Tom Eddershaw finds that improvements in power of terahertz lasers are making them more attractive as industrial products
In February, a team from the University of Leeds in the UK announced in the journal Electronics Letters that they had produced more than 1W of terahertz radiation from a quantum cascade laser, more than doubling the previous power output.
The widespread application of terahertz radiation has been impeded by problems in making the lasers powerful and compact enough to be useful, but the Leeds result offers hope that these obstacles are now gradually being overcome.
Edmund Linfield, professor of terahertz electronics in the university’s School of Electronic and Electrical Engineering, said: ‘The excitement, for us, is threefold. It’s the international aspect for UK-based research, the potential for uptake from industry, and the ability for people to carry out new experiments. This has allowed us to work with international organisations and partners, which is very exciting. Having 1W power also has the potential to attract industrial players and start to get the concept into a more industrial product. From a basic science point of view, high-powered sources open up regions that were previously unavailable.’
He acknowledged that: ‘Although it is possible to build large instruments that generate powerful beams of terahertz radiation, these instruments are only useful for a limited set of applications. We need terahertz lasers that not only offer high power, but are also portable and low cost.’ The quantum cascade terahertz lasers developed by Leeds are only a few square millimetres in size. Linfield explained that there are two main benefits to increasing the power: a stronger signal-to-noise ratio can be collected and observations can be made from further away.
Don Dooley, general manager of Gentec Electro Optics, said: ‘Higher power and higher energy will always attract attention by enhancing performance in certain areas. It could either enhance an existing application or open up new avenues of applications.’
Dooley continued: ‘When we first got involved in this, around seven years ago, most of the customers were from universities doing terahertz research. They were either developing sources or sensors. However, over the past few years our sales have been split between research and industry. This is a sure sign of an evolving market. It parallels the growth of the x-ray market years ago; it started more as a lab curiosity, but they are realising they can do some pretty incredible things with it.’
Dooley pointed to the advantages of the technology: ‘The beauty of terahertz radiation, and the reason it will replace x-rays in certain areas, is that it is non-photoionising. It doesn’t damage materials.’ This has benefits for medical fields, industrial inspection, and security that typically use x-rays, which is photoionising and can damage biological tissue. As Linfield stated: ‘In terms of causing damage, terahertz frequency is much less energy per photon, which means it doesn’t cause as much damage on collision with a surface. Each photon is probing the material; more power means more photons means better signal.’
A major practical application of terahertz radiation is in airport security checks, where the detection technique is time-domain spectroscopy or broadband spectroscopy. Ben Agate, laser sales and service engineer at Photonic Solutions, described a German project to try to surmount the difficulties of portability of terahertz technology by building a more mobile terahertz detector system. The project, carried out between the University of Marburg and Menlo Systems, uses time-domain spectroscopy. The system is able to analyse material, even through a bottle, and indicate whether a substance is safe or not in a matter of seconds. Agate explained that the Marburg-Menlo system ‘is a handheld device connected by a fibre to a laser source that is on wheels. So it is effectively portable where the handheld device emits and receives terahertz radiation.’
However, the way in which the Marburg-Menlo system generates the terahertz radiation differs from the method used by the Leeds University team. At the present moment, Agate pointed out, ‘one of the most common way of creating terahertz radiation is by using an indium gallium arsenide chip pumped with ultrafast pulsed lasers. The Menlo system uses a 1,560nm fibre femtosecond laser, producing less than 100mW. Chips can work with a variety of different lasers.’ Menlo is working with a Fraunhofer institute developing a photoconductive antenna. These antennas allow the user to pump a chip with a laser to produce terahertz radiation. The low temperature antenna is typically made of gallium arsenide for 800nm signals or indium gallium/aluminium arsenide for 1,560nm. These lasers typically produce low power output, but cover a broad spectrum. The quantum cascade lasers used by the University of Leeds produce higher power with a smaller spectral range.
Linfield explained that the team at Leeds had achieved the high power output by improving the semiconductor material quality and uniformity. This meant that all the parts of the structure produce as much light as possible.
To improve the quality of the material, the team grew the semiconductor in a complete vacuum. This reduced the risk of contamination of the cell and allowed the team to take their time in building up the layers of gallium, aluminium, arsenic, and silicon, and allowed very accurate positioning of wells and barriers for the electrons to manoeuvre. These wells and barriers create artificial energy levels when the electrons pass through the material, releasing photons with each step. Linfield explained: ‘For every one electron going through the material, you can get as many photons out of it as there are wells.’ So, by placing the many layers very accurately on top of one another, a higher quality material, and higher power output, can be achieved.
‘That’s half the story,’ he said. ‘The actual device processing, the dimensions, and the packaging itself, are the other half.’
There are thermal issues with terahertz lasers. Linfield explained: ‘It’s not a very efficient process. This is partly because there is a large amount of heat that you can’t dissipate effectively. Driving the efficiency up would obviously be good for the commercialisation of the technology and the less heat you have to deal with, the better.’ However, it is not the quality of the semiconductor that Linfield believes is the issue: ‘You mount the chip on some kind of holder and it’s actually down to how efficiently you can extract this heat from the active region. You have a substrate that has a few microns of semiconductor on it; now that is a resistance to heat getting out. We’re looking at making that better to improve efficiency.’
These issues with heat are also what make manufacturers and the researchers so keen to keep the lasers small. ‘As they get bigger, you get more power but you have to make compromises,’ said Linfield. Agate agreed that, from a commercial point of view, the size is very important: ‘It’s all very well having a high-powered system, but it’s difficult to make it relevant in the real world environment, if it’s not portable.’
Even though terahertz sources have been around for some time, there is still an issue with measurement standards. ‘On the sensor side of things, calibration is a big issue,’ said Dooley. ‘For other laser types there are many detector calibration standards that cover the UV through the far IR. However, for the terahertz spectrum, there are very few calibration standards.
‘When terahertz laser manufacturers come to us, they want absolute power or energy measurements and we have to say “here is the state of the art in terahertz calibration, we can offer reasonably low uncertainty (+/- 15 per cent) in the 30 to 0.7THz range, but beyond that range the best you can hope for is a relative measurement”.’
The Physikalisch-Technische Bundesanstalt (PTB) in Germany offers calibration of terahertz sensors and has done so for around two years. But as Dooley explained, it is the only one in the world to offer this. He added: ‘PTB has recently expanded the spectrum to cover 1THz to 5THz. However, this is discrete wavelength calibration.’
In the USA, a team at the National Institute of Standards and Technology (NIST) is developing a terahertz calibration standard at the moment and plans to offer calibration at one or two wavelengths. NIST has now developed several terahertz calibration standards and is in the process of determining their uncertainty. Once this work is complete, which Dooley predicts could be later this year, NIST will offer calibration to industry.
Gentec-EO and others are trying to reduce calibration uncertainty. One way to improve calibration is by coating the surface of the detector with a non-reflective material. New coatings are being tested which reduce the reflection of terahertz radiation from the detector surface, and so improve the accuracy of a power reading. Dooley explained that ‘what you are shooting for is something that has a very flat absorption over the widest spectrum. This is so you can calibrate at a signal wavelength, and then offer correction factors that have low uncertainty over a good part of the terahertz spectrum.’