Analysing a TiEMPO simulation of a DESHIMA 2.0 observation of the dusty starburst galaxy HFLS3

Abstract

Observing dust obscured high redshift galaxies is vital to understand the evolution of the early universe and the formation of stars and galaxies. HFLS3 is such a galaxy, and its emission spectrum can be detected at submillimeter wavelengths. With the Deep Spectroscopic High-redshift Mapper 2.0 (DESHIMA 2.0) it is pos- sible to observe these high redshift galaxies at the ASTE telescope in Chile. With the recently developed Time- dependent End-to-end Model for Post-process Optimization (TiEMPO) it is possible to simulate observations of DESHIMA2.0.
With TiEMPO it is possible to simulate different sky positions to accommodate different scenarios of wind direction. However, since TiEMPO is fairly new there are still some problems with the simulation of several of the the possible sky positions. More importantly, the model has to be tested with realistic high-redshift sources that are interesting for observations with DESHIMA 2.0. To predict whether a source can be observed with DESHIMA 2.0 it is necessary to estimate what the signal to noise ratio will be. Parameters like the pre- cipitable water vapor and the observation time determine if a source can be measured.
In this report, a solution is introduced to make it possible to use 6 different sky positions while simulating DESHIMA 2.0 measurements with TiEMPO. This solution is part of the current TiEMPO version, and can be used in future simulations.
With the improved model we simulate an observation of the dusty starburst galaxy HFLS3. The CII line in the emission spectrum is fainter than simulations done so far. A new model to simulate the emission spectrum was made to accommodate for this. With the use of beam switching some of the noise from the atmosphere is removed from the data. The simulation is compared to previously made observations as described in [2]. From the resulting signal we calculate the standard deviation σ to determine the signal to noise ratio (SNR). The values found for σ correspond well with the expected relation between σ and integration time. We say that a emission line is detected if the SNR is greater than 5. The calculated SNR of a 16 hour observation with a pwv value of 1.0 mm is 5.2, which shows that HFLS3 can be detected with DESHIMA 2.0. Two simulations of 8 hours with pwv values of 1.0 and 0.5 mm are compared as well. After 8 hours of observation, the SNR is 3.9 for a pwv value of 1.0 mm and 6.0 for a pwv value of 0.5 mm. With this lower pwv value the galaxy can be detected after 8 hours.
The analysis of this project can be repeated on other sources to make a database for future DESHIMA 2.0 simulations. Hereby it is key to have models which can predict the emission lines of a galaxy accurately.