Source: Nature Photonics 1, 257 - 258 (2007)

    Is the terahertz spectral range finally about to be opened up for broad application across the physical and biological sciences? Researchers propose a new source of terahertz waves that could do just this.

    Over the past decade, there has been a staggering increase in the number of research groups worldwide seeking to develop and exploit the terahertz frequency range (which spans from approximately 100 GHz to 10 THz). This has been fuelled, at least in part, by the promise of applications across the physical and biological sciences. These applications include (but are by no means limited to) the spectroscopic studies of pharmaceutical products (for example, distinguishing different forms of a given drug1 or reconstructing chemical distributions inside tablets), the detection of explosives2 and illegal drugs (useful for security screening at airports), the determination of disease (such as cancers) in skin tissue3 and industrial-process monitoring. These applications sit alongside the traditional needs for terahertz sources, detectors and systems for current and future astronomical programmes, as well as for studies of condensed-matter physics and nanoscale structures, for which key energy scales and timescales lie in the terahertz spectral range.

    Given all of these promising avenues, why is it that researchers are still only discussing ＆potential＊ applications and the market ＆potential＊ of the terahertz frequency range? Why have these anticipated applications not yet materialized? After all, the neighbouring microwave and mid-infrared regions of the electromagnetic spectrum are well populated with everyday devices, supporting, among other things, the mobile-telephone industry. The answer to this is principally related to the source technology: there does not exist a compact, room-temperature, high-power, semiconductor source that is well controlled, ideally tunable and suitable for the terahertz frequency range. Or at least, not yet. The technique proposed by Belkin et al.4 aims to address the critical need for a new terahertz source.

    At microwave frequencies, compact electronic devices are typically used as high-power sources. But at frequencies substantially above 100 GHz, as the transit times of electrons in devices become shorter and capacitative effects are more dominant, it becomes increasingly difficult to generate large output powers. Conversely, approaching the terahertz frequency range from the visible side of the electromagnetic spectrum, it becomes progressively more difficult to engineer and maintain a population inversion between two states in a laser, as their energy separation becomes closer and closer to kBT (where kB is Boltzmann＊s constant and T is temperature). For these reasons, most uses of terahertz waves have been demonstrated with broadband terahertz spectroscopy systems that are based on femtosecond lasers; indeed it is this technology that is being marketed to the pharmaceutical industry today. But, given the need for a femtosecond laser, it is unclear whether such systems can be made sufficiently compact and cheap for everyday market opportunities.

    A breakthrough in addressing the terahertz frequency range occurred in 2002, with the first demonstration of a terahertz quantum cascade laser5 (QCL), which followed the pioneering work that led to operation of the first mid-infrared QCL in 1994 (ref. 6). Mid-infrared lasers now operate in continuous-wave mode above room temperature and are being marketed for a range of applications, including monitoring of atmospheric pollutants. Unfortunately, however, unlike its mid-infrared counterpart, it is still unclear whether a terahertz QCL will ever operate at temperatures substantially beyond the values of 164 K seen today7. It is in this context that the work of Belkin et al. makes its mark. Rather than trying to optimize the performance of terahertz QCLs further, they instead refine two well-established mid-infrared designs and combine them in a single semiconductor chip. Terahertz radiation is then generated by mixing the two separate mid-infrared frequencies 〞 but the really neat trick in this case is that the mixing is engineered to occur within the same semiconductor chip that contains the two mid-infrared QCLs, forming a monolithically integrated structure. Previous efforts to produce terahertz waves through mixing techniques have not achieved this, even when two near-infrared diode lasers operating at wavelengths around 800 nm have been used.

    So how well does this new compact, integrated terahertz source from Belkin et al. actually work? At present, it generates radiation at 60 mum (that is, at a frequency of about 5 THz), and the source can be tuned (by changing the output wavelength of one of the mid-infrared lasers). This new source also has the potential to operate at room temperature, although this is still to be demonstrated practically.

    Is it the answer to the long-sought terahertz source? The answer would seem to be no, at least not yet. At present the source only offers an estimated 100 nW of power when operated at 80 K 〞 far too small for most of the applications envisaged for the terahertz spectral range. (This compares with the more than 10 mW of power commonly achieved in terahertz QCLs operating at cryogenic temperatures.) Belkin et al. suggest a number of ways in which the output power from their device can be increased by orders of magnitude, and hopefully this will trigger intense international activity in this field over the next few years. However, only time will tell whether or not this will prove to be the much coveted solution for the terahertz frequency range.

    By Edmund Linfield

References
1. Upadhya, P. C. et al. Spectrosc. Lett. 39, 215每224 (2006).
2. Shen, Y. C. et al. Appl. Phys. Lett. 86, 241116 (2005).
3. Woodward, R. M. et al. J. Invest. Dermatol. 120, 72每78 (2003).
4. Belkin, M. A. et al. Nature Photon. 1, 288每292 (2007).
5. Köhler, R. et al. Nature 417, 156每159 (2002).
6. Faist, J. et al. Science 264, 553每556 (1994).
7. Williams, B. S. et al. Opt. Express 13, 3331每3339 (2005).