Third AO4ELT Conference - Adaptive Optics for Extremely Large Telescopes
Florence, Italy. May 2013
ISBN: 978-88-908876-0-4
DOI: 10.12839/AO4ELT3.13288
Altair at Gemini North: Full Sky Coverage Laser AO
Correction at Visible Wavelengths
Chadwick Trujillo1a , Jesse Ball1 , Maxime Boccas2 , Chas Cavedoni1 , Julian Christou1,3 ,
Dolores Coulson1, Angelic Ebbers1 , Kimberly Emig1 , Inger Jørgensen1 , Stacy Kang1 , Olivier
Lai1 , Anthony Matulonis1, Richard McDermid1,4, Bryan Miller2 , Benoit Neichel2,5 , Richard
Oram2,6 , François Rigaut2,7 , Katherine Roth1 , Thomas Schneider1 , Andrew Stephens1 , Gelys
Trancho2,8 , Brian Walls1,8, and John White1
1
Gemini North, 670 N. A‘ohoku Place, Hilo, Hawaii, 96720, USA
Gemini South c/o AURA, Casilla 603 La Serena, Chile
LBT Observatory, University of Arizona, 933 N. Cherry Ave, Room 552, Tucson, AZ 85721,
USA
Australian Astronomical Observatory, PO Box 915, North Ryde NSW 1670
Laboratoire D’Astrophysique De Marseille, Pôle de l’Étoile Site de Château-Gombert 38, rue
Frédéric Joliot-Curie 13388 Marseille cedex 13 FRANCE
LIGO Livingston Observatory, 19100 LIGO Lane, Livingston, LA 70754
Australian National University, Canberra ACT 0200, Australia
GMTO Corporation, 251 S. Lake Avenue, Suite 300, Pasadena, CA 91101
2
3
4
5
6
7
8
Abstract. We present two recent upgrades to the Gemini North Adaptive Optics (AO) system, Altair.
These two upgrades provide 100% sky coverage for low performance AO suitable for improving the
natural seeing by factors of 25% to 3 from blue visible wavelengths (350 nm) through the near infrared
(2.5 micron wavelengths). The first upgrade, dubbed LGS + P1 “Super Seeing” mode, allows correction
of high order aberrations with an on-axis Laser Guide Star (LGS) while tip/tilt correction is performed
with a more distant peripheral wavefront sensor (P1). Most currently operating LGS AO systems are
limited in their sky coverage, primarily due to tip/tilt star availability. Although P1 provides sub-optimal
tip/tilt correction due to its distance from the science source, its patrol radius allows operation in LGS +
P1 mode anywhere in the sky from declinations of +70 degrees to -30 degrees. This mode was offered for
science use at Gemini North in 2013A. We present typical performance and use from its first semester in
science operation, with a factor 2 to 3 image quality improvement over seeing limited images. The second
upgrade is the commissioning of the AO system to correct at visible wavelengths, which is expected to
be completed in 2014. In this mode, Altair will feed the Gemini Multi-Object Spectrograph (GMOS),
which is an optical imager as well as a long-slit, multi-slit and integral field unit spectrograph. We
intend to replace the current Altair science dichroic with a sodium notch filter, passing only the 589nm
wavelength light from the LGS to the AO system. The rest of the spectrum from 350 nm to the GMOS
red cutoff at 1.1 microns is intended as science capable light. Tip/tilt correction will be performed close
to the science target with the GMOS on-instrument wavefront sensor or with P1 as in the P1+LGS mode
discussed above. We expect an image quality improvement of 25% in this mode over seeing limited
observations. Since exposure time to reach a given signal-to-noise ratio scales roughly as the square of
the image quality, these two upgrades represent a substantial efficiency improvement which is available
to nearly all targets normally observed at Gemini North.
a
[email protected] — to whom correspondence should be addressed
Third AO4ELT Conference - Adaptive Optics for Extremely Large Telescopes
1 Introduction
Altair (ALTitude conjugate Adaptive Optics for the InfraRed) is the facility Adaptive Optics
(AO) system used at Gemini North telescope [1]. It uses a 177 actuator Xinetics deformable
mirror and a separate tip-tilt (TT) mirror to correct atmospheric turbulence sensed by a ShackHartmann wave-front sensor (WFS). Since the original commissioning of its Natural Guide
Star (NGS) mode was completed in 2004, it has been upgraded to add Laser Guide Star (LGS)
functionality in 2007 [2]. Since that time, it has been commissioned to feed three facility class
instruments: the Near InfraRed Imager and Spectrometer (NIRI [3]), the Near-Infrared Integral
Field Spectrometer (NIFS [4]) and the Gemini Near-InfraRed Spectrometer (GNIRS [5]). Here,
we describe two further upgrades, the first extends Altair’s sky coverage to 100% of the sky
above elevation 40 degrees in a low-Strehl “Super Seeing” mode called LGS+P1 and the second
allows the use of Altair in the visible with the Gemini Multi-Object Spectrograph (GMOS) [6].
2 LGS+P1
One of the main user requests for Altair since its inception has been enhanced sky coverage. This
is somewhat difficult to accomodate with even moderately high Strehls (∼ 15% in K) without
a major redesign of the Altair optics. Since the system is still in high demand, comprising
roughly 20% of the science programs in the Gemini science queue, this was considered currently
impractical. The LGS+P1 mode was developed as a compromise feature set that could be added
without considerable downtime to Altair operations.
In the typical Altair LGS mode, a ∼ 20 Watt 589 nm laser is projected onto the ∼ 90 km altitude sodium layer of the atmosphere using a small center launch telescope atop Gemini North’s
secondary mirror. All high-order (Zernike modes greater than 4) corrections are measured using
a Shack-Hartmann WFS illuminated by this laser beacon. Defocus and tip/tilt corrections are
measured using an NGS. The advantage of this mode is that since only a few Zernike modes
are measured using an NGS, the NGS star can be much fainter (magnitude R ∼ 18.5) than when
no laser is used (R ∼ 15). The major drawback of this mode is that it requires that the NGS
TT/focus star be fairly close (25 arcsec) to the science target due to the design of Altair’s patrol
field.
The major innovation of the new LGS+P1 mode is that the Gemini facility NGS wavefront
sensor, Peripheral WFS 1 (P1) is used for tip/tilt and focus. The P1 has a patrol field that
extends about 7 arcminutes from the center of the science field. This allows nearly 100% sky
availability for the tip/tilt and focus star as illustrated in Figure 1. The LGS+P1 mode still
retains the fundamental LGS restrictions that only target elevations of 40 degrees above the
horizon are allowed due to physical limitations of the LGS Zoom lens.
Two hardware modifications were needed to allow the use of LGS+P1. The first modification
was the installation of a 589 nm notch filter that removes the Rayleigh light from the P1 WFS.
Without this notch filter, the P1 WFS was flooded with scattered light from the LGS and guiding
was not possible except on very faint stars. The second hardware change was the replacement
of the previous 6x6 Shack-Hartmann WFS with a 2x2 one. This allowed the use of P1 on
considerably fainter stars and was sufficient for our purposes since it only senses tip/tilt and
focus. There were considerable software changes that were required for this mode of use. The
main change was the ability for P1 to send its measured tip/tilt to the Gemini secondary mirror
(M2) and its measured focus to the Zoom lens (which corrects for sodium layer altitude) in front
2
Third AO4ELT Conference - Adaptive Optics for Extremely Large Telescopes
Fig. 1. The original Altair LGS system sky coverage as a function of galactic latitude using the onboard
STRAP NGS WFS (black) and LGS+P1 sky coverage (red). The LGS+P1 mode offers greatly enhanced
sky coverage, but as described in the text, reduced image quality. For targets near the galactic plane,
the existing system is preferred due to high NGS star availability and increased performance. Near the
galactic pole, many more targets are available using the LGS+P1 mode.
of the Altair LGS WFS. We also tested P1 sending the tip/tilt corrections directly to the Altair
tip/tilt mirror. However, loop functionality was not stable in this configuration because it was
not truly a closed loop feedback system. P1 is before Altair in the optical light path, and thus if
P1 feeds the Altair tip/tilt mirror, corrections are sent by P1 that are applied to the science beam
but are not immediately seen by P1. Repeated tip/tilt corrections applied to the Altair tip/tilt
mirror are eventually offloaded to M2, but the bandwidth of this offload was not great enough
to create a stable loop system using Altair+P1. Therefore, all P1 tip/tilt corrections are applied
directly to M2.
Although the LGS+P1 mode offers nearly all-sky availability, there is significant performance degradation over the original LGS mode using Altair’s onboard avalanche photo diode
based tip/tilt sensor (STRAP). This is primarily due to the effect of tilt-anisoplanatism, where
the tip/tilt produced by the atmosphere isoplanatic patch near the P1 NGS star is different from
that of the science target due to the large (5 – 7 arcmin) distance between the science and P1
targets. Although generally the LGS+STRAP will produce better image quality, this is only if a
star is available near (25 arcsec) the science target. In the absence of a STRAP star, the LGS+P1
3
Third AO4ELT Conference - Adaptive Optics for Extremely Large Telescopes
Fig. 2. LGS P1 on-axis performance using NIRI in the f/32 and f/14 plate scales. Natural seeing is shown
on the x-axis and the corrected FWHM is shown on the y-axis. The image quality is improved by a factor
of 2 – 3. Strehl is quite low (a few percent) in this mode, so image quality is much better described with
typical seeing-limited Gaussian/Lorentzian formulae.
provides superior performance to non-AO tip/tilt correction that is normally employed in the
seeing limited case, as shown in Figure 2. The typical image quality improvement is a factor of
2 – 3 in full width at half maximum (FWHM). Since telescope integration time is inversely proportional to the square of the image FWHM, this represents a substantial reduction in exposure
time of a factor 4 – 9. LGS+P1 has the ability to be used for nearly all targets, so we expect that
it will be a often-used mode of LGS.
4
Third AO4ELT Conference - Adaptive Optics for Extremely Large Telescopes
3 Altair+GMOS
The GMOS instrument is traditionally the highest-demand instrument in the Gemini instrument
suite. The Altair+GMOS upgrade is designed to provide Altair’s corrected beam to GMOS.
Currently, the Altair dichroic sends light with wavelengths less than 850 nm to the Altair WFS
where it is used to estimate image quality aberrations from the atmosphere. Light with wavelengths greater than 850 nm (out to about 2.5 microns) is available for science. Since GMOS
is a visible-light instrument, it has measurable sensitivity in the 350 nm – 1.1 micron range,
meaning that the current Altair dichroic does not allow science use of most of the optical range
(i.e. 350 nm – 850 nm). Thus, the main hardware change to make the Altair+GMOS mode
useful is changing the dichroic. Since the LGS+P1 mode described above only requires light
from the 589 nm sodium D doublet, we plan to install a “notch” dichroic that sends only the
589 nm light to Altair, with the remaining light in the 350 nm – 1.1 micron range going to
GMOS for science. Tip/tilt and focus will then be done with either P1 (as in the LGS+P1 case)
or using the GMOS On-Instrument WaveFront Sensor (OIWFS) which is located just above
the GMOS focal plane. Altair has the ability to select between 2 dichroics, so the user will be
able to select between the 589 nm notch dichroic for GMOS+LGS use and the existing dichroic
for GMOS+NGS use if the science wavelength of interest is greater than 850 nm. A schematic
of the new dichroics appears in Figure 3, where the 589 nm notch is pictured along with an
upgrade to the existing Altair dichroic which will extend the existing science wavelength range
from 850 nm - 2.5 microns by adding a 3 micron – 5 micron capability. This capability is useful
for studying exoplanets about bright host stars.
We have performed on-sky tests of the Altair+GMOS mode using our existing 850 nm
dichroic. We were only able to test the Z (830 nm – 925 nm) and Y (970 nm – 1070 nm)
bands since these are long enough to have sensitivity longward of 850 nm. Results are shown in
Figure 4. The improvement is much greater for the Y band (∼ 40%) than the Z-band (∼ 20%)
most likely due to the shorter wavelength. However, the images were not taken simultaneously
and the seeing did degrade after the Z-band images were taken, so it is possible that some of the
Altair/GMOS
Current Dichroic
<850 nm to Altair
850 nm - 2.5 µm to Science
Upgrades
ZYIJKLM Dichroic
<850 nm to Altair
850 nm - 2.5 µm to Science
3 µm - 5 µm to Science
griZY Notch
589 nm to Altair
350 nm - 1.1 µm to Science
Fig. 3. A schematic of the existing Altair dichroic (top) and the two upgraded dichroics (bottom) which
will allow L and M band use and GMOS+LGS notch.
5
Third AO4ELT Conference - Adaptive Optics for Extremely Large Telescopes
Fig. 4. GMOS with and without Altair for the Y-band (left) and the Z-band (right) as a function of
distance from the AO star. The Y-band shows a ∼ 40% improvement in FWHM even relatively far from
the AO star. The Z-band, a shorter wavelength, shows only a ∼ 20% improvement.
difference is due to the characteristics of the atmosphere at the time the test were performed.
We note that, as expected, the reddest wavelengths show the best improvement. This suggests
that the Altair+GMOS mode will be of greatest use when GMOS is upgraded with red-sensitive
Hamamatsu CCDs in the near future. These CCDs will allow science to be done at useful sensitivities even beyond 1 micron.
Since we cannot test the on-sky performance of Altair+GMOS at wavelengths less than
850 nm, we simulated the performance. The simulations were performed using the Yorick yao
simulation package and were calibrated by reproducing the existing performance on NIRI and
the on-sky tests with GMOS. The results of the simulation appear in Table 1. Even the shortest
wavelengths show improvements of 17% in FWHM which corresponds to gains in telescope
efficiency of ∼ 35% since telescope exposure time is inversely proportional to the square of the
seeing.
Table 1. Yorick yao simulations of Altair+GMOS performance. We note that although image quality
improvement may appear modest, telescope exposure time is a function of the inverse square of FWHM,
suggesting that even the minimum improvement of 17% can lead to gains in efficiency of over 35%.
Filter Name
Wavelength range [nm]
Seeing Limited FWHM [arcsec]
500 Hz LGS/P1 [arcsec]
Image Quality Improvement [%]
Efficiency Gain [%]
g
398 – 552
0.76
0.63
17
37
r
562 – 698
0.73
0.59
19
42
6
i
706 – 850
0.70
0.55
21
46
Z
830 – 925
0.68
0.53
22
49
Y
970 – 1070
0.67
0.50
25
56
Third AO4ELT Conference - Adaptive Optics for Extremely Large Telescopes
We outline the basic modes of use for Altair+GMOS in Table 2. This table includes the
expected demands for the various modes given typical performance, sky coverage and science
wavelength.
Table 2. Expected modes of use for Altair+GMOS
Name
GMOS/LGS/P1/notch
GMOS/LGS/P1
GMOS/LGS
GMOS/LGS/OIWFS/notch
GMOS/NGS
WFS
LGS
LGS
LGS
LGS
NGS
Demand
Very High
High
Medium
Medium
Low
Sky Coverage
100%
100%
15%
15%
2%
Science Wavelength
450 nm – 1.2 nm
850 nm – 2.5 nm
850 nm – 2.5 nm
450 nm – 1.2 nm
450 nm – 1.2 nm
4 Summary
In summary, we describe two upgrades to the Altair system to enhance science utility. The first
is the LGS+P1 mode which allows full sky coverage using LGS. Performance improvements
are modest, but are still enough to improve image quality from the 70 percentile seeing limited
case down to the 20 percentile case. This should provide substantial improvement in signal to
noise as a function of exposure time for the near-IR instruments NIRI, NIFS and GNIRS.
In the visible, we plan to extend the capabilities of Altair to be able to feed GMOS, the facility instrument covering the visible. The main change to Altair will be the installation of a 589
nm notch dichroic that will send sodium laser light to the Altair WFS for high order correction.
Tip/tilt and focus correction will be done as with LGS+P1 or with the GMOS OIWFS. Simulations and on sky tests suggest a useful improvement in image quality with this mode of use.
The usefulness of this mode will be even greater when GMOS is upgraded with red-sensitive
Hamamatsu CCDs in the near future.
References
1. S. Roberts, and G. Singh, SPIE Conference Series, 3353, (1998), 611 – 620
2. M. Boccas, F. Rigaut, M. Bec, B. Irarrazaval, E. James, A. Ebbers, C. d’Orgeville, K. Grace,
G. Arriagada, S. Karewicz, M. Sheehan, J. White, and S. Chan, SPIE Conference Series,
6272, (2006), 62723L
3. K. Hodapp, J. Jensen, E. Irwin, H. Yamada, R. Chung, K. Fletcher, L. Robertson, J. Hora,
D. Simons, W. Mays, R. Nolan, M. Bec, M. Merrill, and A. Fowler, PASP, 115, (2003) 1388
4. P. McGregor, J. Hart, P. Conroy, L. Pfitzner, G. Bloxham, D. Jones, M. Downing, M. Dawson, P. Young, M. Jarnyk and J. van Harmelen, 4841, (2002), 178
5. J. H. Elias, B. Rodgers, R. R. Joyce, M. Lazo, G. Doppmann, C. Winge, and A. RodriguezArdila, SPIE, 6269, (2006), 36
6. I. Hook, I. Jørgensen, J. R. Allington-Smith, R. L. Davies, N. Metcalfe, R. G. Murowinski,
and D. Crampton, PASP 116, (2004), 425–440
7