AURIGA: R&D project on the SQUID

The dcSQUIDs have shown to be the most sensitive
amplifiers on a wide frequency range. The most
frequently used figure of merit for these devices
is the energy resolution per unit bandwidth referred at the input coil;
in this case the energy resolution is given by:
dE = (S_{i} L_{i})/2 where S_{i} is the monolateral spectral density of the equivalent current noise circulating in the input coil of inductance L_{i} (see fig.1left). In AURIGA the signal source impedance of the input circuit is resonant and has a high quality factor: therefore the energy resolution is not sufficient to fully describe the system noise characteristics. A more general model is needed. The SQUID must be considered (see fig.1right) as an ideal noiseless current amplifier with two partially correlated input noise sources: a current noise source i_{n} in parallel with the input port and a voltage noise source e_{n} in series with it. The current noise is superimposed on the input current signal and, in principle, can be easily measured with open input. The voltage noise produces a driving force, which is called backaction, on the signal source with impedance Z_{i}. Unfortunately, with respect to a generic current amplifier, there is a further difficulty for the SQUID: in fact its characteristics depend on the input circuit parameters. The effect of the source on the SQUID characteristics can be neglected when the coupling between the input coil L_{i} and the SQUID loop L is low. This is not the case of the fundamental Physics experiments like AURIGA: as the signals are extremely weak, this coupling must be as strong as possible. An experimental check of the noise theory is lacking just for the low noise dcSQUIDs employed in the resonant gravitational wave detectors: in these experiments the SQUID noise is one of the factors limiting the detector sensitivity. We are performing voltage noise measurements in order to obtain the precise knowledge of the noise sources, necessary for the optimal noise matching and for the estimate of the backaction contribution to the total detector noise. Clik here for a photograph of the SQUID assembly we are now working on. 