Faint flux performance of an EMCCD
ABSTRACT
Thorough numerical simulations were run to test the performance of three processing methods of the data coming out from an electron multiplying charge coupled device (EMCCD), or low light level charge coupled device (L3CCD), operated at high gain, under real operating conditions. The effect of read-out noise and spurious charges is tested under various low flux conditions (0.001 event/pixel/frame < f < 20 events/pixel/frame). Moreover, a method for finding the value of the gain applied by the EMCCD amplification register is also developed. It allows one to determine the gain value to an accuracy of a fraction of a percent from dark frames alone.
1. INTRODUCTION
The advent of electron multiplying charge coupled devices (EMCCD) allows one to apply a gain to the pixel’s charge before it reaches the noisy output amplifier.1 Sub-electron read-out noise levels are thus reachable. However, this kind of signal amplification comes to a price: the stochastic multiplication process induces a noise on the gain level that renders impossible to determine the exact gain that has been applied to a pixel’s charge. This statistical behaviour thus adds a noise factor that reaches a value of 2 1/2 at high gains.2 The effect on the signal-to-noise ratio (SNR) of the system is the same as if the quantum efficiency (QE) of the CCD would be halved.
Some signal processing techniques may allow one to overcome this noise and recover the full silicon QE of the CCD. Simulations are run to test the behaviour of different processing techniques. The simulations and their results are presented in section 2.
Throughout the processing of the signal coming out of an EMCCD, the mean gain of the EM register is a key value that has to be determined with a high accuracy if one wants to acquire absolute photometric information with such a device. In section 4, an algorithm is developed so that the real gain of the EMCCD can be calculated through the processing of dark frames alone.
2. SIMULATIONS
Thorough numerical simulations are run to properly understand the effect of the multiplication register on the pixel signal. This enables the optimization of the signal processing in order to recover the original pixel’s value. Simulations concentrated mostly on low fluxes, 20 photons/pixel/frame being the highest flux simulated.
The simulation process is divided into four sections:
-
The generation of the pixel’s signal. This includes the photon’s signal as well as the spurious charges generated during the vertical transfer;
-
The journey of the pixel’s charge into the EMCDD multiplication register;
-
The simulation of the output amplifier’s noise that is added to the multiplied pixel’s charge;
-
The digital processing of the output signal to try to recover the amount of input photons.
Для продолжения чтения вы можете скачать полную версию материала по ссылке ниже
