Characterizing the noises of InGaAs cameras for astronomical observations

Recently, InGaAs cameras have been utilized in time domain astronomical observations in the infrared bands, taking advantage of their improved performance. However, the noise levels of InGaAs cameras remain high compared to Charge coupled device (CCD) in optical bands, thereby limiting the signal-to-noise ratio for photometry. We characterize noise of Ninox 640 SU InGaAs camera. We test the noise components originating from bias, dark current, and flat, and analyze the readout noise, dark current, non-linearity, and the variation in responses between pixels. Bias and readout noise of CCD in optical band are stable. The frame with the minimum integration time of camera is generally considered as bias, and the readout noise is calculated based on it. However, count and noise of InGaAs camera are still unstable during a short period of integration time, prompting a detailed discussion on bias selection and readout noise calculation methods. The photon transfer curve (PTC), which represents the variance as a function of signal, is commonly employed to determine gain of camera from the slope. However, our PTCs exhibit varying slopes corresponding to different levels of brightness in the light source, indicating a non-constant slope. Consequently, we maintain a fixed integration time while increasing the signal level by intensifying the light source. This adjustment yields constant PTC slopes, consistent with the expectation, thereby suggesting the presence of additional noise that increases with integration time. Therefore, when measuring the gain of InGaAs cameras via PTC, it is imperative to opt for fixing exposure time and changing light brightness. Following the characterization of the noise components, we will develop correction methods for each noise and apply these methods to frames obtained from astronomical time domain observations. Finally, we will discuss the potential of InGaAs cameras in infrared time domain observations.