Highly sensitive detectors are enabling techniques like quantum cryptography and forms of fluorescence spectroscopy. Tom Eddershaw investigates the technology for detecting individual photons
On 24 April, a team of researchers from Toshiba Research Europe, British Telecom, ADVA Optical Networking and the UK National Physical Laboratory announced it had transmitted both a quantum encryption key and useable data down the same fibre cable. The trial is one of the latest advances in Quantum Key Distribution (QKD), a form of data encryption based on quantum mechanics that is much more secure than existing encryption methods.
Quantum cryptography, while still in its infancy, is gathering a lot of interest – the CLEO 2014 event in San Jose, USA will hold a number of talks on the topic. It is a technology that relies, to a certain extent, on the sensitivity of detectors to observe the stream of randomly polarised photons used in QKD.
Andreas Bülter, applications specialist at PicoQuant, said: ‘A couple of companies sell working [quantum cryptography] devices, but they are still only used for very specific purposes – for general use the data throughput is still low and you have to solve the problem that the current systems need dedicated point-to-point networks.’
The low data throughput means the technology currently takes a long time to send a message. Bülter noted that one way to improve this would be to have detectors with higher detection efficiencies, lower dark counts and higher temporal resolution. The photon counting electronics are not really the limiting factor. ‘These days you can get electronics that have temporal resolutions of around a picosecond; so, they are effectively extremely fast stopwatches,’ he said.
He added though that going faster is not really necessary because the best detectors operate at around 20ps – the timing electronics are often faster than the detectors.
The types of detectors used in quantum cryptography are those that are able to distinguish between individual photons. PicoQuant’s vice-president of sales and marketing, Uwe Ortmann, described the detector technology: ‘Photon counting is more sensitive than analogue technology. Analogue measurement needs multiple photons to generate a signal; photon counting needs only a single photon. It’s basically the most sensitive system because it reacts to a single photon. You don’t need hundreds of photons to get a useable system.’
Photomultipliers or avalanche photodiodes are two common detectors that can be used for single photon detection. The principle is relatively simple: a photon hitting the active area of the photodetector will be converted into an electron, which is then amplified into a measurable pulse signal.
A photomultiplier has a dynode structure in which one electron is multiplied leading to a larger, measurable electrical pulse. Photons striking an avalanche photodiode, meanwhile, cause an avalanche effect in a semiconductor material that generates the measureable pulse. For both, Ortmann said: ‘One photon of light hits the photoactive area and under a certain detection efficiency is able to generate an electron. This is amplified inside to create an electrical pulse that could be measured as a single event.’
This means the devices need a high quantum efficiency in order to make the most of the limited light source available. Elliott Chick, a technical sales engineer at Laser Components, explained that the quantum efficiency is dictated by the semiconductor material’s ability to absorb a particular wavelength: ‘Essentially, the quantum efficiency is the percentage chance that a photon will be successfully converted into an electron event. The quantum efficiency at different wavelengths is what discerns the wavelength of light that a detector is suited to.’
Different detectors will have absorption peaks at different wavelengths. However, Chick gave the example of Laser Components’ Count single photon counting module, which would detect an NIR photon even if this is not the peak efficiency of the device. ‘This [the peak efficiency] is normally governed by the structure of the semiconductor and the complex electronics behind the sensor. You know that a photon will give a certain amount of energy to an electron, so you can tailor your bandgaps to give better efficiency,’ he said.
Keeping dark count down
One issue, especially with photon counting detectors in the infrared, is the phenomenon known as the dark count rate. Bülter explained: ‘The number of photons observed increases due to interference from the detector itself in the form of thermal noise. If you have this noise providing 100,000 counts per second while your assembly gives you 1,000 counts per second, you end up measuring a lot of noise and this is where you run into problems. What people have to do in the infrared is design detectors with very low dark counts which usually results in cooling. Certain devices, such as nanowire detectors, provide very low dark counts, but have to be run at around 1K, so liquid helium temperatures, which is extremely expensive and often hard to use.’ He also said: ‘Other detectors can be operated with just peltier cooling, but then these detectors suffer from other problems such as low detection efficiencies.’
He added that this is why 85 to 90 per cent of photon-counting applications operate in the visible spectrum, and only very advanced devices run in the infrared.
Richard Biggs, a sales engineer at Hamamatsu Photonics UK, explained that within the dark count rate there is also random noise that cannot be subtracted out. Instead, cooling the device or increasing the threshold of detection should be considered to reduce the dark count.
Laser Components provides its InGaAs-based Count Q detector, which has a large heat sink fan and thermo-electric cooler built-in. Chick said: ‘The Q is designed for delving into commercial applications, or a small-scale lab that wants to do a free-running experiment without having to take on the expense and materials required for the super-cooled devices.’
Fluorescent lifetime measurements
Quantum key distribution is not the only field that benefits from single photon detectors; life sciences and many other fields can all use their ability to convert a photon into a readable electrical current. Bülter stated: ‘Because we are working with single photons, any applications that work with low light intensity can use it [single photon detectors]. In these instances, where you can’t work with analogue detectors, you need to go to photon counting. Its uses start in fundamental science, such as fluorescence lifetime measurement, and the devices are used through into medical applications.
‘Photon counting is essentially a binary method; it says “is the photon there, yes or no?”. It’s a zero or one answer,’ Bülter continued. ‘If you want to know how many photons are present per second, you count the amount of signals there are per second.’
Ortmann pointed out that photomultipliers have been around for 30 years, as has photon counting. However, now the detectors can do more than purely count the number of photons present. ‘In the past it was only used for counting, measuring the amount of the light. Nowadays, because the timing electronics have become simpler and easier to use, you can now measure more aspects of the photon,’ he said.
Biggs, of Hamamatsu Photonics UK, explained that it is desirable for the manufacturer of such devices to make photon detection efficiency high, but the recovery – or quenching time – as low as possible in order to make rapid, accurate responses. However, he said: ‘Detection efficiency increases with the size of the pixels, but the pixel capacitance also increases with pixel size, giving a longer quenching period. Big pixels give photon counters a high detection efficiency, while small pixels give a faster time response.’
However, Bülter explained that there is no golden detector that gives you good temporal resolution, high detection efficiency, and low dark counts over a large wavelength range. ‘There is always a compromise that needs to be made. You can get a detector with high detection efficiency, but it would have low temporal resolution, or vice versa. If you focus on getting the best temporal resolution you may lose out on other aspects; it’s dependant on the application.’
Chick said Laser Components had turned from being a distributor to a manufacturer of single photon detectors after noticing a number of customers complaining about the high dark count rates. The company builds its own count single photon detector range, which uses the company’s own single photon avalanche diodes (SPADs) with low dark count fluctuation.
‘People are often interested in the intensity from an excited sample, and that’s not a problem,’ said Ortmann. ‘The second part is what kind of detail is there after an excitation – e.g. how long is the energy stored in a chemical compound. This is part of the information you can get out of the photon, which is done with time-resolved photon counting. And that gives you another side of the measurement; it’s not just the pure intensity.’
One such application is fluorescent correlation spectroscopy (FCS). Ortmann explained that, here, a molecule is fired through a laser focus in a microscope causing it to emit a burst of light. This light is detected by a photon counting detector and is used to measure diffusion times. There is also now a new trend called lifetime FCS, which is much more accurate and involves measuring the fluorescence lifetime of the sample. Also, with the technique, the device becomes a lot less vulnerable to artefacts given off by the machine itself.
Ortmann said that measuring just the intensity of the photon is a small fraction of the data that can be collected from an event. ‘If you are looking at the lifetime, you can look at the full information content [from an event].’ This gives more information about the chemical compound or molecule being studied.
Bülter summed up the technique: ‘Basically, it’s not just looking at “if” there was a photon, but “when” there was a photon. If you know the “when”, you immediately know the “if”. This timing information can provide a great deal more information.’