9 September 2014
The Cobalt Light Systems team receiving the MacRobert Award gold medal in July 2014 (from left to right): Dr Guy Maskall, Dr Craig Tombling, Dr Paul Loeffen, Professor Pavel Matousek, Stuart Bonthron
In 2006, UK police and security services foiled what could have been the most significant terror attack in the West since 9/11. A group of men planned to blow up at least 10 airliners using a mixture of liquid explosives smuggled on board inside drinks containers. The long-term effect of the event was the introduction of hand-luggage restrictions, still in place today, that prevent passengers from taking more than 100ml of any one liquid onto an aircraft.
That could soon be set to change, thanks in part to the invention of a technology that can identify a liquid inside a sealed container in seconds. Airport staff can now spot any potentially dangerous substances without needing to take a sample, leading to reduced queues at security checkpoints and greater convenience for passengers. The UK-built system, known as the Insight100, has now been rolled out at 65 European airports including Heathrow and could mean that current liquid restrictions are lifted in the coming years.
The same technology is also used to confirm the identity and perform batch release of pharmaceutical raw materials and products. One day, it could also enable scanners to diagnose breast cancer and bone disease. Its creators at Oxfordshire firm Cobalt Light Systems – a spinout of the Science and Technology Facilities Council’s Rutherford Appleton Laboratory in Harwell, Oxfordshire – was earlier this year awarded the Royal Academy of Engineering’s MacRobert Award, beating fellow finalists Rolls-Royce and Qinetiq.
Cobalt’s system relies on a scientific technique that was developed 80 years ago: Raman spectroscopy. When a beam of light hits a material, a small number of photons interact with its molecules and either gain or lose energy – the Raman effect. Every material produces slightly different energy changes, so by measuring this ‘scattering’ effect it is possible to identify the substance being tested – see Raman spectroscopy.
This method of classifying materials took off after the invention of the laser in the 1960s, making Raman spectroscopy much more practical. But it was still limited to identifying the molecules on the surface of an object; a conventional Raman spectrometer can’t detect any materials behind or inside an opaque or fluorescent barrier. Part of the challenge with using the technique is the weakness of the Raman effect: roughly 1 in 100 million photons are Raman scattered.
Then, in the early 2000s, Professor Pavel Matousek and his colleagues at the Rutherford Appleton Laboratory’s Central Laser Facility made a breakthrough. The team was using lasers to try to understand exactly how energy was transferred between molecules in certain ultrafast chemical reactions that took place in biological processes in solution, for example, processes related to photosynthesis or the ionisation of DNA. To do this, they wanted to use a very short laser pulse to take a spectroscopic snapshot of what was happening to the molecules every picosecond – a period equivalent to one millionth of one millionth of a second.
To make this possible, they also needed a way of blocking out the background energy signal created by fluorescence from the molecules, which occurs predominantly after the Raman effect but swamps its much weaker signal. The researchers hit on the idea of using an optical device known as a Kerr gate, which is operated by a second, very intense laser pulse, to act as a shutter that could open and close fast enough to let through the instantaneous Raman-scattered photons but block the following fluorescence.
When they began employing this technique for chemical company ICI to analyse powdered chemicals, they encountered a puzzling phenomenon. A conventional Raman spectrometer is only able to detect scattered photons when its laser is turned on. But the faster Kerr-gated device was able to see that photons scattered by the powder were visible for 1,000 times longer than expected once the laser was deactivated.
The team realised that some laser photons were bouncing around inside the material before they were re-emitted back towards the spectrometer, increasing the proportion that were in elastically scattered and delaying their re-emission and detection. By taking their snapshot of the signal slightly later, they were able to capture photons that had penetrated deeper into the sample and interacted with molecules below the surface. They now had a way to use Raman spectroscopy to identify a material behind an opaque barrier, as long as it let some light pass through it.
However, the Kerr gate method required a high-intensity laser the size of an apartment, which was a major obstacle to creating a practical system. The problem became more apparent when they collaborated with Professor Allen Goodship, from University College London, and Professor Michael Morris, from the University of Michigan, who wanted to see if the technique could be used to probe through living tissue and detect bone disease. Although they were able to prove it was possible on mouse tissue, they were not able to move their machine to a hospital to do so, encouraging them to think about how to make the technique easier to carry out.
Simplified diagram of how a Raman spectrometer works. A sample is irradiated with monochromatic laser light, which is then scattered by the sample. The scattered light passes through a filter to remove any stray light that may have also been scattered by the sample. The filtered light is then dispersed by the diffraction grating and collected on the detector. This setup works for both the nonresonance and resonance Raman techniques© http://chemwiki.ucdavis.edu
In 1928, Indian physicist Dr C V Raman made a discovery that would win him the Nobel Prize for Physics two years later. For seven years, he had been trying to explain why the sea was blue after proving that it was not simply a reflection of the colour of the sky. His theory was that the water molecules interact with some of the light from the sun and effectively change its colour to blue.
This led him to discover the phenomenon that came to be known as the Raman effect – although several other scientists also independently predicted or demonstrated it. Photons are often ‘scattered’ when they come into contact with molecules in a way that changes their direction but not their energy. But a small number of photons are ‘inelastically’ scattered, meaning that, when they interact with the molecule’s electrons, the photons either gain or lose energy, which alters their wavelength and hence their colour.
Raman showed that the amount by which the wavelength changes is dependent on the type of molecule that the photon interacted with. Other scientists then began recording how much different substances would scatter light in this way, creating a catalogue of data that could be used to identify a material by measuring its Raman effect, a technique known as Raman spectroscopy.
Today, this involves firing a laser at a sample of material and recording the energy shift of the reflected photons on a graph, with peaks along this spectrum representing the molecular vibrations. As different molecules are known to produce a unique set of peaks, each spectrum is like a molecular fingerprint that pinpoints which molecules the sample contains and in what proportions.
In 2004, Matousek realised that some photons travelled sideways through the material – like molecules diffusing in a solution – before emerging. (The effect is similar to shining a point of light through a human finger and causing the light to scatter through all the tissue). This light also contained a considerably larger proportion of photons that had been re-emitted by sub-surface molecules. In practice, this meant that by scanning an area of the material away from the point where the laser was shone, the researchers could detect photons that had travelled below the surface, been scattered by the molecules there, and re-emerged in a different place. This move from a high-speed temporal technique to a simple spatial technique reduced the laser requirement to one the size of a mobile phone.
This Raman signal also contained photons re-emitted by surface molecules. To overcome this, the team took two readings – one at the point of laser entry where only photons scattered by the surface were detected, and used that to cancel out the unwanted signals from a second reading of the subsurface photons, typically taken a few millimetres away from the first. Instead of capturing Raman signals at different temporal intervals, they were taking readings at different spatial locations by moving the laser slightly each time. Therefore, the technique became known as spatially offset Raman spectroscopy (SORS).
Spatially offset Raman spectroscopy. Illustration of two measurements being made, zero and offset where the laser is displaced from the point of measurement. In the zero case, the Raman signal is dominated by signal from the surface of the container. In the offset case, the signal is dominated by photons that have been scattered in the container contents
By taking just two measurements, it is possible to identify a substance behind a thin, partially opaque barrier, for example a coloured plastic bottle. But by taking 20 or more measurements in slightly different locations and using more complicated algorithms based on a technique called multivariate data analysis to cancel out the unwanted signals from each one, it is possible to build up a picture of the multiple layers of a sample more than 1cm in depth. This means that the technique can be employed to study complex systems containing numerous different types of molecule on different levels, such as biological tissues.
With the basic technique established, Matousek’s team began examining ideas for applications in more detail using funding from the technology transfer office STFC Innovations, focusing particularly on the pharmaceutical industry. The new method immediately gave them some advantages because it was much simpler than its temporal gating predecessor and so required only tabletop-sized lasers with much lower intensities, making it more practical and, crucially, safe enough to use in a commercial or clinical setting.
As the work progressed, funding from the Rainbow Seed Fund, Oxford Technology Enterprise Capital Fund and the charitable foundation NESTA was secured. This eventually enabled the group to formally spin out a company (which was originally known as LiteThru) in 2008, attracting more investment including Longwall Ventures. Matousek remains chief scientific officer of Cobalt, but, in 2009, the firm hired Paul Loeffen as CEO to lead a growing technical and business development team. Loeffen came with both a technical background from his time working as a physicist for CERN and ESRF and experience of founding and growing an instrumentation spinout, Oxford Diffraction, which was sold in 2008 for $50 million.
While the pharmaceutical industry provided the first applications for Matousek’s work, Loeffen and the Cobalt team then spotted another opportunity in airport security restrictions. After a conversation with the Home Office about the potential of the technology, the company was invited to apply for funding to develop an airport-specific product through a competition framework in 2010.
This innovative research call, which was the second such initiative by the Home Office to develop new technologies for use in explosives and weapons detection, was not like a typical competition or product brief. The entrants did not know the criteria by which they would be judged. This was partly due to the confidential nature of the European Commission’s requirements for liquid explosive security at airports, which were still being defined at the time.
The Insight100 is a table-top screening device for evaluating the chemical composition of liquids,powders and gels in glass or plastic containers. It is used at airports, other passenger hubs and high security establishments © Cobalt
One element of the technology the firm knew it would have to improve from a commercial standpoint was the time it took to scan a container. Initially this took several minutes, but by improving the system’s optics and data processing, the developers initially brought this time down to under 10 seconds. They also needed to develop the equipment into a more user-friendly format: the original work had been performed in a darkened room while wearing laser goggles, with data manipulation carried out manually over several hours.
The company was a winner in the Home Office competition and was contracted to develop the technology for testing at airports. Once this was deemed a success, the prototype was finally turned into a commercial product and released onto the market. The device that Cobalt finally produced, known as the Insight100, comprises a tabletop-sized cabinet into which drinks bottles or other plastic or glass containers can be placed. An adjustable arm is then positioned below the level of the liquid or gel inside and the door closed. The SORS system scans the container and attempts to match its contents with a library of potentially dangerous substances, such as hydrogen peroxide, providing a unambiguous pass or fail result to the operator in five seconds.
A major aspect of development was determining the limits of the system. The principle of SORS enables the technology not just to classify different materials but to specifically identify them. But Raman signals are inherently weak and the company had to work out how close a particular measurement had to come to a model for a dangerous substance in order to be confident they had a match. This required extensive study of many lethal target materials in a wide variety of conditions and containers. The company now claims its SORS technology can find specified materials close to 100% of the time and produces false positive results on less than 0.5 per cent of occasions (a large benefit when considering passenger flow models for minimising queues).
To create a viable product, the Cobalt product development team had to find a way to automate both the laser system and the signal processing, with the latter providing a particular challenge simply because it had never been done for SORS before. In addition, the laser is not inherently eye safe and has to comply with regulations through being enclosed in an interlocked housing; the system also had to achieve formal approval from the European Civil Aviation Conference through a test compliance process. Improvements in the Raman optical collection efficiency and the detection algorithms finally led to a five-second cycle time for the Insight100.
Early investigation of Spatially Offset Raman Spectroscopy under eye-safe conditions in laboratories at STFC by Professor Pavel Matousek and Dr Kate Ronayne
The system has been in use since January 2014 at airports across Europe, including the major hubs Heathrow, Paris Charles de Gaulle and Amsterdam Schiphol. These operations are paving the way for passengers to be allowed to carry larger containers once more if the new regime being examined is deemed successful.
The success of the Insight100 and Cobalt’s other products has seen the company grow to employ more than 40 people in a science business park that is home to 200 high-tech firms. Now the company carries out not only research and development but also production, performing final assembly and performance verification of the systems. Cobalt has also developed a system for confirming the contents of sealed containers of pharmaceutical manufacturing ingredients: RapID. This raw materials verification is a requirement at all pharmaceutical production plants. The company is now looking at how SORS can be used by customs officers to screen suspicious packages and identify counterfeit alcoholic drinks or medicines. Matousek is now engaged in early-stage scientific work applying the technique to detecting breast cancer and bone diseases such as osteoarthiritis or osteoporosis, opening up the possibility of screening and earlier diagnoses that don’t require biopsies or even X-rays.
For more information visit www.cobaltlight.com
Professor Pavel Matousek is Chief Scientific Officer of Cobalt Light Systems and a Science and Technological Facilities Council Senior Fellow. He is a Fellow of the Royal Society of Chemistry, a visiting professor at the University College London and a co-founder of Cobalt Light Systems.