2018 K-band Resolution

In June 2019 we were made aware of an issue with the IGRINS K-band spectral resolution. This issue impacted spectra acquired between April 2018 and May 2019, when IGRINS was at Gemini South and the Discovery Channel Telescope. The cause of the resolution change was found to be loose fasteners on the back of the K-band detector mount.

In August 2019 we realigned the IGRINS detectors and achieved optimal focus of the optical system. Both the H- and K-band echellograms are again within specified requirements with an average spectral resolution R~45,000.

The text below was sent to all IGRINS PIs who had data acquired between April 2018 and May 2019.

Here is what we now know:

1) The measured line widths of OH sky lines across the K-band echellogram were changed in the commissioning data at Gemini South in April 2018. Prior to this (July 2014 - January 2018) the K-band spectral resolution was reliably >40,000. The line widths were changed for the entire visit to Gemini and the subsequent visit to the Discovery Channel Telescope. Changes in the H-band resolution were +/-5% across the echellogram, with a mean resolution R>40,000. These variations to the H-band echellogram are noticed with each handling of IGRINS and not considered to be an issue. The figure below shows measurements of OH line widths across IGRINS’ spectral orders (in pixel space). While only two epochs are shown (2017 and 2018, before and after the K-band resolution change) we have verified that these trends persist across many nights. The change in the K-band spectral resolution has remained constant since April 2018.

2) The 2D changes in spectral resolution are shown in the figure below. Just as in the plot above, OH line equivalent widths measured after April 2018 show significant variation in the K-band. Wavelengths on the IGRINS detector increase with increasing x pixel and towards lower spectral orders. The largest change to the echellogram (reduction of the spectral resolution to R~20,000) occurs in the short-wavelength end of the long-wavelength K-band orders. This is where the CO bandheads are located (lambda > 2.3 microns).

3) The 1D spectra of GJ 281 are shown below. The CO bandhead in two adjacent K-band orders depict the measured changes in the spectral resolution. The top spectrum is from 2017, before the resolution change, and the bottom is from early 2019. The long wavelength end of the ‘green’ order is unchanged while the short wavelength end of the ‘blue’ order now shows the degraded spectral resolution.

What do we expect to be the cause?

The overall cause of the degraded resolution is a defocus of the K-band echellogram. We suspected a number of optical elements, and in August 2019 we inspected IGRINS in the UT Austin lab. After finding loose fasteners on the K-band detector mount, we repinned it and then tightened everything. Cold testing on August 11th showed that the echellogram focus was corrected back to the design specifications. We will add new test procedures for IGRINS unpacking at Gemini and will ship any needed spares and tools along with IGRINS incase this spectral resolution issue returns.

What if I have IGRINS data impacted by this resolution change?

The first thing to do is to quantify the resolution change in your own data. If it doesn’t impact your science, then you don’t need to do anything. Once the functional form of the change in resolution is known for your data, then it can be accounted for in your analysis. In the future we plan to provide all users with the information that they need to have a better understanding of the IGRINS spectral resolution on a nightly basis as part of the IGRINS Pipeline Package.