Observing the Second Solar Spectrum at IRSOL

In 1960 the University Observatory, USW, in Göttingen, Germany, opened in Locarno its station for solar observations under the earlier name Istituto per Ricerche Solari (Brückner, Schröter, & Voigt 1967). As a consequence of the construction of the new German solar observatories on Tenerife in 1984, the stations on Capri and in Locarno were closed. A private association directed by Alessandro Rima could take over the institute in Locarno and managed to create in 1988 a foundation (FIRSOL) that governs the observatory. The name of the observatory was changed to Istituto Ricerche Solari Locarno (IRSOL). Since 2000 Philippe Jetzer is directing the foundation FIRSOL. Beginning in 1988, the instrumentation has been reconstructed and upgraded with new technologies with the help of USW Göttingen (Wiehr & Bianda 1994), FHS Wiesbaden (Küveler, Wiehr, & Bianda 1998, Küveler et al. 2003), and Institute of Astronomy at ETH Zurich. The close and fruitful collaboration with Jan Stenflo and his Institute allowed the development of a scientific program mainly related to polarimetry within the ZIMPOL project. A historical overview of old and recent results obtained at IRSOL in this field of research is presented here, as well as some information on the ZIMPOL polarimeter itself.

Under the direction of the University Observatory in Göttingen, Germany, several projects in polarimetry were performed in Locarno. The Gregory-Coudé type telescope is well suited for polarization measurements. It has a low amount of instrumental polarization that remains nearly constant over the day. Thus it can be easily compensated with optical devices at the exit of the telescope or during the data analysis.

In 1963 Günter Brückner published his work: Photoelektrische Polarisationsmessungen an Resonanzlinien im Sonnenspektrum (Brückner 1963), reporting observations of the scattering polarization in the Ca i 4227 Å line. In particular he was able to measure the polarization in the line wings. He used a stepwise rotated half wave plate and a calcite crystal as analyzer. The sensor was an IP21 photomultiplier. Jan Stenflo had the opportunity to observe at Istituto per Ricerche Solari already in August 1973 and could confirm and extend the results of Brückner (Stenflo 1974). Probably he did not realize at that time how familiar that place would become for him a few decades later. Further polarimetric investigations were performed by the University Observatory in Göttingen. Eberhard Wiehr (1975) measured scattering polarization in the lines Sr ii 4077 Å, Ba i 4554 Å, and Na i 5890 Å. He also measured the center to limb variation of the continuum in several wavelength windows. His polarimeter was based on a KDP modulator (795 Hz) with which seeing effects could be eliminated.

A summary of the scientific activity performed during the first two decades when the Institute was operated by USW Göttingen can be found in Wiehr, Wittmann, & Wöhl (1980).

The observations performed before 1984 confirmed the suitability of the Locarno telescope for polarimetric measurements. In 1995 together with Jan Stenflo and Sami Solanki it was decided to develop a two beams exchange polarimeter (Fig. 1) similar to the one used by Semel, Donati, & Rees (1993) for stellar observations. The main goal was to perform observations in a complementary way to the first ZIMPOL version, which was not able to explore the violet-blue part of the spectrum, in contrast to the two beams exchange polarimeter. The advantages of the IRSOL two beams exchange system are: its broad spectral performance (limited by the calcite transparency), its simplicity, the good polarimetric sensitivity, and its photon efficiency. There are nevertheless limitations with this technique: two successive recordings are required to measure one Stokes parameter, and this introduces some seeing dependence in the polarization measurements. Moreover, the Stokes polarization parameters cannot be measured simultaneously, and each requires two exposures. Thus six separate exposures are needed to measure the full Stokes vector.

The main result obtained at IRSOL using the two beams exchange polarimeter was the detection of the Hanle effect in the quiet chromosphere in the lines Ca i 4227 Å and Sr ii 4078 Å (Bianda et al. 1998a,b). Several conclusions were presented at SPW2 in Bangalore (Bianda, Stenflo, & Solanki 1999). This technique is now also used at IRSOL for stellar polarimetry of bright stars (Sennhauser 2007).

In 1998 the project to measure the Second Solar Spectrum Atlas with ZIMPOL II started at IRSOL. It was the thesis work of A. Gandorfer. After a careful optimization of the instrumentation for this ambitious project, systematic observations were carried out, which resulted in the first two atlas volumes (Gandorfer 2000, 2002). Since the start of this project, a ZIMPOL system has been permanently installed at IRSOL.

The success of IRSOL is intimately related to the Zurich IMaging POLarimeter, ZIMPOL. The main idea, on which this novel technology is based, was developed by Hanspeter Povel (Stenflo & Povel 1985; Povel, Keller, & Stenflo 1991; Povel 1995). He solved the problem how to combine polarization modulation in the kHz range (much faster than the typical seeing fluctuations) with large 2-D array detectors.

The heart of the ZIMPOL system is the camera with a specially masked CCD sensor, in which the fast demodulation is performed. The polarization state of the incoming light is changed into temporal polarization variations by a modulator. These variations are subsequently converted into intensity fluctuations by a linear polarizer. The masked CCD has alternating open and covered pixel rows. The charges are shifted forwards and backwards between the open and covered rows in synchrony with the modulator. In this way the CCD camera can handle several simultaneous image planes corresponding to different polarization states.

The first version, ZIMPOL I, was mainly used from 1994 to 1998 for observing campaigns at Kitt Peak, Sacramento Peak, and Tenerife. The particular CCD sensors, specially manufactured for the ZIMPOL I cameras, have every second pixel row masked. The camera is synchronized with a photo-elastic modulator, PEM. A ZIMPOL I camera can measure Stokes $I$ and one other Stokes parameter within a single frame. To obtain the complete Stokes vector, three different cameras or sequential observations with one camera are needed. Scientific results with the system installed at Kitt Peak (Arizona) and Sunspot, Sacramento Peak (New Mexico) have been reported in a number of papers, e.g. in Stenflo & Keller (1997), and Stenflo et al. (2000a,b).

A more evolved system, ZIMPOL II, has been available since 1998. With this system, all four Stokes parameters can be measured within one exposure. Three out of each four pixel rows of the CCD sensor are masked, allowing a single exposure to contain four images, which can be combined to reconstruct the complete polarization vector (Gandorfer & Povel 1997). This version is efficient above 450 nm and was used at IRSOL to record the first Second Solar Spectrum Atlas (Gandorfer 2000). The ZIMPOL II version was also used by the Institute of Astronomy in Zurich for international campaigns (Stenflo et al. 2002; Stenflo 2006).

The modulator most often used has been a PEM, with a modulation frequency of 42 kHz. To observe all four Stokes parameters with one exposure, two phase-locked PEMs would be required. So far, however, the synchronization of two PEMs within the required tolerance has not been possible. Thus PEM-modulation with ZIMPOL II has only been possible with one PEM, which allows to measure three Stokes parameters simultaneously. To record the complete Stokes vector, two independent measurements are then needed. The complete Stokes vector can however be recorded with one exposure with a non-resonant modulator based on two phase-locked ferro-electric liquid crystal retarders (FLC), with a typical modulation frequency of about 1 kHz. During the last couple of years ZIMPOL II has therefore increasingly been used with a dual FLC modulation system.

Until the year 2000 the CCD manufacturer was only able to apply a mask on frontside illuminated CCD sensors. Therefore ZIMPOL was not efficient below 450 nm. This wavelength limitation was overcome in 2001 by developing ZIMPOL II-UV cameras, based on CCD sensors with an open electrode structure (Gandorfer et al. 2004). At IRSOL this camera was used to record the data needed to complete volume two of the Second Solar Spectrum Atlas (Gandorfer 2002). The third volume of this atlas (Gandorfer 2005) is based on data recorded with ZIMPOL II-UV at McMath-Pierce Solar Telescope at Kitt Peak.

A drawback of the previous versions of ZIMPOL II has been the low photon efficiency of the masked frontside illuminated CCD. In 2006 a ZIMPOL II-UV CCD sensor could be equipped with a cylindrical micro-lens array on top of the CCD surface, in order to focus all the light on the open pixel rows. The photon efficiency of the ZIMPOL II-UV-micro-lenses camera has thereby improved by a factor of about three to six. A more detailed description of the current version of ZIMPOL II at IRSOL can be found in Kleint, Feller, & Bianda (2008).

Recently a third generation of ZIMPOL has been developed (Fig. 2). It is based on state of the art electronic components, and is a very flexible system that allows instrument configurations for a wide range of polarimetric applications in astronomy. ZIMPOL 3 has been developed at the Institute of Astronomy in Zurich. After the retirement of Jan Stenflo, the University of Applied Sciences of Southern Switzerland, SUPSI, in Lugano-Manno, has taken over the project. A new team is now responsible for the development and setup of ZIMPOL 3 in collaboration with IRSOL. The new system is expected to be available for science by the end of 2008.

Besides the already cited works performed with the two beams exchange polarimeter and the Second Solar Spectrum Atlas, many other observing programs have been carried out, and several of them are still ongoing.

The pointing of the telescope to a desired position is possible through the combination of different devices. The automatic guiding system (Küveler et al. 1998) allows us to maintain the pointing of the telescope within one arcsec. A recently developed encoder system allows precise pointing to desired coordinates within a couple of arcseconds. The image rotator coupled to the ZIMPOL modulator package allows the solar image to be rotated with respect to the spectrograph slit with an accuracy of $0.5^{\circ}$. Further, if observations are performed close to the solar limb (within $50\hbox{${}^{\prime\prime}$}$ inside the limb), a tilt plate system developed at FHS Wiesbaden allows to compensate for image motion effects along the direction perpendicular to the limb, with an accuracy of about one arcsecond.

An adaptive optics system, AO, similar to the infrared AO system used at Kitt Peak (Keller, Plymate, & Ammons 2003), is developed in collaboration with SUPSI. First tests show a clear improvement of the image quality (Ramelli et al. 2006, Figure 5).

Besides the spectrograph observations, a two-dimensional imaging polarimetry program, based on a LiNbO₃ Fabry-Perot filter, has been initiated (Feller, Bianda, & Stenflo 2007). A particular optical design that couples the filter with the spectrograph has recently been tested (Kleint et al. 2008).

IRSOL offers the opportunity to test and improve new instruments to be used in campaigns at major facilities, like at Kitt Peak, Arizona, Sacramento Peak, New Mexico, and Canary Islands. A recent example is the campaign at THEMIS on Tenerife, which was organized by IRSOL. In June-July 2008 the polarimeter ZIMPOL II with a dual FLC modulator was used by several international groups. For polarization observations at wavelengths ranging from about 450 nm to near IR, THEMIS is worldwide the best available telescope. The instrument is almost polarization free because the analyzer is inserted on the optical axis of the telescope, before the folding mirrors. The campaigns showed that the combination of ZIMPOL and THEMIS is excellent for high precision polarimetry in the visible and near IR.

Observing with ZIMPOL at IRSOL, very high polarimetric precision can be achieved from near UV to near IR. The instrumentation of the institute has been developed in order to optimize spectropolarimetric observations. Further performance-enhancing devices (AO, Fabry-Perot filter) are in the realization phase and will be used for scientific observations.

The need for high quality spectropolarimetric data is increasing, also because the scientific community is now developing a number of theoretical tools to better understand Second Solar Spectrum signatures. Observations are showing new aspects that need to be explained. There is also evidence that some lines of the Second Solar Spectrum are changing with the solar cycle. There is therefore much potential for future discoveries and new physical insights about the Sun. A decade ago it was already clear that we have opened a “new window” in solar physics (Stenflo & Keller 1997).

IRSOL has had the opportunity and pleasure to be part of this fascinating development, much thanks to important collaborations like the one with Jan Stenflo and his group. Now it is time to consider new international solutions to optimize the progress of our scientific community in the further exploration of the Second Solar Spectrum. New generation telescopes (ATST, EST, balloon and space missions) will come on line and are needed, but they lie either many years in the future, or are not always optimally designed for scattering polarization work. With already available ground-based instruments much excellent science still waits to be done. Thus there is a need to optimize the facilities and instrumentation that already exist.

Like the ZIMPOL campaign at THEMIS has shown, the collaborations between existing institutes can create synergies of advantage to the whole community. This is the path that we are determined to follow, with inspiration from the thought and art of Jan Stenflo.