Overview

The apparatus has a resonant cavity where the process would take place. The cavity can be tuned from 460Mhz to 810Mhz, based on the mass range which the Axion could have. Since the mass of the Axion is very low in that range (from 10^-2eV to 10^-6eV), only low-energy photons (microwaves) will be produced. This means that we need a way to amplify the signal so that we can actually tell if something is happening. The apparatus has been using an HFET until recently, where it was replaced with a SQUID amplifier. The strength of the magnet that is being used for the apparatus is about 8 Tesla.

However, amplifying the signal is not easy, since electronics have "noise" associated with them, meaning that they pick up whatever ambient signals are in the area and amplify that, also. Ambient signals are any type of electromagnetic field, like one surrounding a computer monitor, cell phone, etc, and of course, the Earth has its own magnetic field and is influenced by the sun and moon. This field interferes with all electronics, but most are not sensitive enough to care. The apparatus, however, is detecting very small fluctuations, so there has to be as little interference as possible.

Ambient temperature is also a source of noise, since anything with heat radiates photons (in the infra-red spectrum, but photons nonetheless) which would interfere with our detection of the low energy photons from the Axion. So, for example, if photons radiated by warm objects have an energy of 100 units (just some made up unit to illustrate the point), and the Axion's photon has an energy of 101 units, you aren't sure if you detected an Axion, or just some noise from the electronics. In order to find out, you have to take many such measurements and average them out, then you will see whether you are detecting anything because the noise will cancel out if you've taken enough measurements.

However, if the radiated photon has an energy of 50 units and your Axion's photon has 101, then you can be much more certain that you saw something, and it wasn't just some fluctuation. This is why we have to cool the apparatus, and also why the HFET amplifier was replaced with the SQUID amplifier -- it has a better signal-to-noise ratio at low temperatures, which helps the apparatus scan for Axions faster.

Specifically, ADMX's sensitivity, and thus how fast ADMX can scan over the possible Axion masses, is inversely proportional to the square of the sum of the cavity temperature and the amplifier noise temperature. In the original experiment, the cavity temperature was around 1.2K and the amplifier noise was around 2K, giving a sum of 3.2K. In Phase 1, the HFET amplifiers are replaced by SQUID amplifiers, which have a noise temperature that is similar to the cavity temperature, so the amplifier temperature is brought down to 1.2K, making a sum of 2.4K. This means we can scan over masses about 60% faster.

First Results

This the first power spectrum taken with the phase 1 configuration of ADMX when cold. It shows the power in the cavity over a small frequency range. Most of the shape is due to the bandpass filter. The peak in the center is a test signal injected into the cavity.