LUCI2 vs. LUCI1
In the 2015 summer shudown it was decided to take LUCI1 offline in order to retrofit it with a new Hawaii-2 RG detector, and offer LUCI2 for partner science in semester 2015B. On this page we discuss the major differences between LUCI1 and LUCI2.
While we are working hard to update all of the necessary tools and documentation (User Manual, Exposure-Time Calculator, Planning Tool) in time for the start of observations this fall, the information on this summary page will serve as a guide to understanding how to adjust your observations to make more efficienct use of LUCI2.
Keep in mind as you go through the information presented here that that LUCI1 and LUCI2 are mechanically and optically identical to within normal manufacturing tolerances. The imaging scales in the N1.8 and N3.75 cameras are about the same, and both instruments have identical sets of filters, a standard set of seeing-limited long-slits, and the pairs of 210_zJHK (G210) and 200_HK (G200) gratings were made from the same masters. To accomodate the option for spectroscopy with adaptive optics, the seldom-used lower-resolution 150_Ks (G150) grating has been replaced by the G040 grating for use with the N30 adaptive optics camera. Since the AO capabilities are not yet commissioned, this capability will be discussed at a later date.
There will be a lot of new information for LUCI2. These links will help you jump around to the different references.
Hawaii-2RG (H2RG) vs. First Generation Hawaii 2
The primary difference between the two LUCIs is the use of a Rockwell Hawaii-2RG (H2RG) detector in LUCI2. This brings some significant advantages, along with some caveats that it will be important to keep in mind when planning your observations.
Quantum Efficiency (QE)
By far the largest difference comes from the significantly higher QE of the H2RG detector. The average QE across all four near-infrared bands (zJHK)QE is about 83% for LUCI2 and 54% for LUCI1. In a comparable exposure time, LUCI2 will go about 0.23mag fainter than LUCI1 at fixed SNR.
The full-well capacity of the LUCI2 detector is about half as large as with LUCI1. Combined with the higher QE, this means that LUCI2 saturates three times faster than LUCI1! Keep this in mind especially when setting up observations of bright sources, like telluric standards. For the lower resolution G200 grating, A0V stars brighter than 8th magnitude WILL saturate in the minimum exposure times. For the higher resolution G210 grating the telluric standards can be a magnitude brighter.
The H2RG detector shows sigificantly higher persistence as compared to LUCI1. There is also some intrinsic structure to the persistence pattern on the LUCI2 detector. You will see this most strongly during a spectroscopic acquisition sequence and in the first few spectra taken on a given source. You can see this in this image.
ALL near-infrared detectors are non-linear! While the canonical HgCdTe detectors are about 4% non-linear at 80% full well, LUCI2 is much higher: about 18% non-linear at 80% full well. Even at 10% full-well (about 5500 ADU) you are already more than 2% nonlinear. Thus, if you work with high-dynamic range data or you care about doing higher precision photometry, it will be very important to linearize your data.
Linearization should be the first step applied to raw LUCI2 data. The equation is:
ADUlin = ADUraw + 4.155×10-6(ADUraw)2
Since any reduction of the data will include subtraction of a dark or temporally adjacent image, I don’t include the order zero constant term from the linearization fit. We have not explicitly measured a linearization coefficient for the MER readout mode, but it should be the same as the equation above. Proceed with normal reduction algorithms using the linearized data.
There is some structure visible in the pixel-by-pixel linearization map, but when these data were taken the illumination of the detector by the calibration unit showed significant gradients and there was a shadow from the W unit beamsplitter cube that affected a fairly large area. When both of these issues are fixed, we will re-derive the linearization coefficients.
At the speed the detector is read out, there is some crosstalk visible from very bright sources (well beyond saturated in one DIT) between the amplifier channels. Because of the architecture of the detector (32 parallel amplifier channels that are 64×2048 pixels in size) the crosstalk signals show up in a vertical line, spaced every 64 pixels. You can see an example of what crosstalk looks like in this image.
LIR vs. O2DCR Readout Mode
With the newer generation of readout electronics and the H2RG detector comes a new and significantly more efficient readout mode: Line-Interlaced Readout (LIR). This is still a double-correlated read, as was the O2DCR mode for LUCI1. But where every DIT on LUCI1 incurred the 4.0 second readout overhead, the line-interlacing readout for LUCI2 incurs a single 2.75s readout overhead independent of the number of NDITs (see this diagram)! As each pixel in a 64-pixel “line” is visited, it is read out as the final read of the preceding exposure, reset, then read out again as the first read of the next exposure, resulting in almost no dead time between adjacent DITs. As an example, consider a 60s exposure made up of NDIT=15 integrations of DIT=4.0 seconds, the clock time expended on the integration (not including time to integrate and save the file to disk) would be:
LUCI1: 1 x [ 15 x (4.0sDIT + 4.0sreadout) ] = 120 seconds
LUCI2: 1 x [ 15 x 4.0sDIT ] + 2.75sreadout = 62.75 seconds
Note that for imaging through all broadband filters and most of the narrowband filters, the LUCI2 detector is fully background-limited in the minimum 2.75s exposure time. Coupled with the higher QE and lower full-well capacity of the H2RG detector, and the efficiency of the LIR readout mode, virtually ALL imaging (seeing-limited spectroscopic acquisitions as well as normal imaging) should be done with a short DIT (~3.0s), using multiple NDITs used to get the required integration (NDIT x DIT) at each dither position. This will also help to keep the detector in the less non-linear regime and reduce persistence.
LIR vs. MER Readout Mode
While the LUCI1 and LUCI2 MER (aka MER10) readout modes are nearly identical, we have run into issues with LUCI1 where MER mode was used inappropriately. MER mode is not the “spectroscopic” readout mode, it is the mode to use when you need to reduce the read noise to reach the background limit in a reasonable exposure time. This can be necessary for faint object spectroscopy or imaging through some of the shorter wavelength narrowband filters. Generally, most imaging and bright object spectroscopy (e.g. telluric standards) are background limited even using the LIR readout mode.
LUCI2 has been set up to have a standard astronomical orientation (north up and east to the left at PA=0). This is horizontally flipped as compared to LUCI1. In spectroscopic mode the red end of the spectra are now to the right. This is more in-line with how astronomical instruments are set up, but do keep this in mind as you set up finding charts and such.
Exposure Time Calculator
The original LUCI exposure time calculator (ETC) was set up to handle both LUCI1 and LUCI2 planning. However, until the calibrations from LUCI2 commissioning are incorporated into the ETC, you will have to use the LUCI1 numbers. Note that because of the higher QE, LUCI2 is about 0.23 magnitude more sensitive than LUCI1 was in the same integration time. However, we recommend that you do NOT adjust for this difference at this time and just use the LUCI1 total exposure time estimates, although you do still need to consider the higher QE and faster saturation in setting your DITs and NDITs. It is better initially to get data a little deeper than needed than it is to find out some unanticipated problem prevents you from completing the science project. We will adjust this recommendation as we gain on-sky experience with LUCI2.