TripleSpec is a cross-dispersed near-infrared spectrograph that provides simultaneous continuous wavelength coverage from 0.95-2.46um in five spectral orders. The primary configuration of the instrument delivers a spectral resolution of R=3500 in a 1.1 arcsecond slit at 2.1 pixels per slit on the spectrograph array. Slits with 0.7", 1.5", and 1.7" are also available. The instrument contains two independent infrared arrays. One provides a 2048x1024 pixel view of the cross dispersed spectrum. The second provides a 1024x1024 view of a 4'x4' region of the sky, including the spectrograph entrance slit, in the Ks (2.16um) band.
| Slits | 0.7x43" (120 microns projected width) 1.1x43" (186 microns) 1.5x43" (261 microns) 1.7x43" (290 microns) |
| Spectral Coverage | 0.95-2.46um |
| Spectral Resolution | 5000(?TBD) for 0.7" slit (undersampled and limited by optical performance) 3500 for 1.1" slit (2.1 spectral pixels per slit) 2800 for 1.5" slit 2500 for 1.7" slit |
| Gain | 3.5 e-/ADU +/- 20% |
| Read Noise | 18 electrons / sqrt(Nfowler) |
| Dark Current | 0.05 e-/s |
| Well Depth | 40000 DN = 140,000 electrons |
| Minimum Integration time | (Nfowler * 0.8 + 0.3) sec on sky (2*Nfowler * 0.8 +0.3) sec to estimate saturation |
| Spectrograph saturation magnitude | 4th (defocus for brighter objects) |
| Background limited exposure time | ~200+ sec |
| Slit viewer/ guider pixel scale (unbinned) | 0.245" /
pixel (175 pixel slit length) |
| Spectrograph spatial pixel scale | 0.39"/pixel (110 pixel slit length) |
| Guider/slitviewer bandpass | Ks only - fixed filter |
| Faintest practical source for acquisition in the slit viewer | Ks ~ 17 |
| Spectrograph continuum sensitivity | J, H, Ks = 17.0, 16.0, 15.5 5-sigma in one hour with 3 pixel spectral smoothing. 0.003 Janskys (using 1.1" slit in good seeing) |
 |
 |
 |
| TripleSpec mounted at the NA2 focus | A spectrum of the dome floor lights through the protective plastic cover on TripleSpec. The wavelengths at the ends of the orders are labeled in the image. The orders, from top to bottom, are 3rd through 7th. | A slitviewer/guider image, which encompasses a 4'x4' field. The slit is offset in the field but nearly centered on the optical axis of the telescope. This configuration enables the use of the rotator to search for guide stars in the rare event that one is not available in the default field of view. |
2. What to expect
Unlike visible wavelength spectroscopy, substantial airglow dominates the near-infrared portion of the spectrum, particularly at wavelengths longward of 1.5um. Beginning at 2.0um and longward a significant ambient thermal radiation component begins to contribute Poisson noise. The spectrum below illustrates these effects on a faint (J, H, Ks ~ 14 mag) object. The animation shows the star observed at two slit positions as is typical for TripleSpec observations. Evident are bright airglow emission lines filling the slit, particularly in the H-band (2nd from top) order. Wavelength increases to the left, and the rising thermal emission in the K-band order (top) is evident as well. Scattered high dark current pixels pepper the array. Pixels that blink on are off are cosmic ray hits. The exposure time for this image is 120 seconds.
Typically spectra are acquired at two slit positions and subtracted to suppress airglow line emission and thermal emission. The difference spectrum below shows the enhanced visibility of the spectrum in such a difference. Residual airglow lines are still present due to temporal variability in airglow, even on timescales of a few minutes. For longer integration near zenith, the small amount of flexure in the instrument (max 1 pixel) can also influence the self-subtraction of the airglow lines. Increasing Poisson noise from K-band thermal background is evident at the longest wavelengths (upper left).
-
a. Spectrograph
- 1. GD153 - an empirical case study
-
GD153 is a DA white dwarf with an effective temperature of
39,000K. The infrared portion of the spectrum of this object
is approximately Rayleigh-Jeans. GD153 has a V-band Johnson
magnitude of 13.35. The UKIRT faint standards list reports the
star (FS33 in the UKIRT list) as 14.24 at K with H-K = -0.078
and J-K=-0.223. The figure at left below shows the counts in
units of DN/s based on eight 120 seconds spectra of this
object. In this figure the flux in different orders has been
merged to give a total system response as a function of
wavelength. The adjacent figure gives the S/N per spectral
pixel for this object in the full 8x120= 960s integration.
Evident is the efficiency function of each of the orders,
peaking near the center of the order. The third figure shows
the calbrated response for this object in Janskys.
Based on the observation of this 14th magnitude white dwarf SNR=5 can be achieved in one hour in a three pixel spectral bin at magnitudes K=15.5, H=16.0, J=17.0.
< |
< |
< |
| Response in units of DN/s for GD153 (J=14.0, Ks=14.2). Counts have been binned where the source is detected in multiple orders. | SNR per pixel in 8x120s of integration on GD153. | Calibrated GD153 spectrum in Janskys. |
2. The limits of detectability
-
The images below shows the K-band detection and extraction of a faint
red source in one hour of integration time. During that hour the
source was observed in twelve five-minute integrations, alternating in
slit position ABBAABBAABBA. The count rate of 0.07 DN/s corresponds
to a K=17.5 mag continuum. With three pixel spectral smoothing the
continuum has an SNR=0.5 per smoothed spectral bin. In order to
produce a spectrum bright enough for extraction, the five minute
integrations had to be binned into 30 minute stacks at the "A" and "B"
positions. Overall, this spectrum may not be very useful
scientifically,
but it does illustrate the limits of seeing and extracting a source
spectrum under good conditions at the 3.5-meter.
Note that although the extracted continuum K-band magnitude is 17.5, the reported source K-band magnitude was 15.8. The observers struggled to guide on this faint target and the lower than expected detected magnitude is likely due to losses experienced in attempting to keep the light going down the slit.
 |
 |
 |
| The image result of differencing two 6x5 minute stacks of "A" vs. "B" position frames. The extracted source magnitude is consistent with a 17.5 mag Ks-band continuum. | The SNR per sets of 3 binned spectral pixels in the H and K spectral orders. The K-band SNR here is about 0.5 per 3 pixel bin. | The extracted spectrum in the H and K-band orders in units of DN/s. The spikes in the spectra are residual airglow lines contaminating the spectrum. |
b. Slit viewer
The sensitivity of the slit viewer dictates the faintest source that can be observed directly and placed on the slit. Below this sensitivity threshold observers will have to depend on blind offsetting to place a source on the slit. In general, if a source is too faint to be seen in the slit viewer its continuum will be difficult to detect in the spectrograph. The integration time for the slit viewer is adjustable, but 30 seconds represents a maximum practical integration for positioning the source and for guiding.
Once again GD153, K=14.2, provides a fiducial for slit viewer sensitivity. The frames below show, first, a raw guider frame with a bright source on the slit at the "B" position. The second figure shows a raw guider frame on a field containing many detectable faint sources - all of which are difficult to see in this view. As outlined below, collecting a background frame, shifting the field-of-view and subtracting that frame from subsequent frames removes all common-mode signals and provides a cleaner view of the sky. In this view (with integration time of 15 seconds) the K=14.2 mag target (just above the slit) is well-detected. The faintest stars readily visible are about 2.5 magnitudes fainter than GD165 or around K=16.5. With 30 second guider integrations it will be possible to see K=17 sources and place them on the slit. Longer guider integration times are likely to be unwieldy.
Boresite guiding requires sufficient spilled light to enable the guider to track the star (e.g. the leftmost figure below). The faintest start that provides sufficient spilled light has yet to be determined, but is probably in the range of Ks=13-14.
Guider sensitivity will depend on conditions - seeing and ambient temperature in particular. The example frames were obtained on a night with T=0C and sub-arcsecond K-band seeing and thus represent nearly the ultimate performance for this channel. At T=15C the system will be approximately 0.7 mag less sensitive than at T=0C under similar seeing conditions due to the increase in thermal background.
 |
 |
 |
| A guide frame from the observation of a bright calibrator. The target is at the "B" position on the slit. The dark features in the frame are either non-responsive pixels (circular region to the right) or defects in the surface of the silicon wafer mirror (e.g. above the slit). | A raw guider frame on a faint object. Field stars are present in this image, but hard to discern due to the pixel-to-pixel response variations combined with significant thermal illumination. The 14th mag target is visible, but is nearly lost because it falls near the reflective defect just above the slit. | The same frame as in the adjacent figure, but this time a displaced frame of identical integration time has been subtracted. The subtraction removes the systematics of the illumination and reveals the faint stars. The telescope focus was slightly off optimal for this exposure. The slitviewer/guider has astigmatism. The good news is that defocus is readily evident and easily distinguishable from poor seeing. The bad news is that defocus is readily evident. |
- Move telescope to target. Typically the source will appear within a few arcsecond of the slit. Since the field of view is 4'x4' it is unlikely the
target will fall outside the field of view. (Note that the slit mirror has a few defects, one of which is close to the primary slit (1.1")). On occasion, the target may land on the defect (or in the slit itself) rendering the source invisible if the target is faint. If the desired source is not evident, a few arcsecond offset may turn it up.
- If the source is faint it may be necessary to detect it in a difference between two guider frames. Although currently this procedure is a command line option, by the time of shared risk observing the ability to buffer and subtract a background frame will be included as a GUI button.
- Place target in the "A" slit position.
- [ctrl-leftclick] on the target will offset the telescope to the marked hotspot (which should be placed at the A position).
- By convention the A position is about 1/3 slit length from the left edge of the slit as seen in the guider display. Similarly the "B" position is about 1/3 slit length in from the right end of the slit.
- The displacement between the A and B positions is about 20 arcseconds.
- Set the number of Fowler samples, integration time, and source name in the TripleSpec instrument/expose windows. Shorter integrations provide better immunity to airglow line variation, however, the background limit - set by the inter-line airglow continuum, is reached only after about 3 minutes of integration. Five minutes of frame integration time is a practical and useful upper limit.
- Either
- Select an ABBA nod sequence (which should automate the positioning of the source at the alternating slit positions) or
- Expose on the "A" position and manually offset and integrate for an ABBA sequence.
- Repeat as desired.
Given that the slitviewer/guider operates at Ks band combined with the high spectral resolution of the spectrograph means that the sky is dark enough for source acquisition and sensitive spectroscopy when twilight is quite bright to the naked eye. Sunlight begins to interfere with five minute spectral integrations when the Sun is 6 degrees below the horizon. The Sun typically reaches this position 30 minutes after sunset or before sunrise. A Triplespec night begins early and ends late. Users should be prepared for initial source acquisition shortly after sunset.
a. Observing outline
b. Source and guide star acquisition with the guider/slitviewer
For bright sources, positioning a target on the slit is straightforward. Once a source is identified in the slit viewer, [ctrl-leftclick] on the center of the source will move the target to the "hotspot" location on the slit. After a few nudges to get the spilled light symmetrical guiding in boresite mode and integration can begin.
For faint sources attention will be required to maximize the source signal in the slit viewer images - which to 0th order is accomplished by increasing the frame integration time. There are two routes to obtaining optimal SNR in the slit viewer. The best practice remains to be determined.
-
i. Background subtraction - In the background
subtraction mode a single frame is buffered and
subsequently subtracted from each incoming guider
frame. The subtraction removes all of the common mode
structure that contaminates a raw frame and makes faint
sources visible at the expense of providing a (+) and (-)
image of each source. The telescope must be offset a few
arcseconds following the acquisition of the background image
otherwise sources will subtract from themselves making them
invisible. By the time of shared risk observing the background
subtraction mode will be implemented by pushbuttons on the
TUI guider GUI.
ii. Raw masked frames - The masks applied by the TUI guider have a pixel-to-pixel scaling that effectively flat fields the incoming frames. Typically thermal infrared dominated frames are not simply flatfielded because of structure arising from non-sky emission (consider the glow from a speck of dust on the window). For the sake of source acquisition and guiding such features are more of an annoyance and it may be possible to acquire even the faintest sources of the flat fielding is effective (and avoid the sqrt(2) noise penalty entailed by background subtraction). By the time of shared-risk observing it may be possible to do faint source acquisition on the masked (proc) guider frames.
The spectra above show that there is substantial airglow contamination across the TripleSpec bandpass. If consecutive exposures place a point source at two well-separated in-slit positions, subtracting these two spectra will, to first order, suppress the airglow line flux while maintaining the full signal from the target. The airglow line intensity can vary substantially even in the course of several minutes. In order to get good subtraction of the airglow lines integration times of less than 5 minutes are desirable. (In theory, the airglow should be removed in data processing as the slit region outside the source is fit and subtracted from the source. In practice the angle of the slit varies with respect to the dispersion direction making clean subtraction difficult - thus the desire to suppress/minimize the airglow signal.)
TripleSpec has four available slits. The primary TripleSpec design implements a 2.3 pixel wide slit which corresponds to a cross-slit spatial dimension of 1.1 arcseconds. The three other slits are 0.7, 1.5, and 1.7 arcseconds wide. These four slits reside on a gold-coated silicon wafer slit mirror supported on an 8 position geneva gear mechanism. A pulldown menu in the TripleSpec TUI Configuration window permits selection of slits (or selection of blocked positions halfway between slits).
-
1. Fowler sampling and read noise
-
The HAWAII-2 spectrograph array can be read non-destructively multiple
times during the course of an integration. Fowler sampling refers to
conducting a burst of N readouts at the beginning of an integration
and an equal burst of N readouts at the end of integration in order to
suppress read noise (FowlerN). Read noise suppression can be
important for TripleSpec, since the dark current is 0.05 e-/s and the
inter-airglow line continuum is weak. For correlated double sampling
(one read at the beginning of an integration and one at the end - also
``Fowler1") TripleSpec read noise is observed to be 5DN or 17
electrons. The TripleSpec TUI configuration menu permits the user to
select a range for the number of Fowler samples. The TripleSpec read
noise is observed to improve, as expected, with the square root of the
number of Fowler samples. With Fowler8 the effective read noise is
about 7 electrons. After the collection of 100 dark current or sky
electrons these factors will dominate the read noise.
2. Integration time and saturation
-
Integration/exposure times for direct readout devices can be
non-intuitive since there is no shuttering and the conversion of the
first and last pixels in one read of the device are staggered in time.
For the TripleSpec HAWAII-2 array the electronics require 792
milliseconds to address all pixels on the array (in 16 128x1024
stripes of pixels read out in parallel). As an example, consider a
Fowler1 integration. The array is reset, establishing the saturation
level and the first pixel of first readout (actually 16 all at once)
is read. 792 milliseconds later the last pixels of the read arrive.
The image captured is thus staggered in time from one end of a stripe
to the other by 0.8s. An 0.3 second delay is enforced prior to the
second readout in order to avoid spurious ``shading" across the image.
If the second readout begins immediately after this 0.3 sec delay (thus
realizing the minimum integration time) the first pixel will be read out 1.1s
after it was read the first time, and so on for all of the pixels on
the array. Despite the fact that 1.9s was invested in acquiring the
data, the image produced by differencing the two readouts has an
on-sky exposure time of 1.1s. The beginning of the last readout
started 1.1s after reset while the end occurred 1.9s after reset - the
saturation threshold varies across the chip! Saturation should be
considered including all of the readout time, not just the on-sky integration
time. For Fowler1 the minimum on-sky integration time is 1.1s, but saturation should be presumed to be estimated from a virtual 1.9s exposure.
Generalizing to FowlerN, N*0.8s is required for the first readout sequence as well as for the second readout sequence. At minimum integration time, currently with an 0.3 second delay between the last pixel of the first sequence and the first pixel of the last sequence, 0.3+2*0.8*N seconds are required to execute the entire sequence while 0.3+0.8*N seconds of integration are obtained on sky (which can be seen also by subtracting pairwise the first read of the first group from the first read of the second group, and so on).
TripleSpec accounts for the readout time in all Fowler modes such that the on-sky time will be the requested duration. This value will also appear as EXPTIME in the frame FITS header (another keyword INTDELAY provides the implemented delay interval between the last pixel of the first read and the first pixel of the last read).
The choice of N dictates the minimum on-sky integration time. A typical value is Fowler8 - enabling integrations as short as 6.7 seconds. Virtually all TripleSpec targets, including calibrator stars and calibration lamps, are observed with integration times longer than 10 sec. N=8 is the recommended Fowler setting for all TripleSpec data acquisition, unless exposures with duration less than 7 seconds are required.
The HAWAII-1 slitviewer/guider array could be read out in FowlerN mode, however the level of thermal background on the array produces thermal Poisson noise far in excess of the system read noise (also of order 17 electrons). In the interest of dynamic range and efficiency, the HAWAII-1 array only operates in Fowler1 (CDS) mode.
- Spectra: /export/images/tspec/PROGRAM/UTDATE/
- Guide frames: /export/images/tcam/UTDATE/
Like DIS, TripleSpec uses independent TUI windows to position the source on the slit and acquire the spectra.
-
a) TripleSpec guiding with TUI
-
Choose "TSpec Slitviewer" from the Main TUI "Guide" pull-down menu.
This guider window behaves functionally like any of the other TUI
guider windows.
b) TUI TripleSpec spectrograph control
Choose "Tspec" from the Main TUI "Inst" pull-down menu.
c) Data directories:
 |
 |
 |
 |
| The TripleSpec TUI guider window. Functions/controls are similar to guiders for DIS/Echelle. Remote users will want to have 2x2 binning activated to minimize download time. Ultimately, sub-framing will provide the most rapid response, but the operation of this mode is still being engineered although it is expected to be available by the time of shared risk observation in 2008Q3. | Main TSpec instrument TUI window showing configuration. In this window the number of Fowler samples can be set and the slit can be rotated to the desired position. Tip-Tilt mode refers to a future mode of the instrument where the spectrum can be steered at the sub-pixel level by a piezo-electric stage under one of the fold mirrors in the system. Array power should, of course be on, and is the usual default. The "environment" button provides instrument internal temperatures and pressure. | TSpec Exposure window: Integration times are set and exposures initiated in this window. The "type" radio buttons simply set a FITS keyword for the recorded data. The system can take multiple exposures, ``#Exp" at the push of one button. This function is particularly useful for bright standards where a set of 5 exposures can be taken consecutively at the "A" slit position followed by 5 exposures at the "B" position to complete the observation. ``Filename" is the root file extension. For example, if "xxx" is chosen for the file name the resulting fits file will be "xxx.yyyy.fits" where "yyyy" is an incrementing frame number that increases steadily through the night. | Nod Script window: Obtained from the "Scripts" pull-down menu, this window has all of the functionality of the Expose window, but also includes a "cycles" option which currently drives the telescope in an "ABBA" slit position seqeuence for each cycle requested. |
6. Calibration
-
a. Flat fielding and wavelength calibration
- No illumination - only continuum thermal radiation evident at the longest wavelengths.
- Neon and Argon lamps on for spectral calibration
- Bright quartz lamp on for continuum illumination.
- (Darks - No darks are required for TripleSpec data analysis since all reductions are done differentially between pairs of images, both of which contain the same "dark" bias and dark current contributions.)
- Cushing, Vacca & Rayner 'Spextool: A Spectral Extraction Package for SpeX, a 0.8-5.5 Micron Cross-Dispersed Spectrograph' -URL
- Vacca, Cushing & Rayner 'A Method of Correcting Near-Infrared Spectra for Telluric Absorption' - URL
Like any spectrograph, TripleSpec requires both continuum and line illumination for flat-field and spectral calibration (ultimately spectral calibration will be carried out with the airglow lines but this function has yet to be implemented in the data reduction pipeline). A typical calibration sequence to be conducted once each night (possibly once per run) consists of ten 60 second exposures with:
b. Telluric and flux calibration
7th-9th magnitude A0V stars are ideal calibrators for most TripleSpec observations. The Triplespectool pipeline and this calibration method are described in:
TripleSpec observers that make use of Mike Cushing's Triplespectool should refer to the tool as a 'modified version of Spextool' in the paper text and reference the above papers.
Aside from broad hydrogen line absorption, A0V stars have few other intrinsic features. The overall A0V spectrum can be well modeled and is fit and removed from the target spectrum in one step of the Triplespectool data pipeline.
Note that for solar-system targets observed in reflected sunlight a G2V standard is reqired instead of an A0V standard. Direct division by the G2V standard will provide correction both for telluric absorption and for the solar absorption spectrum with no modeling require. This option is available in the Triplespectool pipeline.
For proper telluric absorption correction (which can be time variable due to the changing water vapor content of the atmosphere) it is important to observe an A0V calibrator close in time, in airmass, and ideally in angular separation from the science target.
The files below contain thousands of potential A0V and G2V standards to aid in selection of a calibrator .