LRIS
LRIS-B Overview

Chuck Steidel, P.I.
December 2000 (updated June 2002)

Contents

Experienced LRIS users should find it very straightforward to use LRIS-B, since the new spectrograph channel behaves in most ways very similarly to the red side. Even those interested primarily in wavelengths longward of 5000 Å may find that the blue side is very useful for improving observing efficiency (particularly for slit mask setups) and for allowing the use of higher dispersion gratings without sacrificing spectral coverage.

What is LRIS-B?

LRIS-B is a new UV- and blue-optimized spectrograph within LRIS that allows simultaneous imaging or spectroscopy with LRIS-R through the use of selectable dichroic beam splitters. The blue channel receives the reflected beam from the dichroic and allows for grism spectroscopy and/or imaging of the same field as seen by LRIS-R (see Figure CDR-8). LRIS-B is optimized for the 3100-6000 Ångstrom range and has outstanding throughput in the near-UV and visual. With both the red and blue cameras in operation it is possible to image in two bands simultaneously or to obtain multi-object or long-slit spectra covering the entire optical window (3100-10,000 Å) in one integration.

The blue spectrograph has completely independent filters, dispersing elements, and spectrograph camera from those on the red side. The blue camera was designed to produce very close to the same image scale at the detector as the red camera, so that with the new science grade CCD mosaic on the blue side, the pixel scale is 0.015/0.024 times that on the red side, or about 0.135"/pix. The total slit length seen by the blue camera incorporates the entire unvignetted field of the instrument, so that the full 8.0' field is useable on the blue side (as opposed to ~7.3 arc minutes on the red side with the current detector). Each of the dichroic, grism, and filter systems consist of two main subsystems:

The two sides share the same off-axis paraboloidal collimator mirror; this collimator mirror has been coated with a special hybrid coating (developed at LLNL) that reflects as efficiently as Ag above 4000 Angstroms and better than Al in the UV and blue.

Dichroics

LRIS-B receives light by intercepting the collimated beam off the collimator mirror with a deployable dichroic beam splitter (there are 4 different dichroics available, plus an over-coated aluminum mirror that sends all light into the blue spectrograph channel). Wavelengths shortward of the dichroic cutoff (the dichroics are coded by the wavelength of light that corresponds to the 50% point in transmission/reflectance; e.g., the D500 will pass half of the 500 nm photons into the red camera and reflect the other half into the blue camera) are reflected into the blue spectrograph. Typical dichroics will have a transition region that normally spans about 200 Å centered on the nominal 50% point (see the reflectance and transmission curves for the dichroics). Outside of that range, the dichroics generally have transmission efficiencies of better than 95% and reflective efficiencies of better than 98%. Dichroic beam splitters invariably produce low frequency ``wiggles'' in the efficiency of transmitted and reflected light that can be noticed in high S/N spectral observations; these can be removed through flat fielding. The dichroic substrates are 1-inch thick BK7 glass, and are AR-coated on the back side. The dichroics produce no noticeable shift in the focus for the red side of the spectrograph, while there are very small shifts in optimum focus of the blue side spectrograph camera that depend on the chosen dichroic.

At present, for spectrscopy using dichroics, there are ghost spectra that have been determined to be second order light dispersed by the red side grating that is returned to the dichroic (and which passes through because the dichroic is basically a long-pass filter) and then makes it onto the blue side detector as zeroth order light (i.e., it is not further dispersed by the blue grism). The ghosting will thus depend upon red-side grating tilt and by the prism angle of the grism in use. We have ordered a short-pass filter for use in the blue side filter carousel that will block these ghosts for many setups (the ghost line intensity is about 0.5% of the red side line intensity--they are easily seen in arcs but can be quite subtle in taken on the sky)--it is expected to be available by August 2002. In the meantime, if your observations require deep near-UV and blue spectroscopy, I would recommend using the dichroic mirror, which eliminates this ghosting issue. There are no obvious ghosts seen with the 1200 l/mm grism-- because the large prism angle diverts the zero-th order image of the dispersed red light away from the camera. Ghost images should not be a problem for any imaging observations.

Grisms

There is a very important difference between the dispersive elements on LRIS-B as compared to LRIS-R: they are grisms instead of gratings. What this means is that the wavelength coverage you get is fixed for a given slit position in the focal plane. The only way to get different spectral coverage is by either moving the slit in the focal plane (easy to do with slit masks; not so easy for long slits to be viewed with the slit-viewing guider camera). Generally speaking this lack of adjustability will not be as crucial now that we have a double-beamed spectrograph since the spectral coverage that one needs can be covered by two separate channels and there is a wide choice of dichroic options. The one exception is the highest dispersion grism, the 1200/3400, which covers the wavelengths 2990-3980 Å for a slit at the center of the field and 2880-3870 for the long slits; since the shortest wavelength dichroic cuts at 4600 Å, it would not be possible to get complete wavelength coverage using the two sides of the instrument. There is a provision for inserting ``spectrum-shifting wedges'' in the LRIS-B filter wheel, which shifts the spectral format, but no prisms have been manufactured for this purpose for first light. With a specially-designed slit mask placing the slit at the left-edge of the LRIS field, it is possible to shift the wavelength coverage by an additional 300 A to the red.

It is possible to shift between spectroscopic and imaging mode simply by removing or inserting the grism into the beam (choosing deploy or stow on the XLRIS GUI). This can be done in less than 30 seconds, and hence we anticipate that most users whose programs involve multislit spectroscopy will want to do their slit mask alignments using the LRIS-B imaging mode, leaving LRIS-R in spectroscopic mode (with grating in place). This time savings can amount to as much as 35-45 minutes of observing time during a typical LRIS multi-slit night.

Grisms, like gratings, produce "second-order" light. For extremely blue sources, there may be significant second-order light appearing longward of (first-order) 6000 A for the 300 and 400-line grisms now that the UV QE of the CCDs is very high.

Filters

Currently there are 3 imaging filters in the LRIS-B filter carousel: The U filter has a bandpass almost identical to the SDSS u' filter; the B filter is an interference filter. The G filter is very similar to the SDSS g' filter, although it is slightly narrower. A V filter was manufactured for LRIS-B but was defective (delaminating). A new version of that filter should be available by late summer 2002.

Note that on LRIS-B one has the option of pulling any filter out of the beam without selecting a different filter (simply choose "clear" from the blue filter list on the XLRIS GUI to stow the filter in the carousel).

The most efficient means of doing slit mask alignments will be to obtain field images unfiltered using only the dichroic to define the bandpass, so that deploying the grism is the only change to the spectrograph configuration necessary to go from alignment to the start of the spectroscopic integration.

LRIS-B Camera

The LRIS-B camera is optimized for throughput and image quality over the wavelength range 3100-6000 Å, and will produce passable images to about 7000 Å, beyond which the performance falls off noticeably. The end-to-end camera transmission is expected to be 92% or better over the design wavelength range. The camera is capable of producing images with 80% encircled energy in a diameter of 25 µm (about 0.22") and so if everything is in focus the instrument should be able to deliver extremely good blue and UV images. Focus is something that observers should pay some attention to, as the camera design predicts camera focus changes of about 30 µm (just about enough to see a noticeable change in image quality) per degree C change in temperature. A new tool (Xfocus) allows relatively painless measure of camera focus -- by all means take advantage of this.

LRIS-B Overall Throughput

New measurements of spectroscopic throughput were made when the new LRIS-B CCD was commissioned on June 4. Figures ( lrisb_eff.pdf lrisb_eff_wtel.pdf ) summarizing the efficiency curves is linked. These curves were made using the same assumptions as for the curves for red side gratings (i.e., to get total system throughput including the telescope, multiply by about 0.8) to facilitate direct comparisons. Note that the 300/5000 grism provides a peak efficiency of about 56%, about 47% higher than the 300/5000 grating on the red side at visual wavelengths.

Observing Modes with the New LRIS Double Spectrograph

The dual-beam nature of LRIS now allows obtaining spectra with optimized throughput over a very large wavelength range, and it also allows imaging in two bands simultaneously. Some care is necessary in planning the configuration of the spectrograph now that there are so many options for placing optics in the beam(s).

For example, it is possible to image in (u' or B) and (R or I) simultaneously with the right choice of dichroic -- D560 would probably be the best choice in this case. As another example, one can choose to image in V and I simultaneously, with the D680 dichroic in place. There are some issues that remain to be ironed out concerning the best way to focus the telescope for imaging, depending on whether the observer cares more about the blue or red side images (only the red side can currently be used for running MALIGN). Obviously, it is possible to make ``bad'' choices of filter and dichroic combinations (e.g., D560 would be a bad choice for V band imaging on either side, and D680 would be a bad choice for R band imaging on the red side) and at the moment such mistakes are not caught by the instrument control software.

For spectroscopic observations, there are many combinations. The choice will depend upon what resolution is needed in the blue versus red, etc. Some examples appear in the accompanying table.

Grism Dichroic Grating
400/3400 D560 600/5000 [coverage ~3150-8000 Å]
400/8500 [coverage ~3150-9400 Å]
1200/3400 D460 [any red grating; gap in spectral coverage]
300/5000 D680 832/8200 [coverage ~3200-8600 Å]
600/4000 D500 600/5000
600/7500
Instrument Configurations for LRIS-R&B Spectroscopy

Changes in LRIS Post-LRIS-B Installation

You will notice that the edge of the LRIS field on the right side (the side nearest the optical axis of the telescope) is vignetted over the outermost 25 arc seconds of the full LRIS imaging field. This is due to a slight obscuration of the field by the internal rails that were installed to carry the dichroics into the beam. This field vignetting is present whether or not a dichroic is deployed, but should not affect any spectroscopic observations.

Since many new optical components were installed inside the LRIS spectrograph body, there is the possibility of unwanted internal reflections or scattered light being introduced. While every effort was made to add baffles in several places, it is possible that the internal baffling can be improved. If you find evidence for scattered light or internal reflection problems, please alert your instrument specialist and provide as much information (images, ideally) as you can so that improvements might be made.

There is now a very large number of moving mechanisms inside LRIS; all 4 mechanisms on the blue side (focus control, grism, filter, dichroic carousels and transports) can move simultaneously, while 2 mechanisms (grating plus slitmask/filter/focus) can move simultaneously on the red side. All communications with the motor control is done via terminal servers (one for the red side and one for the blue side) and occasionally you might find that a move timed out or was left in an unknown state. We have found that any failures of blue side mechanisms can often be fixed simply by trying the failed move again. Until we have had more experience running both sides of the spectrograph simultaneously it will be necessary for observers to be on their toes to make sure the spectrograph is configured in the desired manner.

Insertion of a dichroic does not alter the LRIS-R camera focus value, but it will have a small effect on position of the beam footprint at the LRIS-R grating (which results in small shifts of the wavelength range seen at the CCD). The dichroics should be essentially identical in terms of these shifts.

Because it is a reflection off of the dichroic that sends light into LRIS-B, and each dichroic is independently adjustable in tilt and piston, there are small focus differences between the different dichroics (and the focus offsets should be redetermined whenever adjustments are made to the dichroics in their cells) which are tabulated for your use. In addition, the footprint at the CCD will vary somewhat from dichroic to dichroic. Thus, if you will be using images of slit masks taken in the afternoon to aid in slit mask alignments during the night, you should be sure that you are using the same dichroic as you will use for your setup images during the night.

Go to: LRIS Home Page - Instruments Home Page - Keck Home Page
Last modified: Wed Jun 20 13:17:42 HST 2001