Table of Contents
Cover
Title Page
Copyright
Foreword
Acknowledgments
The Emergence of 3D Spectroscopy in Astronomy
Scientific Rationale
3D History
3D Technology
Part I: 3D Instrumentation
Chapter 1: The Spectroscopic Toolbox
1.1 Introduction
1.2 Basic Spectroscopic Principles
1.3 Scanning Filters
1.4 Dispersers
1.5 2D Detectors
1.6 Optics and Coatings
1.7 Mechanics, Cryogenics and Electronics
1.8 Management, Timeline, and Cost
1.9 Conclusion
Chapter 2: Multiobject Spectroscopy
2.1 Introduction
2.2 Slitless Based Multi-Object Spectroscopy
2.3 Multislit-Based Multiobject Spectroscopy
2.4 Fiber-Based Multiobject Spectroscopy
Chapter 3: Scanning Imaging Spectroscopy
3.1 Introduction
3.2 Scanning Long-Slit Spectroscopy
3.3 Scanning Fabry–Pérot Spectroscopy
3.4 Scanning Fourier Transform Spectroscopy
3.5 Conclusion: Comparing the Different Scanning Flavors
Chapter 4: Integral Field Spectroscopy
4.1 Introduction
4.2 Lenslet-Based Integral Field Spectrometer
4.3 Fiber-Based Integral Field Spectrometer
4.5 Conclusion: Comparing the Different IFS Flavors
Chapter 5: Recent Trends in Integral Field Spectroscopy
5.1 Introduction
5.2 High-Contrast Integral Field Spectrometer
5.3 Wide-Field Integral Field Spectroscopy
5.4 An Example: Autopsy of the MUSE Wide-Field Instrument
5.5 Deployable Multiobject Integral Field Spectroscopy
Chapter 6: Comparing the Various 3D Techniques
6.1 Introduction
6.2 3D Spectroscopy Grasp Invariant Principle
6.3 3-D Techniques Practical Differences
6.4 A Tentative Rating
Chapter 7: Future Trends in 3D Spectroscopy
7.1 3D Instrumentation for the ELTs
7.2 Photonics-Based Spectrograph
7.3 Quest for the Grail: Toward 3D Detectors?
7.4 Conclusion
7.5 For Further Reading
Part II: Using 3D Spectroscopy
Chapter 8: Data Properties
8.1 Introduction
8.2 Data Sampling and Resolution
8.3 Noise Properties
Chapter 9: Impact of Atmosphere
9.1 Introduction
9.2 Basic Seeing Principles
9.3 Seeing-Limited Observations
9.4 Adaptive Optics Corrected Observations
9.5 Other Atmosphere Impacts
9.6 Space-Based Observations
9.7 Conclusion
Chapter 10: Data Gathering
10.1 Introduction
10.2 Planning Observations
10.3 Estimating Observing Time
10.4 Observing Strategy
10.5 At the Telescope
10.6 Conclusion
Chapter 11: Data Reduction
11.1 Introduction
11.2 Basics
11.3 Specific Cases
11.4 Data Reduction Example: The MUSE Scheme
11.5 Conclusion
Chapter 12: Data Analysis
12.1 Introduction
12.2 Handling Data Cubes
12.3 Viewing Data Cubes
12.4 Conclusion
12.5 Further Reading
Chapter 13: Conclusions
13.1 Conclusions
13.2 General-Use Instruments
13.3 Team-Use Instruments
13.4 The Bumpy Road to Success
References
Index
End User License Agreement
Pages
1
2
3
4
5
6
7
8
9
10
11
13
14
15
16
17
18
19
20
21
22
23
24
25
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
81
82
83
84
85
86
87
88
89
90
91
92
93
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
115
116
117
118
119
120
121
122
123
124
125
126
129
130
131
132
133
137
138
139
140
141
142
143
144
145
146
147
151
153
154
155
156
157
158
159
160
161
162
163
164
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
199
200
201
202
203
204
205
206
207
208
209
210
211
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
245
246
247
248
249
250
251
269
270
271
272
273
274
275
276
Guide
Cover
Table of Contents
Foreword
Begin Reading
List of Illustrations
The Emergence of 3D Spectroscopy in Astronomy
Figure 1 HST Ultra Deep Field (credit AURA/STScI). This is a small (2.4 across) “randomly” located multiband image taken with an unprecedented depth by the Hubble Space Telescope over the 0.4– wavelength domain. Apart for a handful of forefront Milky Way stars, the field is dotted by many galaxies, some dating back to a mere million years after the birth of our 13.6 billion-year old Universe.
Figure 2 The pinwheel galaxy Messier 101 (credit European Space Agency & NASA). This Hubble space telescope image shows a prototypal spiral galaxy, made of about 100 billion stars orbiting around the galaxy center and surrounded by a huge dark matter halo. The fat central bulge holds the oldest stars dating back from the galaxy formation some 13 billion years ago. It is surrounded by a flat rotating disk of stars, gas, and dust with continuous star formation till now. The UV-bright young disk star clusters are enveloped by large ionized gas regions, easily studied spectroscopically from their bright narrow emission lines. Typical radial velocity range covered by the ionized gas clouds and the stars is .
Figure 3 Long-slit spectrum of our nearby galaxy cousin, the Andromeda Nebula (Messier 31). The slit is vertical, with spectral dispersion in the horizontal direction. Central part : galaxy spectrum with strong calcium absorption lines, coming from the integrated light of a vast number of stars. Upper and lower part : comparison emission lines spectrum of a gas discharge hydrogen lamp, used to precisely derive the nebula absorption line wavelengths. This 8.8 h-long observation was done by E. Fath in 1908 on the 38 in. diameter Lick telescope.
Figure 4 Fabry–Pérot rings illuminated by the 372.7 nm singly ionized oxygen doublet, projected on an image of the Orion Nebula, a nearby (1350 light-years away) gas cloud ionized by hot blue stars. Ring radii, compared to those given by a laboratory gas discharge lamp, directly gave the gas radial velocity. This 40 min integration was performed by Fabry et al . in 1914 at the 80 cm telescope at Marseilles Observatory. Credit Observatoire de Marseille.
Figure 5 Schematic three-dimensional (sky coordinates , ; wavelength ) illustration of the current three main 3D phyla spatiospectral coverage. These are multiobject spectrography (MOS, Section 2.1), scanning Fabry–Perot spectrography (SFPS, Section 3.3.3) or Fourier transform spectrography (Section 3.4), and finally integral field spectrography (IFS, Section 4.1). Note the very different shapes of the data “cubes” delivered by the instruments.
Figure 6 First 1971 echelle spectrum (order 67 to 57 from top to bottom), obtained at the Pine Bluff 36” Telescope on the very bright star Capella. It roughly covers the 360–430 nm UV-violet spectral range. Wavelength increases from left to right and from top to bottom. Many absorption lines, coming in particular from Ca, K, and Fe atoms, can be seen. Credit University of Wisconsin, USA.
Figure 7 First multifiber data in the optical range within the galaxy cluster Abell 754 by Hill et al . in 1979 at the Steward Observatory 2.3 m telescope. Spectral dispersion is horizontal, with 37 fibers distributed along the vertical axis. Of these, 32 fibers were put on the sky objects in the 0.5 patrol field, and 5 were locally illuminated by a calibration HeAr gas discharge lamp.
Figure 8 First multislit (actually multiholes) data in the optical range within the galaxy cluster 0054+1654 by Butcher et al. in 1982 on the Mayall 4-m telescope at Kitt Peak National Observatory. The spectra cover the 500–900 nm range with a spectral resolution . The picture shows a portion of the full data set. Spectral dispersion is vertical. The bright spectrum on the left is of an alignment star.
Figure 9 The 1980 first light of a fiber-based integral field spectrograph demonstrator at the Mauna Kea 2.2 m UH telescope by C. Vanderriest [11] on the active galaxy 3C 120 nucleus. The hexagonal science field is paved by 169 fibers, with four additional rows of 9 fibers each to monitor an empty sky region. Each 0.8 diameter fiber gives a spectrum over the 390–690 nm spectrograph range with spectral resolution. (a) Fiber-based image dissector input. (b) A few successive spectra from a 20-min exposure on the object, including one spectrum (bottom) from an empty sky location.
Figure 10 First light of the lenslet-based TIGER integral field spectrograph by Adam et al. 1989 [8]. The field is paved by 70 lenses, giving 70 spectra in the 370–689 nm range with spectral resolution. (Left) Direct visible image of the Einstein Cross, the four-element gravitational mirage of a distant quasar by the nearby galaxy located at the center. (Right) Reconstituted image from the 70 spectra in the redshifted 190.9 nm doubly ionized carbon emission line, showing three of the four mirages. Credit Observatoire de Lyon, France.
Figure 11 First light of the SPIFFI slicer-based integral field spectrograph by Krabbe et al . [12] at the 2.2 m La Silla telescope. The field of view is sliced in 256 slits, each giving K-band spectra at spectral resolution. Central square: Reconstituted K-band (1.85–2.4 ) image of the central star cluster in our Galaxy. Surrounding boxes: Extracted K-band spectra of stars, orbiting around our Galaxy 3.5 million solar masscentral black hole.
Chapter 1: The Spectroscopic Toolbox
Figure 1.1 Light propagation at an air–glass interface.
Figure 1.2 Visualization of the infinitesimal etendue component .
Figure 1.3 Schematic illustration of the 2D etendue conservation for any optical system and the 1D etendue conservation for any centered system.
Figure 1.4 Principle of the optical fibers. Light entering the fiber core is trapped by total reflection at the core-cladding interface and propagates to the fiber end.
Figure 1.5 Principle of the Fabry–Pérot interferometer. Light rays are trapped by multiple reflections inside a cavity of depth e. The cavity acts as a spectral filter as only rays undergoing wavelength-dependent constructive interference are transmitted; all the others being reflected by the etalon.
Figure 1.6 (a) Set of rings ( to 5) from monochromatic light uniformly illuminating a Fabry–Pérot etalon, with exact phasing at the center ( ). The successive ring radii obey the canonical equation , which for small angles translates into ring radii . With rings FWHM , rings etendues are all the same. (b) Transmission function for two etalons of respective finesses 2 (blue) and 10 (red).
Figure 1.7 Prism's principle: The Figure shows a prism used at minimum deviation for the central wavelength (green ray). This is a symmetrical configuration with similar beam incident and emergent angles . The extreme wavelength beams are shown in red and blue.
Figure 1.8 Grating's Principle's: A parallel polychromatic light beam (black rays) falls on a plane reflection grating at incidence angle . In that illustration, first order green rays (not shown) are diffracted back at the same angle along the input beam (Littrow condition), while extreme wavelength beams are shown respectively in red and blue.
Figure 1.9 3D view of the long-slit grating spectrograph concept. (a) The 2D image at telescope focus is sliced by a long narrow horizontal slit. (b) Light on the detector is dispersed in the vertical direction. Note that we follow in this book the two usual conventions for Figure showing light propagation inside an optical system: (i) whenever possible, light goes from left to right (top to bottom for a vertical optical axis) and (ii) the horizontal and vertical scales are generally not the same, usually exaggerating the angles of the optical beams for better clarity.
Figure 1.10 This shows the state-of-the-art , pixels CCD231-84 from e2v, one of the leaders in the field. Since this high performance device is four-side buttable, it can be used as a building block for the development of extremely large mosaics. Credit Paul Jordan [14], e2v, the UK.
Figure 1.11 Illustration of the deleterious effect of non-telecentricity. (a) Three telecentric beams (with parallel optical axes) fall on a detector. Non-perfect flatness of the detector degrades the images, but does not move their centers of gravity with respect to each other. (b) The same, but for non-telecentric beams (optical axes not parallel). There is a similar image degradation, but now their centers of gravity are displaced with respect to each other. This leads to significant measuring errors, typically a few micrometers for, say, flatness deviation.
Figure 1.12 Illustration of the Schmidt mounting, in essence a spherical mirror with the entrance pupil located at its center of curvature. Two parallel light beams at two different inclinations are shown. Owing to the system's full rotational invariance, all input parallel beams are imaged with the same (small) aberration irrespective of their 3D inclination, a trick first discovered by the philosopher (and lens maker) B. Spinoza in 1600 and implemented by B. Schmidt in the 1930s. This field-invariant aberration can, for example, be canceled by adding an aspheric window on the pupil; its correction effect then varies over the field, but only as a cosine function, which in most cases is good enough. Again, because of rotational invariance, the images are located on a spherical segment with its center of curvature on the pupil. A field flattener (a thick convergent lens possibly doubling as the detector entrance window) can also be placed just before the focus.
Figure 1.13 The so-called three mirror anastigmat (TMA) is a centered optical system (i.e., with a common optical axis) made of three highly aspheric mirrors, the shapes and positions of which are tuned to give extremely good images on a flat focal plane over a wide field of view. In real life, to avoid 100% beam obscuration by , only off-axis cuts of the three mirrors are used. The TMA can be used as a camera with light reflected successively by , , and , or as a collimator when used in reverse.
Figure 1.14 This classical kinematic mounting features three optically polished sapphire spheres glued at to the underside of a component and three right-angle hardened ground steel grooves at on top of its mounting plate. This gives the required six contact points, ensuring incredibly stable and repetitive positioning. The upper component can be removed and then put back (gently) in place, and without any adjustment repositions itself within a fraction of a micron. Note that this requires only very lax absolute accuracy (say only at the millimeter level) for the relative positions of the spheres and grooves.
Figure 1.15 The three main telescope foci are shown, namely prime focus, Cassegrain focus, and Nasmyth focus. Additional mirrors are needed for the folded Nasmyth and coudé foci. As the telescope tracks during the night along two orthogonal axis, much like a warship turret, instruments at prime focus and Cassegrain focus move along, and on top usually rotate to cancel field rotation. At Nasmyth focus, owing to the rotating tertiary mirror, light is sent to a horizontal rotating platform when the instrument sits; field rotation has still to be canceled, though.
Figure 1.16 This is a schematic view of the classical three-mirror derotator in the case of a parallel beam input. It works also with a convergent beam, for example, with an image of the field on mirror #2. Field rotation is nulled by counterrotating the derotator around its horizontal axis. For small enough light beams, a prism with three internal reflections can be used instead.
Figure 1.17 Project Funnel. This small cartoon illustrates how starting from a broad concept, any instrumental project becomes more and more tightly defined as it moves through successive stages toward start of operation. Along the way, uncertainties, in terms of cost, timeline, and/or performance drastically decrease, that is for a successful project.
Chapter 2: Multiobject Spectroscopy
Figure 2.1 Slitless spectrograph concept . An optical relay re-images the sky field at the focal plane of the telescope on the 2D detector at an appropriate focal ratio, typically in the F/1.5 to F/3 range. It also gives an intermediate pupil on which a prism or grism is inserted to provide spectra from any object in the field.
Figure 2.2 (a) 115–362 nm UV image of the young star cluster NGC 604 in the nearby M33 galaxy taken with the STIS instrument aboard the Hubble Space Telescope. (b) STIS slitless prism spectra in the same region. Dispersion is nearly horizontal, and spectral resolutions range from 2500 at 115 nm to 50 at 362 nm.
Figure 2.3 Principle of the multislit spectrograph. Many short narrow slitlets are put on selected objects in the field of view. An optical relay re-images the slitlets at the telescope focal plane on the 2D detector at an appropriate focal ratio, typically in the F/1.5 to F/2.7 range. It also gives an intermediate pupil on which a prism or grism is inserted to provide spectra from every slitlet. With no disperser inserted, the instrument gives direct sky images that can be used to measure the accurate positions of potential targets for further multisilt observations.
Figure 2.4 First VIMOS exposure in its multiobject mode. This optical instrument was built by a European Consortium led by O. LeFèvre (LAM, Marseille) for the European Southern Observatory (ESO). Two of the four quadrants are shown. Dispersion is in the vertical direction. A total of 221 low-resolution spectra on as many galaxies have been obtained in this single exposure. ©The European Southern Observatory.
Figure 2.5 Picture of a TI. DMD used in digital projectors.
Figure 2.6 Two close-up views of the JWST micro-shutter waffle-like array. The entire array is made of four quadrants, each the size of a postage stamp. Credit NASA.
Figure 2.7 Principle of the multifiber spectrograph. Optical fibers are put on selected objects in the field of view and rearranged in a pseudo long-slit at the entrance of a classical long-slit spectrograph. The spectrograph gives a spectrum for each of the selected targets.
Figure 2.8 Picture of the two-degree field multifiber positioning system installed at the prime focus of the 4-m diameter Anglo-Australian Telescope. The pick and place robotic arm is on top, with the currently addressed focal plane plate at the bottom.
Figure 2.9 A highly schematic view of the fishermen pond multifiber positioner system.
Figure 2.10 The Echidna positioning systems concept developed by the Australian Astrophysical Observatory. It uses piezo-activated fiber-carrying spines distributed evenly at the telescope focal plane. On-target pointing of each spine inside its own small 2D patrol field is achieved by spine flexing.
Figure 2.11 Highly schematic view of the Starbug concept, developed by the Australian Astrophysical Observatory. Piezo-driven semi-autonomous robots walk on the focal plane window to their designed target locations. They are prevented from falling by vacuum suction.
Chapter 3: Scanning Imaging Spectroscopy
Figure 3.1 First scanning long-slit data cube on the galaxy NGC 5128 (Centaurus A) at the 4-m Anglo Australian Telescope by Wilkinson et al . 1986. (a) Optical image of the galaxy field, credit the European Southern Observatory, PR image eso0005b. (b) Data cube covering the blue to yellow optical region with spectral resolution. Strong magnesium (Mg) and sodium (Na) lines are indicated. The slit length is horizontal. The vertical direction has been filled by 71 successive sky integrations, moving the slit on the galaxy each time by about one full slit width.
Figure 3.2 First 2D Fabry–Pérot data on an external galaxy by Carranza et al . [39]. One sees etalon ring fragments in the 656.3 nm line of ionized hydrogen, fed by the extended ionized gas regions in the Messier 33 galaxy that happens to have the right wavelength–radius combination to be transmitted by the etalon. 1024 different radial velocities were extracted in the 6 arcmin diameter field, giving the first global ionized gas radial velocity field in a galaxy.
Figure 3.3 1974 first 3D scanning Fabry–Pérot data [4] at the KPNO 84 in. telescope. (b) One of the 16 successive exposures in the 6562.3 nm ionized hydrogen line in the Messier 51 galaxy, showing ring fragments fed by the galaxy ionized gas regions. (a) Illustration of the scanning concept. The etalon cavity is changed by 1/16 of an order for each of the 16 successive exposures and the light intensity curve on each spatial pixel is extracted. This gives a data “slab” made of a small spectrum for each of the spatial points in the field.
Figure 3.4 Scanning Fabry–Pérot spectrograph concept. The 2D square or circular field is collimated, with the pupil image put on a Pérot–Fabry etalon. One etalon order is selected by an interference filter (not shown here) and the field is re-imaged by the camera on the 2D detector. Scanning the etalon builds the same small-size spectrum at each field/detector pixel.
Figure 3.5 (a) Schematics of the simplest incarnation of a Fourier transform spectrometer. For astrophysical use, a second detector is added on the horizontal return beam to avoid losing 50% of the light. (b) Output signals (three red dots) for three moving mirror positions, respectively with fully constructive, half constructive, and fully destructive interference.
Figure 3.6 Schematics of retroreflectors used for Fourier transform spectroscopy. (b) Corner cube prism with three orthogonal faces, which reflects back any incoming ray parallel to itself. (a) Cat's eye using a parabolic main mirror and a small plane mirror at its geometrical/optical focus, with the incoming light beam first reflected by , then by , and finally sent back to the beamsplitter by . Any moderately tilted incoming ray is reflected back in an almost parallel direction.
Figure 3.7 An early result of the Sitelle Fourier Transform Spectrograph at CFHT on the diameter nearby galaxy Messier 51. (a) Reconstructed image in the ionized hydrogen 656.3 nm line. (b) Corresponding color-coded radial velocity field, mainly due to rotation of the ionized gas clouds around the galaxy center.
Chapter 4: Integral Field Spectroscopy
Figure 4.1 The lenslet-based integral field spectrograph instrumental concept. Left to right : An enlarger plus field lens combination produces a highly magnified telecentric image of the small field of view at the entrance of a 2D microlens array. Each lens samples a very small spatial field and produces a small circular exit pupil, called a micropupil. These micropupils act as the equivalent of the slitlets at the entrance of a classical multislit spectrograph, which gives a spectrum stack (one per each micropupil) on the detector. The (single) exit pupil at the level of the disperser is actually filled by highly enlarged stacked images of each small fraction of the field of view sampled by the microlens array.
Figure 4.2 Illustration of the uniformly illuminated spatial point spread function of one lenslet micropupil. One sees the telescope central obscuration by its secondary mirror.
Figure 4.3 Examples of cross-dispersion profiles for values of 2 (solid line), 1.75 (dashed line), and 1.5 (solid-dashed line) in the case of constant intensity spectra. Increasing cross-talk as spectra are more closely packed on the detector can be clearly seen.
Figure 4.4 The picture shows the microlens array of the OASIS integral field spectrograph, made of 1500 identical hexagonal lenses. With a number of exchangeable enlargers before the lens array, sky samplings ranged from to . Adaptive optics corrections (Section 9.4) were applied when working with every sampling but the coarser ones.
Figure 4.5 The SAURON integral field spectrograph at the Cassegrain focus of the 4.2 m William Herschel Telescope (La Palma, 2000).
Figure 4.6 Fiber-based integral field concept (Courtesy C. Vanderriest). The Figure shows its key ingredient, a multifiber image dissector that changes the input squared field format into a narrow slit-like output format, well-suited to directly fill a classical long-slit spectrograph. Credit Observatoire de Paris, France.
Figure 4.7 MPE 3-D/SPIFFI mirror slicer concept, shown here with 3 of its original 16 slices. (b) Telecentric beams from the telescope fall on a “staircase” made of a stack of thin (0.1–0.2 mm) tilted flat primary mirrors at the telescope focal plane . The stack slices the small 2D field and sends the light back to flat secondary mirrors that redirect all light beams parallel to their common initial optical axis. With the secondary mirror vertexes lying on a paraboloid of focus , Fermat's principle (constancy of the optical paths , here all nulls) ensures that all output light beams originate from the virtual slit shown here. (a) Illustration of field slicing and its reorganization into a staggered (virtual) pseudo-slit.
Figure 4.8 Advanced image slicer concept. Telecentric input beams (bottom right) from the telescope fall on a staircase stack, made of thin curved primary mirrors , which slices the small sky field. small curved secondary mirrors are put on the respective pupil images given by the primary mirrors. They re-image the field slices on a real pseudo-slit. Field lenses or mirrors are added near this pseudo-slit to give telecentric output beams (top left) that input the associated spectrograph. Figure from the Gemini Telescope web site (http://www.gemini.edu/node/21).
Chapter 5: Recent Trends in Integral Field Spectroscopy
Figure 5.1 November 2013 direct detection in the K band of three giant ( 10 Jupiter mass) exoplanets around the young (60 million-year old) star HR 8799 by the Gemini-South GPI instrument. Image credit: Christian Marois (NRC Canada), Patrick Ingraham (Stanford University) and the GPI Team. From the Gemini Telescope website (www.gemini.edu/node/12314).
Figure 5.2 This picture shows the MUSE instrument installed at one Nasmyth focus of the ESO VLT 8.2 m diameter Unit Telescope #4. The instrument is largely hidden behind the small forest of cables needed to operate its main body and the 24 subunits.
Figure 5.3 MEIFU Concept. Spectra are dispersed at a angle to avoid spectral overlap. Each spectrum covers pixels in the blue and pixels in the red region on the detector ( per pixel). Interspectra gaps are 26 pixels (spectral) and 3 pixels (spatial).
Figure 5.4 MUSE image slicer and spectrograph unit. This drawing shows light propagation inside one of the 24 MUSE spectrographic Units. Note that the drawing is to scale.
Figure 5.5 Picture of one of the MUSE image slicers.
Figure 5.6 Schematics of the FLAMES deployable integral field units (IFU) system.
Figure 5.7 Photographic image ann12071a of the KMOS focal plane, showing its pickoff arms and image slicers.
Chapter 7: Future Trends in 3D Spectroscopy
Figure 7.1 Night-sky emission spectrum from 1 to at the 4200-m high Mauna Kea Observatory, in dry conditions. Given their huge range, OH line intensities are shown in a semi-logarithmic plot, with the flux unit in ph/sec/(arcsecond) . Note the interlines low intensity level, at least 1000 times fainter than the brightest OH lines. The strong positive slope beyond is due to atmosphere thermal emission.
Figure 7.2 OH-suppression filter concept. (a) Schematic of an individual fiber with a laser-written internal Bragg grating. (b) Rejection of a large number of airglow lines in the H-band. Note that 10 dB corresponds to a rejection factor of 10, and 20 dB to 100.
Figure 7.3 Arrayed waveguide principle. (a) Illustration of the wavelength dispersing effect of this component. (b) Picture of an actual AWG device.
Figure 7.4 Photonics Fourier transform spectrometer SWIFTS concept.
Chapter 8: Data Properties
Figure 8.1 Examples of various analytical forms of 2D PSF. As shown in (a), all PSFs have been set to the same FWHM. However, their contribution can be very different when proper normalization by the total flux is taken into account (b). Note the impact of the very extended wings of the MOFFAT model with , which decreases the contrast of the central peak.
Figure 8.2 Example of resolving power as a function of wavelength for two types of dispersers: a grating and a prism.
Figure 8.3 Examples of LSF for various types of spectrographs: slit-based, fiber-based, lenslet-based, Fabry–Pérot, and FTS.
Figure 8.4 Examples of various detector noise patterns: (a) original source, (b) photon noise, (c) readout noise, (d) pickup noise.
Figure 8.5 Example of impact of cosmic rays in a thick red-sensitive CCD.
Chapter 9: Impact of Atmosphere
Figure 9.1 Atmospheric transmission in the optical and near-IR windows.
Figure 9.2 Astronomical seeing. Light wavefronts coming from a point-like astronomical object are almost perfectly plane. Images given by a “perfect” telescope would then be very small, limited by light diffraction to sub- values for meters-size telescopes. Unfortunately, wavefronts observed at ground level are severely distorted by turbulent layers in our Earth's atmosphere. This degrades image quality to much larger values, almost never below , even at the best world sites.
Figure 9.3 Turbulent Stellar Images. (a) Image of a point-like star taken through the Earth's atmosphere, with an exposure time short enough (a few ms at optical wavelengths) to freeze atmospheric turbulence. The irregular speckle pattern observed is due to light wavefront corrugations. Individual speckle size corresponds to the telescope diffraction limit, a few hundreds of an arcsecond for a large telescope at optical wavelengths. (b) Corresponding long time exposure image, that is, long enough to smooth out the dancing speckle pattern in (a), viz., a few seconds at optical wavelengths. Image size is at best a few tenths of an arcsecond across, even at the very best sites.
Figure 9.4 Principle of an adaptive optics control loop. The turbulent wavefront (top) falls on a deformable mirror (DM) and goes to a beam splitter. Light from the science object is sent to the instrument; light from a reference star in the field of view goes to the wavefront sensor (WFS). Signals from the WFS detector are processed by the real time computer (RTC), which sends appropriate voltages to the DM actuators to flatten the wavefront.
Figure 9.5 Single conjugate adaptive optics (SCAO). This is a highly schematic view of the simplest adaptive optics system. A single reference star in the observed field illuminates a single wavefront sensor. The real time computer sends commands to the single deformable mirror usually conjugated with the telescope pupil (hence the SC in SCAO), giving near diffraction-limited image correction in the immediate vicinity of the reference star. Typical correction parameters for an NIR AO system on a 4- to 8-m class telescope are response time a few milliseconds, number of corrected elements a few hundreds to a thousand. (Used Under Creative Commons License: https://creativecommons.org/licenses/by/4.0/.)
Figure 9.6 The Figure shows 2D image profiles of a (point-like) star in the H-band given by the Keck1 10-m diameter telescope adaptive optics imager in 2001. (a) Natural seeing image, exhibiting a shallow light distribution with FWHM. (b) AO-corrected image with two basic components, a sharp peak ( FWHM) carrying 23% of the total star light, and a shallow halo, similar to the natural seeing PSF and carrying the remaining 77%. Note the spectacular contrast gain, between the two image peaks. (Courtesy W.M. Keck Observatory.)
Figure 9.7 Principle of the piezo deformable mirrors. A thin glass plate is bonded on a 2D array of piezoelectric actuators. High voltages (hundreds of volts) are applied to the actuators in order to deform the front glass plate up to a few microns, hundreds of times per second. (Reproduced with permission of Tokovinin.)
Figure 9.8 Principle of the 2-D Shack–Hartmann wavefront sensor. An image of the telescope pupil is paved by a 2D lenslet array, each giving an image of a reference star on the detector. Top : star images for a plane wavefront. Bottom : star images for a distorted wavefront. Locations of the multi-images of the reference star measured on the fly by the sensor reflect the instantaneous wavefront local slopes. (Reproduced with permission of Tokovinin.)
Figure 9.9 AO Isoplanatic field limitation for a single natural reference star. The Figure schematically shows the negative impact of the natural reference star offset angle from the science target, as the wavefront sensed by the reference star beam is increasingly different from the science target ones as the offset angle increases, especially for the high altitude turbulence layers. (Used Under Creative Commons License: https://creativecommons.org/licenses/by/4.0/.)
Figure 9.10 Upper atmosphere illumination by the ALFA Laser Guide Star system at the GEMINI-S 8-m telescope. The small ( diameter) pseudo-star at the center of the Figure comes from excited sodium atoms in the high atmospheric layer above the Earth. The extended parasitic light at the bottom right comes from Rayleigh scattering of the upcoming laser beam in the atmosphere above the telescope, up to about 15 km. (Reproduced with permission of Gemini Observatory.)
Figure 9.11 AO isoplanatic field limitation with a laser guide star. The Figure schematically shows the poor matching of the light cylinder from the science target (blue) by the light cone coming from a single laser guide “star” (yellow), except close to the ground. (Used Under Creative Commons License: https://creativecommons.org/licenses/by/4.0/.)
Figure 9.12 Multiple laser guide stars. The first ever laser guide star “constellation” launched in January 2011 at Gemini South. The image shows the 50-W laser beam as it shines upward toward the 95-km-high atmospheric sodium layer to create a pattern of five artificial guide stars (upper left) used to sample atmospheric turbulence for the Gemini Observatory GeMS adaptive optics system. The yellow-orange beam visible from the lower right to the upper left is caused by scattering of the laser's light by the Earth's lower atmosphere. (Used Under Creative Commons License: https://creativecommons.org/licenses/by/4.0/.)
Figure 9.13 Laser tomography. The picture schematically shows the turbulence layers affecting the science object covered by two laser cones. In practice, fair corrections can be achieved with three to four 20 W, continuous lasers. (Source : http://www.eso.org/sci/meetings/2015/EriceSchool2015/Erice_Marchetti_1.pdf Used Under Creative Commons License: https://creativecommons.org/licenses/by/4.0/.)
Figure 9.14 Multiconjugate adaptive optics (MCAO). This is a highly schematic view of the MCAO principle. Multiple wavefront sensors, each illuminated by a natural or artificial star (hence the star-oriented label), are used here to sense separately the two main turbulence layers, at least one close to the ground and one 4–5 km above the telescope. Commands are sent to two deformable mirrors optically conjugated to these two layers, achieving diffraction-limited correction over a few arcminutes field. (Used Under Creative Commons License: https://creativecommons.org/licenses/by/4.0/.)
Figure 9.15 Ground-layer adaptive optics (GLAO). This is a highly schematic view of the GLAO principle. Multiple wavefront sensors, each illuminated by a natural (a) or artificial (b) star, are used to sense the turbulence layer close to the ground. Commands are sent to a single deformable mirror optically conjugated to the ground, thus erasing the ground layer contribution over a few arcminutes field. Courtesy E. Marchetti, ESO. (©The European Southern Observatory.)
Figure 9.16 Sinfoni Galactic Center data. (a) Reconstructed K-band image of a small field around the center of our galaxy (open circle) with a spatial resolution of . (b) Typical H-band spectrum of these central stars. Their hydrogen and helium absorption lines are used to get stellar radial velocities and physical parameters. (Reproduced with permission of Frank Eisenhauer.)
Figure 9.17 Multiobject adaptive optics (MOAO). This is a highly schematic view of the MOAO principle. The left image shows the correction principle for one science object (covered by an Integral Field Unit) flanked by two wavefront sensors. The right image shows the field of view coverage with open-loop deformable mirrors on the science objects and closed-loop deformable mirrors + wavefront sensors on reference stars. (Used Under Creative Commons License: https://creativecommons.org/licenses/by/4.0/.)
Figure 9.18 Measured Paranal extinction curve in the visible range [99]. (Used Under Creative Commons License: https://creativecommons.org/licenses/by/4.0/.)
Figure 9.19 Air index of refraction as a function of wavelength for normalized conditions (sea level, temperature ).
Figure 9.20 Optical night sky brightness at the Mauna Kea site in units of . (Reproduced with permission of Gemini Observatory.)
Figure 9.21 Near-IR night sky brightness at the Mauna Kea site in units of . Note the dense emission lines forest over a much fainter continuum. (Reproduced with permission of Gemini Observatory.)
Chapter 10: Data Gathering
Figure 10.1 Example of a pointing chart used for MUSE observations of the Hubble Ultra Deep Field at the VLT. The MUSE field is visible (green rectangle), surrounded by the slow guiding area (dotted green curve). Note that the center of the field, the North-East orientation arrows, and the candidate guiding star are also shown.
Figure 10.2 The ESO VLT control room during day calibrations (Credits ESO).
Chapter 11: Data Reduction
Figure 11.1 Schematic of MOS or IFU data reduction process.
Figure 11.2 MUSE field splittings and data organization.
Figure 11.3 MUSE pipeline schematic.
Figure 11.4 MUSE raw detector calibration exposures: (a1) channel 12 of bias exposure, (b1) 30 mn dark exposure of the same channel. A zoomed window, displayed as a red square in (a) and (b), is shown in a2 (bias) and b2 (dark). Note the impact of cosmic ray in the dark exposure.
Figure 11.5 MUSE raw flat fields calibration exposures: (a1) channel 12 of internal tungsten lamp, (a2) zoom of the small window in a1, (b1) twilight sky exposure, and (b2) zoom of the small window in b1. In addition to the sun typical absorption spectral features, one can see the strong telluric absorption in the red due to the Earth's atmosphere.
Figure 11.6 Example of trace mask. The green and red curves give the edge location of each slice on the detector.
Figure 11.7 MUSE raw wavelength calibration exposures: (a1) channel 12 of the internal Ne arc lamp of 0.7 s exposure (zoomed view of the window marked in red is shown in a2). (b1) Example of a science 400 s exposure centered on NGC 3379 elliptical galaxy nuclei. Note the [OI] brightest sky line that can be seen on top of the galaxy bright continuum.
Figure 11.8 Example of MUSE wavelength calibration result. Here we see the residuals for one slice.
Figure 11.9 Examples of raw science exposures. For each object, one full channel is shown plus a zoom on a specific region (red rectangle). (a) The planetary nebulae NGC 3132 (60 s exposure). Note the bright emission lines all over the field. (b) Planet Saturn (1 s exposure). The deep absorption bands can be seen in the red region. (c) A 25 mn exposure in the Hubble Deep Field South. The raw image is dominated by the bright sky OH emission lines. (d) The brightest stars in the globular cluster NGC 6752 (120 s exposure) are clearly visible with their strong continuum.
Figure 11.10 Example of sky subtraction in a deep field 25 mn exposure. Spectrum of an ‘empty’ location averaged over a 1 arcsec aperture before (a) and after (b) the sky subtraction. Note that while proper sky subtraction indeed removes its average contribution, it does not (and cannot) remove its Poisson noise contribution. Flux are units.
Figure 11.11 Spectrophotometric calibrations: instrument response (in blue, arbitrary units) and telluric corrections (in red).
Figure 11.12 The spectrum of the sdO type HD49798 standard star (in blue) compared to a classical K star (in green).
Figure 11.13 The outskirts of the NGC 3201 globular cluster used for astrometric calibration (courtesy Sebastian Kamann). The MUSE field of view is displayed in red. The isolated stars (with less than 1% contamination from close neighbors within 2 arcsec) are identified as green circles overlaid over an HST image of the cluster.
Figure 11.14 Schematic of the drizzle scheme in 2D with three input data points that overlap onto one output pixel. The parts of the input data that overlap the output pixel and therefore contribute to it are drawn as opaque regions.
Figure 11.15 Spatial correlated properties in the MUSE data cube after drizzle interpolation. Each image shows the correlation map with its spaxel neighbors.
Figure 11.16 Monitoring of the readout noise levels of the 96 quadrants of MUSE detectors.
Figure 11.17 Evolution of the MOFFAT FWHM with wavelength measured on the deep observation of the Hubble deep field south with MUSE [101].
Chapter 12: Data Analysis
Figure 12.1 An example of a DS9 data cube view. The main DS9 window shows two frames: the right one (green color scheme) is the traditional white light image coming here from the MUSE Hubble deep field south observation. One monochromatic plane image (or slice) of the corresponding data cube is shown on the right frame with a blue color scheme. The wavelength (index 2286) has been selected from the cube with the slider window at the bottom right. Using the region type cursor, a small aperture was selected on the image (see the zoom window at the upper right) and the corresponding spectrum is shown on the right window (Plot3d).
Figure 12.2 QfitsView example session. The upper window shows the image plane view and the lower ones the spectral view. In this example, we have used the QfitsView capability to perform continuum subtraction by selecting the line and continuum emission wavelength range in the spectra window. The resulting narrow band image is displayed on the image view where only emission line objects at the selected wavelength show up.
Figure 12.3 Different views of the Orion nebulae MUSE data cube. (a) Color composite using three emission lines fluxes, in blue: , green: [N II] 6584, and red: [S II] 6731. (b) Color composite, showing emission line fluxes in red: Paschen 9, green: , blue: . (c) Enhanced color image, showing the three ionization levels of oxygen: red: [O III] 5007, green: [O II] 7320, blue: [O I] 6300.
List of Tables
Chapter 2: Multiobject Spectroscopy
Table 2.1 List of optical multislit spectrographs on 8–10 m class telescopes with their main characteristics: host telescope (Tel.), field of view (FoV) in arc minutes, maximum multiplex M, spectral resolution , and spectral range in
Table 2.2 List of NIR multislit spectrographs on 8-10 m class telescopes with their main characteristics: host telescope (Tel.), field of view (FoV) in arcminutes, maximum multiplex M, spectral resolution , and spectral range in
Table 2.3 Wide-field optical MOS survey capabilities with their main characteristics: instrument name, host telescope (Tel.) name and diameter, spectral range in , start date, patrol field area in square degree, multiplex M, spectral resolution , and main current or planned surveys
Table 2.4 Wide-field NIR MOS survey capabilities with their main characteristics: instrument name, host telescope (Tel.) name and diameter, spectral range in , start date, patrol field area in square degree, multiplex M, spectral resolution
Chapter 4: Integral Field Spectroscopy
Table 4.1 List of lenslet integral field spectrographs with their main characteristics: name, host telescope, start and when applicable end year, number of spatial elements, , number of spectral elements and , spectral length in pixels, , spectral resolution, range, spectral range in , and Ref., bibliographic reference
Table 4.2 List of fiber-based integral field spectrographs with their main characteristics: host telescope, start and end date, , spaxels number, , spectral pixel number, , spatial sampling ( ), , spectral resolution, and maximum spectral range ( )
Table 4.3 List of mirror slicer based integral field spectrographs on large telescopes with their main characteristics: host telescope, start date, , spaxels number, , spectral pixel number, , spatial sampling ( ), Range, spectral domain ( ), , spectral resolution, Ref., bibliographic reference
Chapter 5: Recent Trends in Integral Field Spectroscopy
Table 5.1 List of deployable multi-integral field spectrographs with their main characteristics: start date, , number of probes, , number of spaxels per probe, PF, patrol field in square degrees, , spectral resolution, and spectral range in micrometer
Chapter 6: Comparing the Various 3D Techniques
Table 6.1 Tentative ratings of the various 3D spectroscopic instrumental flavors
Chapter 11: Data Reduction
Table 11.1 Example of a pixel Table content created by the scipost pipeline recipe. XPOS and YPOS are the spaxel coordinates (here as offset from the field center) and LAMBDA is the wavelength in
Optical 3D-Spectroscopy for Astronomy
Roland Bacon and Guy Monnet
Authors
Roland Bacon CRAL – Observatoire de Lyon 9, avenue Charles André 69230 Saint-Genis-Laval France
Guy Monnet CRAL – Observatoire de Lyon 9, avenue Charles André 69230 Saint-Genis-Laval France
Cover Galaxy: (c) NASA
All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.
Library of Congress Card No.: applied for
British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.
Bibliographic information published by the Deutsche Nationalbibliothek Die Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.d-nb.de.
© 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany
All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law.
Print ISBN: 978-3-527-41202-0
ePDF ISBN: 978-3-527-67485-5
ePub ISBN: 978-3-527-67484-8
Mobi ISBN: 978-3-527-67483-1
oBook ISBN: 978-3-527-67482-4
Cover Design Schulz Grafik-Design, Fußgönheim, Germany
The development of astronomy has gone hand in hand with advances in technology. A celebrated step was taken by Galileo Galilei in 1609 when he used the newly invented telescope to observe the night sky and discovered mountains on the Moon, the phases of Venus, and four moons orbiting Jupiter. This revolutionized our world view, and was made possible by the increased light-gathering power and image sharpness provided by the 5 cm telescope lens compared to using the naked eye.
Since then, a number of such transformational steps enabled by new technology have occurred. The move from the lenses of refracting telescopes to the mirrors of reflecting telescopes allowed those telescopes to be of much larger diameter. Replacing the human eye as the detector behind the telescope, first with photographic plates and subsequently with almost perfectly sensitive electronic detectors, made it possible to collect light over time and hence observe much fainter objects. Dispersing the light into a spectrum revealed the physical nature of objects through the study of absorption and emission lines. The detection of infrared light and radio waves from the ground expanded astronomers' view beyond the wavelengths of visible light, and the launch of telescopes into space gave access to the entire electromagnetic spectrum. With the ability to detect particles and, most recently, gravitational waves emitted by celestial objects, astronomers now have even more ways of observing the Universe.
Spectroscopy in the optical and near-infrared regions was initially possible with a single aperture, which was adequate for observing stars. The development of spectroscopy with a slit allowed a more efficient study of extended objects and, more recently, the ability to perform spectroscopy over an extended area has once again provided an enormous jump in capabilities. This latest revolution is the topic of this book. The techniques for integral-field spectroscopy in the visual and infrared wavelength region have now matured to a level where the angular resolution of the spectroscopic observations can be as high as is achievable in direct imaging, and many telescopes have been equipped with such integral-field spectrographs, with others under development for the next generation of giant telescopes. A comprehensive overview is hence timely.
The authors are world-renowned experts who have had a major role in driving the development of integral-field spectroscopy from initial prototypes such as TIGER on the Canada France Hawaii Telescope and SAURON on the William Herschel Telescope to the transformational MUSE