An optical / near-infrared detector array to help study stars, galaxies and planet formation
15 November 2016
A Microwave Kinetic Inductance Detector (MKID)
As astronomers usually want to look at the faintest objects in the universe, they need very sensitive detectors and ideally every photon has to be captured. In addition, they also often want to break the light up into its constituent colours, i.e. obtain a spectrum, and time the arrival of the photons to see if they are pulsed or bunched in any way. A detector that satisfies all of these requirements would truly be heaven sent.
The modern astronomer currently uses charge-coupled devices (CCDs) as the detector of choice, replacing the photographic plates of old. These arrays, typically of tens of millions of pixels, are similar, but obviously larger, than those used in the camera of your mobile phone.
Unlike your mobile phone, however, astronomical detectors are cooled to liquid nitrogen temperatures (around -2000C) in order to reduce noise. The effects of noise in a CCD are all too familiar to anyone who tries to take a picture with their phone at night in poor lighting conditions.
Another problem with CCDs is that they take time to read out (perhaps as much as a second or more) and, if they are clocked faster for higher read-out speeds, it comes at the expense of even more noise. Thus they cannot easily do very high time resolution astronomy, e.g. looking for intense but very short pulses.
A final problem is that there is no colour information if you use a single CCD on its own, unless a filter or spectrograph is placed in front of it. That said, CCDs have amazing sensitivity, almost 80-90 per cent of photons that hit it are registered. This compares with 2-3 per cent in the case of a photographic plate.
With the invention of CCDs, telescopes more or less overnight became 30 times better in terms of light-gathering power.
Just over a decade ago, however, Peter Day, a microdevice engineer at the California Institute of Technology’s Jet Propulsion Laboratory, and his colleagues  came up with a novel technique to detect photons that relied on superconductivity rather than the doped silicon of CCDs.
Certain elements, such as aluminium, and alloys like titanium nitride become superconductors if the temperature is lowered below a critical temperature (TC). TC varies depending on the substance, but is usually a few degrees above absolute zero.
The phenomenon occurs because once the temperature is low enough, electrons couple together through the positive charges of the ion lattice to form so-called Cooper pairs. These pairs, or more precisely the ensemble of pairs, then offer zero resistance, at least to a DC voltage.
Varying voltages however are a different matter. A coiled superconductor, just like a normal conductor, will have an inductance (L) due to the motion of the charges, a phenomenon known as kinetic inductance. In addition, separated superconductors, like normal conductors, can store charge, i.e. can be made into a capacitance.
Combining these two elements together, it is therefore possible to set up a superconducting LC circuit. And just like the ones in your transistor radio, each LC circuit is associated with a unique resonant frequency and bandwidth, related to its Q (Quality) factor. When the circuits are made similar in size to the pixels in your mobile phone, the resonant frequencies however are much higher than those in your radio, and are typically in the 1-10 GHz range.
What Peter Day and others , realised, however, is the importance of how such a circuit reacts to light. When a photon strikes the circuit, Cooper pairs are broken and so-called quasi-particles, i.e. free electrons are formed.
Of course, providing the radiation is not too intense, and the temperature is kept well below TC (typically at TC/10) these electrons quickly form new Cooper pairs, losing their excess energy as photons throughout the lattice.
The recombination time is of order a millisecond but during this period, the resonance circuit has a resistance and its inductance changes (normally the light, i.e. photons, are focused on the inductance element). The result is both a flattening, effectively a decrease in amplitude, and widening of the Q curve and a phase shift in its response.
All these effects can be measured. Moreover, because the photon energy required to break Cooper pairs is so small (around 10-3eV), a typical optical/near-infrared photon (with energy around 1eV) produces many quasi-particles and the number produced is clearly a function of the photon’s energy.
Thus, an individual resonating circuit can detect not only the time of arrival of a photon with high precision but also its energy (colour).
Of course, so far we have only explained how photons hitting one resonating circuit (pixel) may be read out but, in a working detector, an astronomer needs to read at least a thousand. How is this done?
Unique resonant microwave frequency
The secret is to design each pixel in the array so that it has a unique resonant microwave frequency. Obviously, it is a bad idea to have adjacent pixels with frequencies that are too close, to avoid cross-talk, but that is easily avoided.
Multiplexing is achieved by sending a range of frequencies (typically around 5-10 GHz) down a pair of wires to the array. Each pixel then only responds by resonating at its peak frequency and over a narrow band of at most a MHz. Clearly, one can readily address several thousand pixels in this manner.
As the system is reading data out very quickly, and there is a lot of data, processing is done in parallel using fast Field Programmable Gate Arrays (FPGAs). It turns out the type of technology used to read out these detectors is similar to that used in radio astronomy.
Radio astronomy today observes over very wide bands (several GHz wide) and over many channels. Obviously, the wider the band the better the sensitivity, but multiple channels are used to detect lines, for example from molecules in space, and to remove the effects of radio frequency interference (RFI).
This means we can use the high level technology currently under development for giant arrays of radio telescopes such as the Square Kilometre Array (SKA), as the basis of a read-out system.
The Dublin Institute for Advanced Studies, with support from Science Foundation Ireland, is in the process of building up the first MKID group in Europe specialising in optical/near-infrared detectors. Aside from using modified SKA technology, with assistance from the University of Oxford, to read out the arrays, we also plan on using a novel method to build the detectors themselves using a combination of titanium and titanium nitride (the material that gives your drill bits that golden colour) as superconductors.
The design of the arrays will be carried out in collaboration with Maynooth University and the detectors will be manufactured in the Centre for Research on Adaptive Nanostructures and Nanodevices (CRANN) in Trinity College Dublin with the assistance of the School of Physics.
While these arrays are being built with astronomy in mind, they can be used in a wide range of applications where sensitivity, high time resolution and spectral information are required. A major advantage is that while these arrays require cooling to very low temperatures, i.e. cryogenics are involved, all the electronics can be housed at room temperature.
Moreover, the flexibility and versatility of MKIDs mean they can be used in a host of environments. For example, an MKID array in space can reduce the number of moveable parts in a satellite by eliminating the need for filter wheels. At the same time, they can carry out multi-wavelength exposures more efficiently. The next few years should be exciting times, and we look forward to Ireland playing a lead in the development of these next generation detectors.
Author: Tom Ray, Dublin Institute for Advanced Studies
1. Day, PK, LeDuc, HG, Mazin, BA, Vayonakis, A & Zmuidzinas, J. ‘A broadband superconducting detector suitable for use in large arrays.’ Nature 425, 817–821 (2003).
2. Doyle, S, Mauskopf, P, Naylon, J, Porch, A & Duncombe, C. ‘Lumped Element Kinetic Inductance Detectors.’ J. Low Temp. Phys. 151, 530–536 (2008).
3. Cardani, L et al. ‘New application of superconductors: High sensitivity cryogenic light detectors.’ Nucl. Instruments Methods Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc. Equip. (2016). doi:10.1016/j.nima.2016.04.011
4. Broekema, PC, Nieuwpoort, RV van & Bal, HE. ‘The Square Kilometre Array Science Data Processor. Preliminary compute platform design.’ J. Instrum. 10, C07004–C07004 (2015).