Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.arcetri.astro.it/~lt/preprints/nics/nics.ps.gz Äàòà èçìåíåíèÿ: Tue Sep 11 16:56:54 2007 Äàòà èíäåêñèðîâàíèÿ: Sat Dec 22 06:53:35 2007 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: optical telescope

Astronomy & Astrophysics manuscript no.
(will be inserted by hand later)
NICS: The TNG Near Infrared Camera Spectrometer ?
C. Baffa 1 , G. Comoretto 1 , S. Gennari 1 , F. Lisi 1 , E. Oliva 1;4 , V. Biliotti 1 , A. Checcucci 1 , V. Gavrioussev 2 ,
E. Giani 1 , F. Ghinassi 4 , L.K. Hunt 2 , R. Maiolino 1 , F. Mannucci 2 , G. Marcucci 3 , M. Sozzi 2 , P. Stefanini 1 ,
and L.Testi 1
1 Osservatorio Astrofisico di Arcetri, Largo E. Fermi 5, Firenze, Italy
2 Centro per l'Astronomia Infrarossa e lo Studio del Mezzo Interstellare--CNR, Largo E. Fermi 5, Firenze, Italy
3 Dipartimento di Astronomia e Scienze dello Spazio, Universit`a di Firenze, Largo E. Fermi 5, Firenze, Italy
4 Centro Galileo Galilei, Santa Cruz De La Palma, TF, La Palma, Spain.
Received..........; Accepted..........
Abstract. NICS (acronym for Near Infrared Camera Spectrometer) is the near­infrared cooled camera--
spectrometer that has been developed by the Arcetri Infrared Group at the Arcetri Astrophysical Observatory, in
collaboration with the CAISMI­CNR for the TNG (the Italian National Telescope Galileo at La Palma, Canary
Islands, Spain). As NICS is in its scientific commissioning phase, we report its observing capabilities in the
near--infrared bands at the TNG, along with the measured performance and the limiting magnitudes. We also
describe some technical details of the project, such as cryogenics, mechanics, and the system which executes data
acquisition and control, along with the related software.
Key words. Instrumentation: spectrographs, Instrumentation: polarimeters, Near Infrared
1. Introduction
The 3.5m Italian National Telescope Galileo (TNG)
(Barbieri 1995), under operation on La Palma (Canary
Islands), included a general purpose near--infrared cam­
era/spectrometer in its Instrument Plan for first--light op­
erations (Fusi Pecci et al. 1992, Fusi Pecci et al. 1994).
The telescope itself is not optimized for the thermal in­
frared bands, so NICS was designed to operate at near--
infrared wavelengths, from 0.95 ¯m up to 2.50 ¯m, avoid­
ing the spectral range where the ambient blackbody ra­
diation could degrade the signal--to--noise ratio of the ob­
servations. Moreover, this spectral range is conveniently
covered by HgCdTe (or MCT for Mercury--Cadmium--
Telluride) large format focal--plane array detectors cur­
rently available, which, at present, offer the best combi­
nation of quality and low read­noise. NICS was the only
infrared instrument for the first light of the TNG; as a con­
sequence, we decided to incorporate a sufficient degree of
operation flexibility by adopting a collimator/camera opti­
cal scheme along with a good pupil image where a number
of analyzers (filters, grisms, polarizers) can be easily ac­
commodated. This configuration allows a large number of
photometric and spectroscopic observing modes, and the
observer can switch rapidly between different modes in re­
Send offprint requests to: C. Baffa
? Based on observations taken at TNG, La Palma.
mote operation, for instance, adapting the observations to
the seeing conditions of the night.
2. Observing modes
The instrument is provided with the following imaging and
spectroscopic observing modes:
-- wide­field imaging with a plate scale of 0:25 00 /pixel and
a total field, as projected on the sky, of more than
4 0 \Theta 4 0 ; wide-- and narrow--band filters are available for
photometry and in--line imaging;
-- small­field imaging with a plate scale of 0:13 00 /pixel
(¸ 2 0 \Theta 2 0 field of view), for better sampling under
excellent seeing conditions;
-- medium­ to low­dispersion long--slit (4 0 slit) grism spec­
troscopy with a resolving power between 300 and 1300;
-- very low­dispersion long--slit (4 0 slit) spectroscopy with
a resolving power ú 50, by means of an Amici prism;
-- imaging polarimetry for both wide-- and small--field
imaging mode. Polarimetry imaging is performed on
only 1/4 of the field of view, but simultaneously on
four directions of polarization angle (0, 45, 90, and
135 degrees), with a clear gain of relative sensitivity.
-- spectro--polarimetry with a reduced (25%) slit length,
but with four directions of polarization angle (0, 45,
90, and 135 degrees) measured simultaneously.

2 C. Baffa et al.: NICS: The TNG Near Infrared Camera Spectrometer
Fig. 1. Spectral response and efficiency of NICS wide--band
filters.
Three of these modes, low­dispersion long--slit spec­
troscopy and simultaneous angle polarimetry/spectro--
polarimetry are unique to NICS, and provide for the first
time the possibility to perform reliably these measure­
ments in the infrared.
The simultaneous mode polarimetry is obtained by
means of wedged double Wollaston prisms and a special
field stop (Oliva 1997). High spatial resolution imaging at
the diffraction limit of the telescope is possible by means
of an external adaptive optics module (Ragazzoni 1996),
using the re­imaging optics of the adaptive module, which
has an f/33 output beam. Two reduced plate scales (about
0:08 00 and 0:04 00 per pixel, respectively) are available in
connection with the wide field and the small­field optics
of NICS.
Figure 1 shows the spectral response and the efficiency
of the wide band filters which are presently available; be­
sides the standard filters for J, H, and K bands, NICS
offers the 1¯m filter centered at 1.030 ¯m in correspon­
dence with a fair atmospheric window, the Jn filter as de­
fined by the Gemini project, the K' filter which cuts the K
band portion where the thermal emission dominates the
background flux. There is also a band pass filter (labeled
SW) for general purpose observations and pointing and a
high--pass filter (labeled LW) which, along with the spec­
tral response of the detector, acts as band pass for longer
wavelengths.
The narrow--band filter set includes Br fl and Fe II fil­
ters, plus the associated K and H narrow­band filters to
sample the near--by continuum. It is possible to insert in
the beam several accurate grey filters to reduce the flux
from very bright sources. Figure 2 shows the resolving
power (when associated with the 1 00 slit) and the efficiency
of the available grisms which are resin--replicated Milton­
Roy gratings on IRGn6 prisms (Vitali et al., 2000).
Fig. 2. Resolving power and efficiency of NICS grism.
3. General Description
The optical scheme comprises the single collimator and
the two cameras; all the optical components reside in a
vacuum at a temperature of about 80 K, inside a suitable
cryostat. The focal plane masks (field stops and slits) are
mounted on a wheel. Immediately after the focal plane, a
further removable field stop makes possible polarimetric
measurements.
The collimator is an achromatic doublet lens, which
images the entrance pupil of the TNG and provides a par­
allel beam at the pupil plane where Lyot stops can be
placed. Immediately after the pupil plane, two adjacent
wheels carry filters and grisms. Two interchangeable op­
tical systems relay the image of the focal plane on the
detector with the desired magnification.
A third optical system images the entrance pupil on
the detector, for the purpose of an accurate alignment of
the telescope with the instrument pupil, and is not nor­
mally used in routine observations. A short discussion can
be found in Gennari et al. (1995).
The spectroscopic mode makes use of the wide--field
camera by inserting one of the grisms, located on the sec­
ond filter wheel, into the parallel beam after the pupil.
The rejection of stray and thermal light is left to the
TNG baffles in J, H, and K' bands. At longer wavelengths,
where the thermal radiation could prevail, it is possible to
insert a cold stop precisely positioned at the pupil image.
The sensitive element of NICS is the Hawaii 1024 \Theta
1024 pixels HgCdTe array detector (Rockwell Science
Center). It has a 18.5 ¯m/pixel pitch and is sensitive to ra­
diation at wavelengths between ¸ 0:90¯m and ¸ 2:5¯m.
Its performance in term of dark current, efficiency, and
read noise, is comparable or better than the 256 \Theta 256
NICMOS 3 (e.g. Lisi et al. 1996).
The electronic noise is dominated by the detector and
by the first cold amplifier and is ¸ 25 e \Gamma , if a suitable num­
ber of detector resets (more than 32) is performed at each

C. Baffa et al.: NICS: The TNG Near Infrared Camera Spectrometer 3
wedges
Pupil
image
Wollastons
Pupil
image
q=45
q=135
q=0
q=90
e
o
e
o
Optical axis
q
Wedges
Wollaston down Wollaston up
Input field
Output images
Fig. 3. Logical sketch of polarimetry operations. (Adapted
from Oliva, 1997, but with real data.)
integration. In the most common read strategy (double
sampling), during the integration at least two measure­
ments are performed, one at the beginning (to sample the
reset bias level), the second at the end of the integration
ramp.
NICS is mounted at the same focal station as the op­
tical CCD camera; the two instruments share an adapter
that carries also the adaptive optics module. A plane mir­
ror (M4), mounted on a remotely--operated sliding bench,
deflects the beam from the telescope to the entrance of the
IR camera, that has its optical axis perpendicular respect
to the telescope beam. NICS is mounted on the adapter by
means of a set of spherical joints that allows for a limited
adjustment of the optical axis alignment respect to the
entrance beam. The fine alignment is handled by the M4
mounting hardware, which is designed to allow for small
adjustments in matching the telescope optical beam to the
camera optical axis.
4. Optical Design
The collimator used for all the scales and the observing
modes is a detached doublet (BaF 2 --IRG2) (Oliva and
Gennari 1998) with spherical­surface lenses that trans­
form the f/11 TNG beam into a 22 mm parallel beam
(f/ larger than 2000) with a total level of aberrations be­
low 0.4 mrad. The blur of the entrance pupil image has a
maximum size of 0.21 mm at 90% encircled energy, which
guarantees an accurate background rejection when the
Lyot stop is inserted. The aberrations of the pupil im­
age have been reduced by a factor 2 by means of a mildly
aspheric entrance window, which is designed to leave the
pupil plane position unchanged.
The two optical trains for the cameras and the pupil
re--imaging system are mounted on a wheel, accurately
driven by an external motor. These optics are based on
detached doublets (BaF 2 --IRG2) followed by one or two
lenses made of the same materials.
The wide­field camera (0.25''/pix) consists of four
lenses. The spot blur on the detector is good (90% of the
energy within a pixel inside a ú 120 00 radius circle, 75%
in the corner). The exit pupil is positioned 88 mm behind
the focal plane of the camera, which renders the system
not completely telecentric; this introduces some cos 4 losses
(around 4:5% at the corners of the field). In addition, the
system suffers a moderate distortion of the order of few
percent at the field corners.
The small­field camera (0.13''/pix) consists of three
spherical lenses. Its overall image quality is much better
than that of the wide­field camera: the spot blur is smaller,
distortion is basically absent (! 0:16%), and the exit pupil
is at about ­350 mm, which makes the system almost tele­
centric.
The filters are mounted on two wheels located in the
parallel beam just after the pupil plane; the second wheel
carries also the grisms, slightly tilted with respect to the
optical axis to achieve their best efficiency. The second fil­
ter wheel also hosts the wedged double Wollaston prism,
made of LiYF 4 , that deviates the rays of the parallel beam
into four different beams corresponding to the polarization
angles of 0, 45, 90 and 135 degrees; as a consequence, the
input field of view is split into four images on the de­
tector, one for each of the four polarization angles (Oliva,
1997). This way, one can perform simultaneous photomet­
ric measurements of the polarized flux at different angles
as required to derive the first three elements of the Stokes
vector (see figure 3).
For polarimetry, a suitable field stop right after the fo­
cal plane limits the field of view to about 1/4 of the total
field, to avoid overlapping of the four polarized images on
the detector plane. The second wedged double Wollaston
prism is made of LiNbO 3 . This prism is designed to per­
form spectro­polarimetry when associated with one of the
available grisms. The polarizer works with the same prin­
ciple as the imaging mode, simultaneously delivering four
polarized long­slit spectra at angles rotated by 45 degrees.
The 4 0 slit normally used in the spectroscopic mode will
be masked to have an equivalent length of about 50 00 , to
avoid any overlapping of the four polarized spectra.
In order to allow focusing capabilities, the detector is
mounted on a motorized base. For each optical configura­
tion, the exact detector position has been measured and
the control software automatically performs the internal
focus setting.
5. Cryogenics and Mechanics
The cryo--mechanical design is based on a cylindrical vac­
uum shell, with a size dictated by the total length of the

4 C. Baffa et al.: NICS: The TNG Near Infrared Camera Spectrometer
Fig. 4. Mechanical Layout of Nics cryostat.
optical system and the wheels diameter. This shape has
the advantage of low weight and high stiffness, associated
with a reasonable cost.
The cylindrical case, with its symmetry axis orthog­
onal to the TNG exit beam, is rigidly connected to the
Nasmyth adapter flange and supports the internal cooled
structure via thermally isolating, rigid elements. The radi­
ation shield is connected to the cold structure; the volume
between the pupil plane and the detector is protected from
stray light by means of a second radiation shield. The in­
ternal structure supports all the optical functions, that
is the four wheels (mounted on ball bearings), the col­
limator, and the detector, keeping them at the required
relative positions. Several external micro--stepping motors
take care of wheels and slides positioning, as required by
the selected observing mode.
Operations at the de­rotated Nasmyth focus necessi­
tate the rotation of the instrument itself, making imprac­
tical the use of liquid cryogenics for cooling and main­
taining the operative temperature; instead, we based the
cryogenic system on a closed cycle cooler.
Because of the relatively large mass of the instru­
ment, cooling it down to the operating temperature is best
achieved by means of a continuous flow of liquid nitrogen
inside a pipe welded to the optical bench; during normal
operation the liquid nitrogen is not necessary. The detec­
tor, in good thermal contact with the cold plate, has an
active temperature control. An activated charcoal cryo--
adsorption pump is installed on the cold surface to guar­
antee a good vacuum in the presence of out­gassing or
small leaks for periods greater than 90 days. For main­
tenance purposes, the molecular sieve is equipped with a
suitable heating element and thermal switch.
All measurements of temperature and pressure inside
the cryostat are transmitted to the control computer,
where the software can look after the regularity of op­
erational parameters at scheduled intervals of time.
6. Electronics
The data acquisition and control system is based on the
controller developed by the CCD Working Group of TNG,
suitably modified to adapt it to the architecture of infrared
arrays (Comoretto et al. 1995, Comoretto et al. 1999). The
controller is based on a set of Transputer processors, which
are responsible for handling data and sending commands,
and on a DSP Motorola 56001, which generates the syn­
chronized clock pattern needed for accessing and reading
the array multiplexer.
The analog signal read on each pixel of the four quad­
rants is buffered by four FET amplifiers located on the
same board that hosts the detector. After that stage, there
is a set of four 16­bit A/D converters, which converts the
pixel intensity of the four quadrants in parallel. The paral­
lel outputs from the converters are translated to the trans­
puter serial protocol using a dedicated programmable logic
chip from Xilinx.
The transputer stage sends the digital data to a Linux
PC by means of a fiber link which exploit the fast se­
rial connection capability built into each Transputer. The
controller takes care of telemetry and stepper motors by
means of dedicated RS--232 serial ports.
7. Software
The low--level software involves relatively complex inter­
actions between the Motorola DSP and the cluster of
Transputers modules, which operate in a multitasking con­
figuration, each performing an elementary task in parallel.
The transputer network is organized as a linear chain with
the possibility of single node branches (we call them left
branches), as described above.
Due to the intrinsic multitasking nature of transputers,
we organized low level software as a collection of modules,
each performing an elementary task, all acting in parallel.
Complex tasks (as data acquisition, handling and trans­
fers) are realized by the cooperation of many modules,
often running on different CPUs.
One of the biggest issues we faced in developing trans­
puter software was the inter­processes communication. We
developed a simple packet switching solution, in a way
that roughly resembles the IP protocol (Internet way).
All communications are performed by means of fixed--size
packets. Each packet starts with a header stating the node
of origin, the destination node and destination process,

C. Baffa et al.: NICS: The TNG Near Infrared Camera Spectrometer 5
and the command and sub­command(s). The packet has
also a large (1024 short integers) data area.
Each node has a process which examines the header of
each packet and then dispatches it to the three possible
outward directions or to the destination process for execu­
tion. The modified linear chain enables us to make all rout­
ing by means of only local fast comparisons, which makes
the system very efficient and suitable also for data dis­
patching. For further details on the architecture of the in­
terprocessor communication, see Baffa et al. (1999), Baffa,
(2000), and references therein.
The high--level control program, which comprise the
telescope interaction, the data handling and storing and
the human interface of NICS (Xnics) is based on a simi­
lar interface developed several years ago for the Arnica IR
camera. Xnics provides the observer with the environment
to define the parameters of measurements and to start
the desired sequence of integrations (such as single frame,
multiple frames, mosaics, scans, scripting language), along
with sky--source subtraction and preliminary reduction for
quick--look purposes. Several tools are available to con­
trol the overall quality of data during the measurement,
while the program is in charge of validating the param­
eters which the observer has chosen before starting the
observation. In the background, a task is always active
which monitors messages and error flags coming from the
low--level software.
All the low level handling is performed by a concurrent
program, NICSgate. It consists of several object­oriented
processes, each one controlling a special portion of the
hardware functionality: telescope, motors, Transnix ini­
tialization and programming, acquisition as a defined task
and all types of communication in real time. Each process
maintains its inner state and can be activated at any time,
when an external event needs its special functions.
Due to the intrinsic network awareness of both Xnics
and NICSgate, a distributed execution of the software
is possible: NICSgate on the local acquisition computer,
Xnics on a remote one.
The software developed for this instrument is ``layer
organized'', that is to say organized as a stack of many
layers of subroutines of similar levels of complexity. To
accomplish its task, each routine needs relies only on the
immediately adjacent level and on global utility packages.
Such a structure greatly simplifies the development and
maintenance of the software.
Our efforts were aimed at several different require­
ments. Our first priority was to have a flexible labora­
tory and telescope engine, capable of acquiring easily the
large quantity of data a panoramic IR array can produce.
Another main goal was to produce an easy­to­use soft­
ware with the smallest ``learning curve''. Our idea was
that data acquisition must disappear from observer atten­
tion, giving him/her the possibility to concentrate on the
details of the observations; in this way, observing efficiency
is much higher. The human interface is realized through a
fast menu interface. The operator is presented only with
the options which are currently selectable, and the menu
Filter JS H K'
Zero point 22.1 22.3 21.8
Efficiency 0.21 0.28 0.32
Lim. mag. 22.5 21.4 21.2
Average sky 15.5 13.2 13.2
mag/arcsec 2
Table 1. Zero points, efficiencies and limiting magnitudes of
NICS at the TNG measured in October­December 2000. Lim.
Mag. is the point source limiting magnitude for a 3oe detec­
tion in a hour of on­source integration with a seeing of 1 00
(integration aperture of 2 00 ) in LF mode. Average sky is the
approximate sky brightness, in magnitudes/arcsec 2 of the var­
ious measurements.
is rearranged on the basis of user choices or operations.
We have also implemented automatic procedures such as
multi--position (``mosaic''), multi--exposure (stack of many
frames) and a scripting language capable of performing a
fairly complex set of measurements with only a ``quality
control'' from the observer.
8. Characterization and performance
NICS was tested and characterized both in the laboratory
and, at the moment of writing, during four commissioning
runs at the TNG.
For what concerns the detector, the current read­out
noise (double sampling) is about 25 e \Gamma , but is expected to
improve once the forthcoming multi­sampling mode will
be available. The dark current is about 1 e \Gamma /sec and the
well capacity is about 10 5 e \Gamma . With the current electronics
setup the conversion factor is about 8 e \Gamma /ADU.
The efficiency of the instrument was measured in the
laboratory and at the telescope. In the laboratory, 2¯m
efficiency was measured by means of a blackbody located
in front of the window. The efficiency in imaging mode is
about 60%, in agreement with what is expected from the
combination of the efficiencies of the detector and of the
filters. The overall efficiency of the instrument and tele­
scope (and atmosphere) was measured during the commis­
sioning runs. The total efficiency critically depends on the
cleanness of all optical surfaces, which, for various reasons
could be less than perfect. The efficiencies and sensitiv­
ities given in the following, refer mostly to the last two
commissioning runs, after the cleaning of M3 (the relay
mirror to the Nasmyth focus), but without cleaning of the
entrance window. The final assessment of NICS' perfor­
mance at the TNG must await the cleaning of all optical
surfaces. In Table 1 we give the zero points in the three
main broad bands and the corresponding (total) efficien­
cies. We report also the limiting magnitudes for a detec­
tion at 3oe in a hour of on­source integration with a seeing
of 1 00 (integration aperture of 2 00 ).
The efficiency in the spectroscopic modes agrees with
the efficiencies given in Table 1 convolved with the re­
sponse curves given in Fig. 2. Particularly interesting is
the efficiency measured through the Amici prism, since

6 C. Baffa et al.: NICS: The TNG Near Infrared Camera Spectrometer
Fig. 5. Global efficiency of the system (instrument + telescope
+ atmosphere) measured through the Amici prism.
Disperser Central Resolution Limiting Magnitude
Amici 50 19.8
JK' 350 18.4
IJ 500 18.6
JH 500 18.3
HK 500 18.1
I 1250 17.8
J 1200 17.6
H 1150 17.6
K 1250 17.2
Table 2. Limiting magnitudes of NICS at the TNG in dif­
ferent spectroscopic mode. Limiting Magnitude is the point
source limiting magnitude for a 3oe detection in a hour of on­
source integration with a seeing of 1 00 and with a 1 00 slit in LF
mode, averaged on the wavelength range. The sky luminosity
is assumed to be the same as in Table 1.
the latter disperser has a nearly flat efficiency over the
whole near­IR range and, therefore, gives a global view of
the efficiency of the system. Fig. 5 gives the (absolute) effi­
ciency obtained through the Amici and, more specifically,
by dividing the Amici spectrum of a standard star by the
intrinsic spectrum of the star (therefore, this is global ef­
ficiency of the Amici prism + instrument + telescope +
atmosphere). Note the main atmospheric windows which
are marked in the figure. Also worth noting is the efficiency
drop in the J and 1¯m bands which can be ascribed to a
drop in efficiency of the detector at short wavelengths.
A broad estimate of the limiting magnitude in po­
larimetry can be derived from broad--band values by sub­
tracting 1.5 magnitudes (pupil light is divided in four
parts). In Table 2 we give preliminary values for spec­
troscopic limiting magnitudes using large fields, one hour
of exposure, and with a 1 00 slit.
As mentioned in Section 4, the wide­field camera op­
tics suffer from distortion at the level of a few percent at
the array vertices. During commissioning, this distortion
was characterized at the telescope by measuring crowded
stellar fields with known astrometry. It turns out that the
distortion can be well approximated by a symmetric ra­
dial sixth­order polynomial, and the coefficients for the
forward and inverse transformations were derived from as­
trometric measurements. Results show that, during com­
missioning, the optical center of the array is within one
pixel of the center of symmetry of the distortion, and the
amplitude of the measured distortion is consistent with,
perhaps slightly smaller than, the design specifications.
9. Conclusions
After several months of testing at TNG, NICS proved able
to provide the entire set of observing modes included in
the design with the desired performance. At present, the
TNG is equipped with a near­IR facility well suited to
a 3.5m­class telescope, ready to serve the astronomical
community.
Two of the available observing modes, polarimetry and
low resolution simultaneous 0.9--2.5 ¯m spectroscopy are
unique to NICS and make the TNG + NICS system the
only facility available for this kind of observations.
Acknowledgements. Most of the instrumentation projects for
astronomy can be successfully achieved only with the help of a
team of skilled engineers and astronomers. The authors would
like to emphasize that the NICS project received (and is still re­
ceiving) support, help and technical input from several people;
they thank in particular A. Marconi, L. Miglietta, G., Tofani,
F. Fusi Pecci, L. Corcione, G. Nicolini, J. Licandro, the TNG
development team, the CCD Group, and many people from
Arcetri and from the astronomical community for useful dis­
cussions. The authors would like to particularly acknowledge
the support coming from the TNG staff, always ready to help
during the difficult and busy time of the commissioning at the
telescope.
References
Barbieri C., TNG Technical Report n. 53, November 1995
Baffa C., and Comoretto G., 1996, Arcetri Technical
Report N.3/96
Baffa C., Comoretto G., Gavrioussev, V., Giani, E., Lisi
F., 1999, Mem. Soc. Astron. It. in press
Baffa C., 2000, Arcetri Technical Report N.2/2000
Comoretto G., Baffa, C. Gavrioussev, V., Giani, E., Lisi
F., 1999, Mem. Soc. Astron. It. in press
Comoretto G., Baffa C., and Lisi F., 1995, Arcetri
Technical Report N.4/95
Fusi Pecci F., Stirpe G.M., TNG Instrument Plan, 1992,
Osservatorio Astronomico di Bologna.
Fusi Pecci F., Stirpe G.M., Zitelli V., TNG Instrument
Plan:II. A Progress Report, 1994, Osservatorio
Astronomico di Bologna.

C. Baffa et al.: NICS: The TNG Near Infrared Camera Spectrometer 7
Lisi F., Baffa C., Biliotti V., Bonaccini D., Del Vecchio C.,
Gennari S., Hunt L.K., Marcucci G., Stanga S., 1996,
PASP, 108, p364.
Gennari S., Vanzi, L., Lisi F., 1995, SPIE, 2475, 221.
Maiolino R., Rieke G.H., Rieke M.J., 1996, AJ, 111, 537
Oliva E., 1997, A&AS, 123, 589
Oliva E., Gennari S., 1998, A&AS, 128, 5890
Oliva E., Origlia L. 1992, A&A, 254, 466
Ragazzoni R. (ed.), AdOpt Yearly Status Report, 1996,
Osservatorio Astronomico di Padova
Vitali, F., Cianci, E., Lorenzetti, D., Foglietti, V.,
Notargiacomo, A., Giovine, E., Oliva, E., 2000, SPIE,
4008, 1383.