection at double the frequency. These "jumps" are reflected in the "Journal: Etude courante" again and, when the intensity lowers rather than improving, then we back one step and are on the central fringe. A comparison with the side values and with the values of the initial scan confirms or not that the process has been successful. This is simple and efficient. With a limited number of fringes (which is a definite flux advantage since the spectral bandpass is larger) the identification process, even when contrasts are as low as 1%, has never failed. Only when numerous fringes are considered is the identification more difficult since fringes have almost the same amplitude. As reported in Table 1, very high stabilities were achieved by this method (³ l/100) and, for extended, Sun-like objects, with better results when using small flux/high contrast (small holes) rather than larger flux/low contrast (large holes ³ to the interbaseline resolution).
Important to this process of acquisition and stabilization is, indeed, the control software and the acquisition card beside the "pure" hardware: diode amplifiers, synchronous detections, frequency generator, piezoelectrics, step motors, etc. Since 1994, we developed and use a dedicated C++ program (20 000 lines of code) to perform, under Windows 3.11 (using the multimedia - video - clock) this real-time process of acquisition and stabilization at the 300 Hz rate of the modulation (sampling at 3 KHz). With the need to control not only one but several interferometers simultaneously we have developed a new software (still in C++ and Visual Basic) but this time under a "real" multitasking and multithreading preemptive system, Windows NT 4.0. This has a definite advantage in terms of response and control of the tasks since the control to the instruments (the GPIB/IEEE link) is faster with a very limited risk of collisions and since, moreover, all the internal tasks of acquisition or fringe jumps can be programmed as independent "threads" in the program. Beside the evident gain in terms of multi-telescopes / interferometers control and real-time capacities, better stability performance is also expected since, in addition, a new and faster (500 Ksamples/s) acquisition card has been provisioned allowing gain adjustments on the different entry channels (overcoming the noise/dynamic problem encountered with the present card). This new acquisition software (and hardware) will be used by mid-May 1997 on the 3-Telescope Imaging Breadboard. With this system we will truly enter the world of controlled interferometric imaging.
Although the laboratory results obtained in 1994 (cf. Table 1)
were excellent and hardly contestable, they were still doubts that
the laboratory conditions could reproduce without bias the exact
solar conditions. Since the cophasing is performed in the visible,
either in the Space Instruments or in the ground programme, we
therefore moved the whole experiment to the "Grand Sidérostat
de Foucault" at Meudon Observatory from May to July 1995 for a
complete demonstration of the cophasing on the Sun and obtained the
first solar directly cophased interferometric fringes with
independent telescopes (cf. Damé, 1995b and 1996). Our major
problem was the lack of pointing control (no reference of the
refractors' position) since, despite the double laser metrology and
the autocollimating lens (cf. Fig. 7), only alignments from the field
stops to the interferometers could be mastered (but well mastered:
the two lasers being aligned, the two interferometers are aligned up
to the selection holes right at the focus of the refractors).
However, if an objective moves, the solar image moves and the fields
do not overlap anymore (selecting holes are a few µm) resulting
in no interferences. Because of that situation, pre-alignment was
made, first, with a collimating mirror in front of the objectives
and, second, by the use of a perfect point source: a star. We
observed at night Arcturus and Altair (for this we controlled the
speed of the siderostat motor by a frequency variator) and during the
night of July 6, we aligned and then cophased our interferometer on
Altair with a measured stability of l/225
but a limited contrast (40%). At 7:00 AM the 7th, we cophased the
interferometer on the Sun using a 10 µm hole. We measured a
fairly low contrast of 4 % but nevertheless achieved a stability of
l/140 (at lref.
= 550 nm). With an improved system (Figs. 11-13 show the setup used
at Meudon up to March 1997 using a 1.2 x 1.8 m bench), the
observations (on the Sun but also on Jupiter, Mars and the Moon) were
also carried in 1996/1997 (we largely improved the stabilities and
contrasts - over 50% gain - though not yet to the laboratory values).
The initial alignment was still obtained with a collimating mirror
but the constraining alignment on a star was eased by a pre-alignment
on a pseudo point-source, a laser diode installed 200 m away on the
top of the Meudon Solar Tower. With this system, the fine adjustment
of the reference aperture stops using the star (still necessary
because of the laser diode finite size) is made in a matter of
minutes rather than hours... Beside stabilities and contrast
measurements, these observations allowed to test the dual filtering:
large field, 30" or more, at telescope level for the scientific
field-of-view, and reduced field-of-view, a fraction of the
resolution, in the reference interferometer where the phase
(cophasing control) is measured.
At this point, the major demonstration is made: we observed and cophased fringes on the Sun and we are convinced that our major difficulties are linked to the lack of fine pointing (seeing is bad at Meudon, small fields are not always superimposed, and the small size of the refractor is sensitive to scintillation when observing stars). Accordingly, by September 1997, the system will be upgraded to three telescopes, and fine pointing (active mirrors) will be implemented
Fig. 14 shows the optical layout of the 3-Telescope Imaging
Breadboard which is currently integrated at Verrières on a
large optical bench (1.5 x 3 m) before being moved to Meudon for
further observations. This new breadboard will also use the small 60
mm refractors (cf. Fig. 13) to simulate the entrance telescopes and,
alike the experiment carried in laboratory in 1995, a collimator (8
inches Celestron), will be used to illuminate the objectives during
the integration and testing phase. New hardware and software will be
implemented for the 3-Telescope Imaging Breadboard financed on one
side by CNES R & T funds and, on the other, by ESA Contract
Optical Aperture Synthesis Technologies 2 (OAST2). In particular, a
second delay line and 3 active mirrors have been provisioned that can
be controlled up to 100 Hz.
Fig. 11 - The
2-Telescope Breadboard of the Solar Interferometer at the "Grand
Sidérostat de Foucault" at Meudon Observatory in March
1997.
Fig. 12 - Top and
side views of the 2-Telescope Breadboard at Meudon Observatory in
March 1997.
Fig. 13 - The two "small" objectives (60 mm diameter) which simulate the entrance telescopes of the 2-Telescope Breadboard.
With the 3-Telescope Breadboard we will apprehend the system
aspects of multiple telescopes control and imaging. Control can
either be achieved by a cascade of servo-control with strictly
similar reference interferometers (the method that we favor) or with
some more complexity by a triple modulation and closure phase
technique when associating 3 telescopes together. The 3-Telescope
Breadboard will allow to investigate the performances of the two
methods. However, we lack of an appropriate CCD camera to fully
qualify the different imaging approaches (phase closure for low flux
or high rate acquisitions, etc.). It would be a plus for the
performances evaluation to acquire a reliable CCD system for the
ground system. This has been asked to INSU in France.
Fig. 14 - Schematic layout design of the 3-Telescopes breadboard as it is currently implemented on our new 1.5 x 3 m optical table. Note that by applying the corrective error signal to Delay Line "A" rather than "B", one can switch from the "cascading" mode of interferometers control ("B" servo-controlled on "A" and "C" on "B") to the "simultaneous" mode ("A" and "C" servo-controlled on "B"). Note also the 3 Active Mirrors and the new imaging channel where the 3 beams are recombined.
Fig. 15 -
Application to THEMIS of the Cophased Interferometry approach of
SOLARNET. Because the THEMIS pupil is already a full pupil (no need
to add telescopes) the proper choice is to use the full pupil and
consider individual segments as the entrance subpupils to be
cophased. Optimum THEMIS 90 cm telescope are 5 or 7 segments in order
to account for an r0 of 40 to 50 cm or so. Note that direct cophasing
interferometry is much more efficient that adaptive optics since the
method has been proved to work on extended objects and also since
only 6 phase measurements (6 reference interferometers) and, then, 6
servo-controlled delay lines are required to phase the whole
pupil.
The application to ground-based solar astrophysics is direct. The
cophasing method is working in the visible on any diaphragmed field
on the Sun. It is, then, directly applicable to THEMIS. With access
to the THEMIS pupil at the level of the present Active Mirror, we can
envisage replacing this simple system by a 7 pupil's segments
cophased pupil (7 reference interferometers) helped by 7 active
mirrors (one per subpupil). Our current control application
(COPHASE), based on Windows NT 4.0, is already dimensioned to control
7 reference interferometers as well as the pointing of 7 active
mirrors.
The prospects are therefore excellent to achieve, permanently, 0.1 arcsec spatial resolution without back effects (since the pupil is full all observing modes, dispersed spectra or spectro-imaging, are allowed) with THEMIS in the near future. It is our intention to propose an interferometric mode for Permanent High Resolution Observations for the second generation of THEMIS instrumentation.
We have developed a complete design for a solar interferometer
suitable to represent a major breakthrough in the Solar Physics
findings of the next century. We proved that the major assumption of
the overall concept, the cophasing of the array, is feasible and,
moreover, that performances to expect are very high. We are working
on a 3-Telescope Breadboard, cophased and pointed, which will, by
September 1997, demonstrate the system aspects and show the first
interferometric images of the Sun made with independent telescopes.
With continuous observations of the Sun with 0.1" spatial resolution
on a 40" FOV, a breakthrough will be achieved in Solar Physics on
ground if this system is implemented in THEMIS (instruments of second
generation). More details on the instruments (on ground and in
Space), cophasing techniques, image reconstruction algorithms and
performances, double monochromator design (focal instrument of the
Space's programmes), and our laboratory and "sky" results are
available on our web server: http://must.aerov.jussieu.fr.
Acknowledgments. We are grateful to MATRA MARCONI SPACE for financial support in 94 and 95 that allowed to complete the 2-Telescope Breadboard and to carry the solar tests at Meudon Observatory. This work is supported by CNES R & T Grants since 92. ESA Contract OAST2 is supporting most of the evolution of the Breadboard from 2- to 3-Telescope.
Coradini, M., Damé, L. et al.: 1991, "Solar, Solar System and Stellar Interferometric Mission for Ultrahigh Resolution Imaging and Spectroscopy (SIMURIS)", Scientific and Technical Study - Phase I, ESA Report SCI(91)7
Damé, L.: 1997, "A Revolutionary Concept for High Resolution Solar Physics: Solar Interferometry", ESA Symposium on Scientific Satellites Achievements and Prospects in Europe, Paris, France, 20-22 November 1996, (Proc. to be published by the AAAF)
Damé, L. et al.: 1997, "Solar and Planetary High Resolution UV Imaging by Interferometry, SOLARNET", Proposal submitted to ESA Opportunity Call SP-1201, December 1996, for "Externally Mounted Payloads during the Early Space Station Utilisation Period"
Damé, L.: 1996, "The MUST/3T Solar Interferometer: An Interferometric Technologies TestBed on the International Space Station", ESA Symp. on Space Station Utilization, Darmstadt, 30 Sept.-2 Oct. 1996, Ed. T.D. Guyenne, ESA SP-385, 369
Damé, L., Derrien, M., Kozlowski, M., Antonucci, E., Ragazzoni, R. and Tondello, G.: 1996, "SIMURIS: a UV and XUV Mission for High Resolution Solar Physics", Advances in Space Research, 17(4/5), 377-380
Damé, L., Derrien, M., Kozlowski, M., Ruilier, C., Appourchaux, T. and David, F.: 1995a (September), "Observation Directe et Asservissement des Franges avec un Interféromètre Solaire à Deux Télescopes au Grand Sidérostat de Foucault à l'Observatoire de Meudon", Rapport Interne du Service d'Aéronomie, v1.0
Damé, L., Derrien, M., Kozlowski, M. and Ruilier, C.: 1995b, "Instrumental Prospects in Solar Interferometric Imaging", JOSO 27th Annual Meeting, Benesov, Tcheque Republic, 12-15 November 1995, JOSO Annual Report 1995, 52-59
Damé, L.: 1994, "Solar Interferometry: Space and Ground Prospects", in Amplitude and Intensity Spatial Interferometry II, Ed. J.B. Breckinridge, Proc. SPIE-2200, 35-50
Damé, L., Derrien, M. et Kozlowski, M. : 1994 (September), "Démonstration de Faisabilité de l'Interférométrie Spatiale Solaire par Asservissement Synchrone sur les Franges en Lumière Blanche", Rapport Interne du Service d'Aéronomie, v1.0
Damé, L.: 1993: "Actively Cophased Interferometry with
SUN/SIMURIS", in Spaceborne Interferometry, Ed. R.D.
Reasenberg, Orlando, Proc. SPIE-1947, 161
Damé, L. et al.: 1993a, "A Solar Interferometric Mission for Ultrahigh Resolution Imaging and Spectroscopy (SIMURIS)", Proposal to ESA Call for the "Next Medium Size Mission - M3"
Damé, L., Marti\o(c;´), M. and Rutten, R.J.: 1993b, "Prospects for Very-High-Resolution Solar Physics with the SIMURIS Interferometric Mission", in Scientific Requirements for Future Solar Physics Space Missions, Oslo, 12-13 January 1993, Ed. B. Battrick, ESA SP-1157, 119-144
Damé, L., 1992: "Demonstration and Performances of
Real-Time Fringe Tracking: a Step Towards Cophased Interferometers",
in Solar Physics and Astrophysics at Interferometric
Resolution, Eds. L. Damé and T.D. Guyenne, ESA
SP-344, 277
Damé, L. and Marti\o(c;´), M.: 1992, "Study of an Optimized Configuration for Interferometric Imaging of Complex and Extended Solar Structures", in Targets for Space Based Interferometry, Ed. C. Mattok, Baulieu, ESA SP-354, 201
Damé, L. and Rutten, R.J.: 1992, "Prospects with SIMURIS",
in Solar Physics and Astrophysics at Interferometric
Resolution, Eds. L. Damé and T.D. Guyenne, ESA
SP-344, 21
Damé, L., et al.: 1989, "A Solar Interferometric
Mission for Ultrahigh Resolution Imaging and Spectroscopy (SIMURIS)",
Proposal to ESA for the "Next Medium Size Mission - M2"