Thierry Robin and Clémentine Bouyé, Tematys
Terahertz technologies – which use the band between 30 µm and 3 mm – are bulky and expensive systems, used mainly in R&D and lab applications, especially astrophysics. But great technical progress in recent years has enabled commercialization of terahertz technologies for industrial uses such as nondestructive testing (NDT), industrial process monitoring and pharmaceutical quality control (QC); other possibilities include security markets or, in a few years, telecommunications and biomedical markets. Terahertz is in fact well suited for industrial applications, and this field is the most promising short-term market for the technologies (after the existing research market).
Terahertz technologies are already starting to be adapted for industrial requirements. They allow implementation of the first online application of terahertz spectroscopy for paper-thickness measurements and analysis as well as other applications in at-line NDT of composites and powders. The reinforcement of regulations will open applications such as QC of food or pharmaceuticals.
Figure 1. Terahertz components and systems have evolved over the years. NDT = nondestructive testing.
The adoption of manufacturing execution systems (MES) by manufacturers will contribute to the use of terahertz products for online NDT of semiconductors or composites. But technical improvements are needed before terahertz technologies can be widely adopted: We must increase the acquisition speed, improve the reliability of terahertz measurements, and create an extensive database for a better interpretation of spectroscopic data. Finally, in eight to 10 years, when the technology reaches a certain maturity, investments expected from telecom companies should reduce the costs of terahertz components and allow a wider adoption in the industrial field.
Ruggedness, ease of use: Short-term opportunity
At their beginning in the 1980s, terahertz systems were developed by researchers for laboratory applications. They were complex, expensive scientific systems that could be used only by highly qualified staff. However, over the past 10 years, great technical developments have been made toward the implementation of easy-to-use, cost-effective systems through the development of more compact and reliable components. For example, the maturity of femtosecond lasers allows the development of less bulky, less expensive sources.
As for terahertz detectors, the tendency is to implement detector arrays such as bolometer arrays, allowing the combination of smaller size with great detection performance.
As components have become cheaper, more reliable and easier to integrate, systems have become smaller, less costly and less complicated. Today, several manufacturers of terahertz systems for spectroscopy offer their products at a price less than €100,000 (Emcore, Menlo Systems, Hübner and others). Thanks to all these improvements, terahertz techniques are now entering the commercial market, as the first products for applications in industrial process monitoring have been sold.
But terahertz technologies have not yet reached maturity, and some challenges remain for their widespread use in industrial NDT and process monitoring. In terms of power performance, there is still important work to do: Most terahertz sources deliver less than 1 mW at 1 THz, which is often too low for industrial NDT or process monitoring applications. Moreover, for wide acceptance and use in online process monitoring, terahertz systems and components must be more reliable and work without any staff intervention. Also, a system must cost less than €20,000 to €50,000 (about $27,000 to $68,000), depending on applications, to be widely implemented in online processes.
Terahertz spectroscopy, in particular time-domain spectroscopy (TDS), is the most mature terahertz technique and the one most likely to be adopted in industrial applications. However, the systems’ speed must be increased because high throughput in spectra acquisition is needed. Current progress in time-delay design (mechanical, electronic, optical) paves the way for accurate, rugged solutions. Moreover, the use of terahertz spectroscopy is hampered by the lack of reference spectra in current databases. For example, the terahertz Riken database contains 1550 spectra, compared with the 1.4 million spectra of the IR Bio-Rad database.
QC, process monitoring: Mid-term opportunity
The intensification of regulations in the industrial field is an opportunity for terahertz technologies in two ways: First, more and more regulations force industries to monitor their process and control the quality of their product (before, during and after the process). Because of its ability to penetrate through barrier materials (clothes, packaging, etc.), to perform noncontact and NDT, terahertz technology is an adequate candidate for inspection, control and monitoring in the industrial field. Second, the technologies used for control and monitoring will be regulated. Nonionizing, environment-friendly technologies will be supported over incumbent technologies such as x-rays or nuclear technologies. Terahertz technologies fit these safety constraints.
Picometrix (an API company) demonstrates the importance of the regulations on safer technologies for the establishment of terahertz technologies in industries. In the paper manufacturing industry, online monitoring of the paper’s thickness is required. Until now, this was done by nuclear gauge, but recently the regulations strengthened, and using nuclear gauge became more complex because of recycling constraints.
Terahertz TDS systems are very sensitive to thickness, so Picometrix developed an online terahertz product called the T-Gauge to replace the nuclear gauge in the process. The first T-Gauges were sold in 2012. This kind of substitution is expected to happen more and more in the coming years.
However, Figure 2 shows that there are numerous nonionizing competing technologies. For example, terahertz techniques must compete with techniques such as ultrasound – methods that present drawbacks (low propagation in air, highly skilled operator required) but which are mature technologies more likely to be adopted by manufacturers. Terahertz technologies have to prove their reliability compared with these techniques before being accepted.
Figure 2. Terahertz competes against other nonionizing techniques for process monitoring in the industrial field.
The main challenge for a wide adoption of terahertz technologies will be to find the application(s) where they answer unfulfilled needs and/or where they can replace ionizing techniques such as nuclear gauge or x-ray systems.
Manufacturing execution systems: Long-term opportunity
An MES is a computer system that gathers production data in real time with the goal of optimizing production throughout the process, from the order to the end product. The collected information can be used across different function areas: management of product definition, product QC, product track and trace, production performance analysis and so on. The use of an MES leads to more efficient manufacturing processes, waste reduction, maintenance reduction and increased uptime.
In recent years, manufacturers have started to use MES more intensively to improve their productivity. For this purpose, accurate and reliable measurements – and, consequently, accurate and reliable technologies – are required. This is a great opportunity for the implementation of terahertz spectroscopy and imaging systems over the long term.
The most important technical feature for industry is the ability to perform online, real-time and noninvasive measurements. Terahertz measurements are contactless and have begun to be operated online; the main hurdle today is the acquisition time, which is too slow for a wide adoption of terahertz systems in industrial process monitoring.
However, many developments are being implemented to adapt to the needs of the market. The ultimate aim is to attain real-time measurement. In particular, TDS is getting faster and faster, as seen in Figure 3. TDS is one of the most promising terahertz techniques for implementation in the industrial field; for example, in composition analysis of powders or in NDT of composites. Figure 3 shows that the first commercially available systems were very slow (less than 100 waveforms per second), taking several minutes to several hours to acquire an image – a rate impossible for manufacturing. At that time, technical improvements allowed an acquisition time of a few minutes.
Figure 3. A look at the past and future evolution of the acquisition speed of time-domain spectroscopy systems, with examples of manufacturers.
Today, with disruptive technologies in development, it is possible to reach acquisition speeds of 1000 to more than 100,000 waveforms per second – so real-time applications are conceivable. These kinds of systems are expected to be commercially available around 2020 and could lead to a wide adoption of terahertz systems in industry.
Market for industrial applications
The terahertz market is expected to grow from €34 million in 2012 to €350 million in 2022 ($47 million to $482 million), with a strong compound annual growth rate of 26 percent. Industrial applications hold the biggest market share, representing around 50 percent of the total terahertz market with €15 million in 2012 and €175 million ($21 million to $241 million) expected in 2022.
The first terahertz products are now commercialized for several applications in the industrial field (Figure 4), such as analysis of paper and powders or analysis of composites near the process line. More and more products are expected to be available on the market in five to eight years in key sectors such as pharmaceutics or semiconductor manufacturing.
Figure 4. A road map of terahertz industrial applications.
The industrial field is complicated to enter with new technologies, and the terahertz market will not rocket forward here because technical developments are very long term. Indeed, for wide adoption, it is necessary to have high technical performance at low cost. It took about 15 years to see the first terahertz commercial products for industrial applications, and it will take 10 more years for widespread use of the technology by manufacturers.
As a comparison, the telecommunications market for terahertz technologies is expected to rocket between 2018 and 2023. What takes time in the telecommunications field is creating the new standards needed to use the terahertz band. Once standardization is created, the entrance to mass-market applications will be possible, especially as important investments are made to allow the development of cheap devices. The good news is that investments in telecommunications will have a positive impact on the adoption of terahertz technologies by the industrial field, as they will speed up cost decreases.
Meet the authors
Thierry Robin is chief technical officer at Tematys in Paris; email: [email protected]. Clémentine Bouyé is an analyst at Tematys; email: [email protected].
Three reasons terahertz could become a technology of tomorrow for industrial applications
1. Technical evolution is making terahertz systems more rugged and easier to use; this is a short-term opportunity.
2. Strengthening of regulations brings the need for safe and reliable tools for QC and process monitoring; this is a mid-term opportunity.
3. Terahertz fits the trend toward manufacturing execution systems in the industrial field; this is a long-term opportunity.