FEATURE

Food for thought

Portable spectrometers are being used in more ways than ever to analyse food and, this year, consumers will even be able to purchase handheld spectrometers to check food before they buy. Jessica Rowbury looks at how these spectrometers are being used to test food

In the food industry, spectrometers are not just coming out of the laboratory into the processing plant but even into the shops and supermarkets. Miniaturisation and better manufacturing methods have allowed small, robust, and cost-effective instruments to carry out analysis at different stages of food production without having to take samples to a laboratory. One such product, Ocean Optics’ IDRaman Mini, has been named as a finalist for the SPIE Prism Awards 2014, whose winners will be announced at Photonics West in February.

Handheld spectrometers are also being developed for the general public. A team of researchers at the Fraunhofer Institute for Photonic Microsystems (IPMS) in Germany has created a prototype near-infrared (NIR) spectrometer smaller than a sugar cube, which in the future could be integrated into a smart phone. This would enable consumers to check the quality of food before they buy it – whether it be the fructose content in fruit or the fat content in meat.

This September a Canadian company, TellSpec, plans to release a handheld device for the public, so that a consumer can simply point at an item of food to analyse its composition.

Fraunhofer IPMS’s miniature spectrometer emits light from 950 to 1,900nm, which is diffracted on a movable optical grating and collected by a single Hamamatsu uncooled InGaAs photodiode detector. The central element of the spectrometer is a micro-electro-mechanical systems (MEMS) chip that contains the grating and both optical slits and measures just 9.5 × 5.3 × 0.5mm. The individual gratings and optical gaps are manufactured directly on silicon wafers and diced into individual spectrometers. With this method, a single six-inch wafer is large enough to produce several hundred spectrometers and scientists do not have to adjust the components piece by piece as with conventional spectrometers. ‘That’s the idea for mass production for the spectrometers that would be used in mobile applications,’ said Dr Heinrich Grüger, business unit manager of Sensor and Actuator Systems at IPMS.

The next stage for IPMS is to find a partner who can integrate the technology into a marketable device such as smart phone. Pricing is the main challenge, according to Grüger: ‘We have spoken to all the technology scouts from companies that visit our post in the USA. They would like to integrate the spectrometer – as soon as we can provide it for less than €5. This will take a lot of work.’

Grüger believes that new technology is required to manufacture high volumes of optical components and drive the price down. ‘We can already produce the MEMS component at IPMS cheaply and in high volumes, and Hamamatsu can manufacture the InGaAs detector at even higher volumes,’ explained Grüger. ‘But the optical components are quite tricky – they are currently produced by ultra-precision milling out of a metal substrate. What needs to be invented is some sort of plastic moulding technology that meets the requirements for the optical performance.’

It will be around three years before the system will be ready for introduction to the market. Grüger believes that the first spectrometers marketed to consumers for food-checking will be in a different form: ‘I do not think the very next device will be a mobile phone, but some kind of handheld, integrated system with a similar platform.’

TellSpec has been developing such a product. Due to enter the market in September, the product will be small enough to fit on a key chain and will be able to identify calories, macronutrients such as carbohydrates, as well as allergens, chemicals and other nutritional information. From a coeliac who needs to avoid gluten, to a person wanting to count calories or carbohydrates, this device could be highly beneficial to a large variety of consumers.

The system will consist of a handheld spectrometer, a unique cloud-based algorithm created by Dr Stephen Watson, CTO at TellSpec, and a smart phone app. The data is uploaded to the Cloud – which processes the information, compares it to pre-set food fingerprints, interprets the results and downloads the information to a smart phone.

Having completed the first round of investment, the team is in the process of dealing with the miniaturisation of the hardware and configuring a beta design. A variety of Raman and non-Raman spectrometers supplied by different companies have been tested, but a final choice has yet to be made on the spectrometer and manufacturing company that will be used for the final device. ‘We’ve tested several spectrometers and our algorithms performed very well with all of them,’ said Isabel Hoffmann, Tellspec CEO. ‘In the next month we are making a decision regarding who our main optics manufacturing partner will be and we have narrowed it down to two companies.’

The light source is a key factor, according to Hoffman: ‘We would prefer not to use a laser because there would be less safety concerns – whether to use a laser or not is the choice we have to make.’ Tests are being carried out to check that the algorithms will work with the proposed broad bandwidth light source, and to test its feasibility within the final product. ‘We have tested the devices at the actual two locations of our potential partners. We had to go a little bit backwards because of the possibility of having no laser involved, but it is worth taking the time if we can achieve this.’

The production capacity of the two companies will also affect the decision, even though it is likely that some of the manufacturing will have to be outsourced.

‘Although the companies we are in discussions with are large and very solid in the industry, they are simply not geared up to supply 5,000 to 10,000 spectrometers in such a short amount of time,’ explained Hoffman. ‘It would be great if the company we decide to work with will be able to produce this number of units at once – but I don’t think it really exists. So, we will have to outsource some of the manufacturing portion, likely from an Asian company geared up for large assembly.’

However, handheld spectrometers are generally less suited to large-scale food production, because of highly variable samples and difficulties in focusing. To make a handheld device more robust for these types of applications, Ocean Optics has developed a spectrometer that can be used to inspect inhomogeneous food samples at different stages of production. The IDRaman mini has been named as a finalist for the SPIE Prism Awards 2014.

‘Normally, with samples such as rice where you have food particles that are spaced oddly or are an odd size, focusing becomes difficult with handheld instruments,’ said Dr Michael Allen, director of marketing and product development at Ocean Optics. The sampling technique in the device, raster orbital scanning (ROS), overcomes these difficulties by scanning a tightly focused beam in an orbital pattern, allowing data to be collected from a larger area of the sample. ‘By scanning over a larger area – something that’s millimetres in diameter as opposed to microns – you will get a better picture of what’s going on in the food sample.’ An additional complementary technique that is currently under development by Ocean Optics is surface enhanced Raman (SER), which uses a substrate to enhance weak Raman signals and produce stronger peaks in the presence of a particular compound.

The advances in sampling allow the handheld device to provide increasingly reliable analysis that is comparable to a bench-top Raman system. ‘These technologies are going to make it easier for someone to analyse a box of food with a Raman spectrometer and SERs and get reliable data, instead of having to sample multiple times, or go back to the lab, which makes the sampling process very inconvenient,’ explained Allen. ‘That’s really been the driving force of our Raman products in the food market – although it is still early days for us.’

So, how is it that handheld spectrometers are able to give the same performance as a laboratory-based machine, but at a fraction of the size and cost? Better manufacturing processes and miniaturisation of components have been major factors, according to Allen: ‘Technologies such as the cell phone have driven linear CCD array detectors down to the point where they use very little power and are very small. These detectors are being mass-produced, even in the small form factor, making it easy to acquire high-quality detectors at a reasonable price.

‘Once you have a very small detector, the other optical components can be easily miniaturised and scaled down,’ said Allen. ‘Really you’re just leveraging all of the miniaturisation and low-power efforts from all of the individual components of the spectrometer to be able to put that into a very small form factor device.’

Not only are spectrometers becoming ever more compact, but they are being designed to be faster and more robust to perform analysis on food production lines. Spanish flour supplier Emilio Esteban uses a NIR spectroscopic system to inspect the flour that flows through its pipes at a rate of 25 metres per second. The spectrometer in the system, supplied by Ibsen Photonics, is placed on one side of the flow opposite a lamp on the other side, and a measurement is taken that lasts for one tenth of a second. Compared to conventional methods where samples had to be taken to a spectrometer in a laboratory, the processing time is significantly reduced.

‘Most of the reason to do in-line process control is to shorten the feedback link – you can immediately stop the production line if there is a problem,’ said Dr Thomas Rasmussen, vice president of sales and marketing at Ibsen Photonics. ‘This allows you to create more consistent products.’

The development of grating-based NIR spectrometers with non-moving parts has been significant in making it possible to measure at these high-speeds. ‘In the larger bench-top devices, the technology involves a moving part or a scan, which would take seconds or minutes,’ said Jason Pierce, director of business development at StellarNet. ‘Our devices use a grating and a detector array, where you have a set exposure time to be able to capture the same amount of information in milliseconds.’

Moreover, the type of grating used can make a difference to its performance in a food production environment. ‘Transmission gratings made out of glass help with the speed – the more light you get through, the shorter time it takes to measure,’ explained Rasmussen. ‘These gratings are also thermally stable, which means that even with vibrations and temperature variations that exist in an industrial setting, you still get the high throughput with no changes in the angle of the diffracted light.’

As seen with the Raman device, it is high-volume mass production of miniature components that has led to compact and cost-effective NIR spectrometers. ‘We make the transmission gratings ourselves, by producing 200 to 300 gratings on a single wafer and then dicing them out,’ said Rasmussen. ‘Similar approaches can be used to produce the lenses and mirrors and, also, very small standard mirrors can be supplied in high volume by lots of Chinese suppliers.’

A similar technology suited to the high speeds of a production environment is integrated wavelength-selective detectors. The PixelSensor from Pixelteq has the option to place eight narrow bent spectral filters in front of an array of eight photodiodes. So, a food supplier is able to select and measure only the wavelengths related to their food item instead of measuring the whole spectrum, which reduces the cost, size and time of measurement. ‘In the future I think this kind of technology will be used very heavily for in-line inspection because a company will know exactly what wavelengths they are looking for,’ said Marco Snikkers, director of sales and marketing in Europe at Pixelteq. ‘The device is limited to eight critical wavelengths, so the analysis can be carried out at very high speeds because the processing time goes down.’

In the future, spectrometers used in food quality control will be able to analyse different types of food samples at once, according to Rasmussen. ‘I think that the same spectrometer will be capable of monitoring several process lines at the same time. So, there would be five optical fibres as input to the spectrometer, and the ends of these fibres would be placed at different locations for testing.’ Pierce agreed that the way forward is for a single spectrometer to be capable of testing different food samples: ‘StellarNet is currently making case-carrying systems that have everything included in it. The same device could be used to do quality control both at the production line and at the agricultural field. Companies will be able to monitor products several times during production by using the same spectrometer.

‘I definitely feel that miniature spectrometers are going to revolutionise production and efficiency,’ Pierce added. ‘The entire industry, from the field in agriculture, all the way to the inspection line of a final product, is going to be affected by these low-cost mini spectrometers being implemented in all these steps in production.’

About the author

Jessica Rowbury is a technical writer for Electro Optics, Imaging & Machine Vision Europe and Laser Systems Europe.

You can contact her on jess.rowbury@europascience.com or on +44 (0) 1223 275 476.

Find us on Twitter at @ElectroOptics, @IMVEurope, @LaserSystemsMag and @JessRowbury.