Andy Extance investigates the new sources that could broaden adoption of terahertz imaging
As of today, terahertz imaging and sensing is yet to fully live up to the hopes that potential users in security applications might have had for it. The term terahertz became popular among spectroscopists referring to the electromagnetic spectrum between the infrared and microwave in the 1970s. Interest from researchers in using technology producing and detecting such light for identifying hidden threats grew through the 2000s, as better terahertz sources became available. The promise they saw is today being exploited in a few airport scanners – but that’s only a very narrow niche. However, higher-power sources have entered the market; could they drive broader adoption?
Terahertz’s appeal comes because it interacts with matter differently to other types of light. Whereas infrared, for example, induces molecules to perform bending and stretching motions, terahertz light causes collective motions of groups of polar molecules like water. Consequently, it’s potentially great for detecting what others want to keep from you. Because it’s not absorbed by non-polar materials like cardboard and clothes, terahertz offers similar capabilities for imaging what’s within things to X-rays, but without the health risks. Explosives, chemical or biological weapons do absorb terahertz light, giving unique spectroscopic ‘fingerprints’.
Yet to be practical, security systems have to be fast, reliable, robust, and reasonably inexpensive. Combining all these requirements has proven ‘difficult to achieve’, explained Anselm Deninger, director for terahertz technologies at Toptica Photonics in Munich, Germany. ‘With most commercial terahertz systems it takes at least a minute to record a high-quality spectrum,’ Deninger said. ‘If you want to scan hundreds of envelopes per hour, this is clearly too slow.’ Cost is also not amenable to wide deployment, he added.
Consequently, the significance of the non-destructive testing market for Toptica’s terahertz systems is greater than the market for security applications. ‘Via time-of-flight techniques, pulsed terahertz systems can quantify the thickness of paint layers, or wall thickness of plastic pipes or bottles,’ Deninger said. ‘This is a much more dynamic field right now. I do believe that the terahertz market will grow. One might debate whether or not the growth will be driven by defence and security – in my view, likely not.’
Toptica produces two complementary terahertz system brands: TeraFlash and TeraScan, with maximum output power of 65µW. TeraFlash is a pulsed system, based on the company’s FemtoFErb, a 1,560 femtosecond pulsed erbium fibre laser. The laser pulse is split in two; one part travels to a semiconductor-based terahertz emitter, creating a terahertz beam that then interacts with a sample before travelling to the detector. The other part serves as a ’readout’ pulse at the detector where it samples the incident terahertz field, much like a sampling oscilloscope does. Deninger noted that Teraflash can offer either very broad bandwidth, covering frequencies as high as 6THz, or speed, recording a complete spectrum in 20ms. ‘Of course, we trade measurement speed for spectral bandwidth – this is true for any terahertz system, but even at maximum speed, we still obtain an impressive signal,’ he said.
TeraScan combines beams from two distributed feedback (DFB) semiconductor lasers, obtaining continuous wave terahertz light whose frequency is the difference between the two input lasers. Its main advantage is the spectral resolution, distinguishing absorption lines down to single megahertz amid the terahertz range. ‘In both systems, the dynamic range of the terahertz power is very high,’ Deninger said. ‘Thus, you will still see a signal in the case of highly absorbing samples, within physical limits, of course.’
Toptica’s DFB lasers were used in a terahertz system designed to identify toxic chemicals by Goodrich ISR Systems in Danbury, Connecticut. ‘This worked well, but for reasons not disclosed to us, the project was discontinued,’ Deninger commented. Now, his company is taking part in a German consortium – including the city of Mannheim’s fire brigade – looking at trace gas detection. ‘They are keen to identify gases released in an industrial disaster, such as a fire in a factory. This has a direct impact on the protective gear firefighters need. We still need to work out how we best bring gas samples to the spectrometer, but the new TeraScan seems to be a great instrument, owing to its signal quality, spectral resolution, and frequency repeatability. The project has already detected gas on the parts-per-million level.’
The need to bring samples to the spectrometer highlights an inherent challenge facing this technology. ‘Terahertz light is strongly attenuated by water vapour – omnipresent in air – so remote sensing simply does not work: after a few metres, only selected “window frequencies” survive,’ Deninger explained. This is not sufficient for ‘stand-off spectroscopy’ applications detecting threats at a distance, he added. ‘I have talked to people who wanted to detect buried land mines with the help of terahertz light, and I had to tell them terahertz rays will not pass through soil, so this does not work either. Applications that I consider realistic include detecting trace amounts of toxic gases in public places, buildings, or subway stations. Also, the spectroscopic analysis of mail envelopes seems feasible. This does not require extended path lengths, and paper is reasonably transparent, so one might be able to check envelopes for explosives, or illicit drugs inside.’
More power to them
Alan Lee, who co-founded Mountain View, California’s Longwave Photonics in 2010, agreed that security and defence applications using terahertz spectroscopy technology, although interesting and very promising, ‘are still quite limited’. ‘For now, the need for high resolution terahertz spectroscopy is still mainly driven by laboratory and industrial internal R&D,’ he said. ‘However, with the maturation of terahertz technology and with more researchers adopting our high-quality Easy-QCL source, I can see a strong need might emerge in this area.’
Longwave’s semiconductor quantum cascade laser (QCL) is currently a test platform for research and development, enabling applications in the 2 to 5THz frequency range to be explored. These include homing in on the narrow frequency windows where terahertz light does propagate through atmospheric moisture. ‘One large challenge is producing lasers that operate in these windows, which we can do by creating DFB lasers for specific frequencies,’ Lee explained. ‘We’ve been able to develop DFB devices that have milliwatt average power levels with nice beam patterns and single frequency operation. Milliwatt power is also sufficient to use room temperature direct detectors like pyroelectric detectors and microbolometer focal plane arrays for real-time imaging. Otherwise, terahertz sources that produce microwatts of average power must typically use either a more complicated heterodyne detection technique or a high-sensitivity liquid helium cooled detector.’
However, confining the electrons responsible for emitting the terahertz light in semiconductor QCL structures also requires very low temperatures. ‘To date, the maximum operating temperature of a terahertz QCL is 200K, but they really work best below liquid nitrogen temperatures of 77K,’ Lee explained. Yet Longwave packs its lasers in compact pulse-tube cryocoolers that provide closed-cycle refrigeration, without needing cryogens like liquid nitrogen or even – they claim – maintenance. ‘We’ve made these systems to be flexible so that researchers can exchange laser modules to access different parts of the 2 to 5THz frequency range, or take advantage of DFB or even higher power Fabry-Perot devices,’ Lee added.
This technology allows spectroscopy at a distance, such as in remote sensing of atmospheric gases and observing emission from molecular gases in astronomy. ‘We recently prepared a device that was made at MIT for use on the STO-2 NASA balloon mission in Antarctica,’ Lee said. The final instrument, assembled by the Netherlands Institute for Space Research (SRON), will use a 4.74THz QCL to provide some of the first observations of emission from neutral oxygen.
Lee is optimistic that similar capabilities can help in security applications. ‘I would argue that the ultra-high frequency resolution and high power of our terahertz source could lead to an advanced buried explosive detection system,’ he said. ‘Not only could it generate a binary, true or false result, but the ability to do fine spectrum analysis would reveal a chemical fingerprint which is crucial for explosive identification.’
Working on the terahertz image
Higher-power QCLs could also enhance the prospects of replacing X-ray security imaging, according to Pierre Gellie co-founder of Paris, France’s Lytid. Lytid, which was spun out of Paris Diderot University in 2015, also offers a milliwatt-output ‘TeraCascade’ QCL. It emits specifically at 2.5THz, but provides both continuous wave and pulsed operation from the same system. Higher power sources can illuminate more pixels on a detector, Gellie emphasised. ‘That’s very interesting in imaging – you could perform several million measurements per second,’ he said. ‘That could enable terahertz cameras producing several hundred thousand signals at one time and several tens of images per second. Eventually scanners could work quickly.’
Closed-cycle, ‘maintenance-free’ refrigeration is also included in Lytid’s TeraCascade QCL systems. ‘Our market study shows people want a terahertz source with higher power that is very easy,’ Gellie said. ‘We’ve worked on integration and user-friendliness a lot – it’s easier to use than a smartphone. It’s one press on the touch screen, you wait 20 minutes for it to cool down and then you’re ready to go. You have all you need in a single box – a cooler, all the electronics to drive the cooler and the quantum cascade laser. That’s why it’s bulkier than just a diode laser – and also makes the system more expensive. Obviously it has to be as transparent as possible for the end user in any real-world application. This is what we’re aiming for, and what’s missing in other sources on the market.’
Having been established so recently, Lytid is initially targeting the better-developed industrial non-destructive testing market, where it provides real-time terahertz imaging. Currently that’s the greatest interest in TeraCascade from the defence industry, Gellie explained. ‘It’s a tool for thickness measurements, finding faults in very high grade designs in aerospace engines, and also on armoured vehicles. Of course, going towards industrial applications, having a fully integrated system, the most reliable components are a must-have. But they don’t need to worry about the source any more – they can focus on their application.’ TeraCascade won a Prism Award in the Scientific Lasers category at Photonics West 2016, Gellie added.
TeraCascade’s reliability would be well suited to use airport scanners, Gellie suggested – although even at the milliwatt level it doesn’t yet have the necessary power. ‘Airport scanners now use millimetre wave technology that provides low-resolution image patches and can only detect small areas,’ he explained. Consequently, security applications either require manual scanning or systems using emitter and receiver arrays. ‘With terahertz you could have much higher resolution – you could actually see the proper shape of the object,’ Gellie explained. ‘But nowadays it’s difficult to implement; there are few sources, they are not powerful enough and also detectors are not quite sensitive enough. You’re talking about needing 1W of power with actual receiver technology today. If they improve too you might be able to do something with a few tens of milliwatts of power.’
Lytid is working on power improvements, but the need for better detectors highlights the key barrier to broader uptake. ‘For terahertz technology to go mainstream you have to do more on components and the whole system,’ Gellie said. ‘There are no terahertz optical fibres right now, there probably won’t be any time soon. We hope that there are some advances still to be made. ’ Similarly to how Lytid came to commercialise the TeraCascade, he feels that those advances are most likely to involve technology transfer from academic labs.
Producing systems that operate at other frequencies is another area that Lytid would like to explore. Gellie believes this would take terahertz imaging in a highly desirable direction for security use. ‘Using different frequencies you can go towards spectroscopic imaging,’ he explained. ‘That would be the holy grail for this application, not just being able to detect what kind of object is hidden, but also the chemical composition, for example finding explosives and drugs. This has been shown in academic publications – it’s still in an infant stage, but it’s very promising. It will come.’
Andy Extance is a freelance science writer based in Exeter, UK