05 Mar 2004
Scientists around the world are busy trying to develop a practical semiconductor laser that emits terahertz waves. Rob van den Berg reports on the progress being made in the race to raise the temperature of operation and output power of terahertz sources.
From Opto & Laser Europe March 2004
That said, a narrow region of the electromagnetic spectrum between the infrared and microwaves has so far remained largely untouched by human innovation. This void is known as the "terahertz gap"and is approximately 1-10 THz in frequency (equivalent to a wavelength of between 300 and 30 µm).
Diverse applications Demand for a convenient, cost-effective source in this region has increased in recent years, as scientists have discovered that terahertz waves have a wealth of potential applications in security screening and medical imaging. Materials such as fabric, plastic and wood appear transparent in the terahertz region, which could be useful for revealing the presence of concealed metal items such as guns or knives. In addition, the distinctive terahertz absorption signature of cancerous cells and pharmaceuticals could lead to important clinical applications.
As a result, researchers worldwide are on a quest to develop semiconductor-based terahertz lasers that will mimic the performance of their infrared and visible counterparts. The problem is that conventional laser diodes have a large bandgap and thus emit light at far too high a frequency. Even diodes that are based on narrow-gap materials cannot easily be extended to wavelengths of several tens or hundreds of microns.
The solution may lie with bandgap engineering. The first big breakthrough was made almost two years ago by an Anglo-Italian collaboration from the National Enterprise for Nanoscience and Nanotechnology (NEST-INFM) at the Scuola Normale Superiore in Pisa and the Cavendish Laboratory at Cambridge University. Rüdeger Köhler and his colleagues succeeded in making a semiconductor terahertz laser based on quantum cascade (QC) engineering - a bandgap engineering concept developed at Bell Labs in the 1990s for making mid-infrared lasers.
The first working device from the team produced intense radiation at 4.4 THz (corresponding to a wavelength of 67 µm) and delivered a maximum output power of 2.5 mW. The downside was that it had to be cooled to a temperature of 50 K.
Since this prototype was unveiled, there has been enormous progress in the field in terms of wavelength, output power and operating temperature, which has now increased way above liquid nitrogen temperatures.
Köhler explained: "By alternating ultra-thin layers of semiconducting materials like gallium arsenide (GaAs) and aluminium gallium arsenide, multiple quantum-wells can be generated, which restrict electron motion. The electrons can only jump from one state - called a sub-band - to the other by discrete steps, emitting photons of light. The spacing between the steps depends on the width of the well and increases as the well size is decreased."
An active region at the heart of the device creates a population inversion (which is necessary for lasing) between two of these sub-bands. This region consists of a series of quantum wells in which electrons are selectively injected into the upper lasing state and are extracted as fast as possible from the lower lasing state.
At the Cavendish Laboratory, Köhler first calculated the optimal band structure before growing the samples by molecular beam epitaxy (MBE) on a substrate of semi-insulating GaAs. The net growth time is 12-14 h for more than 1500 layers, each of which is only a few nanometres thick and is precisely doped. The laser cavity is obtained by cleaving the material to form facets that act as mirrors.
Waveguide challenge One of the big problems was how to guide the terahertz radiation within the device. "In the mid-IR region, one can still use a conventional dielectric waveguide, consisting of only a few layers on top and bottom. In the terahertz region, however, this is no longer possible as the thickness of the waveguide has to be comparable to the wavelength - some tens of microns," said Köhler. "This is impossible to grow by molecular beam epitaxy. Moreover, there would be enormous optical losses due to absorption by free carriers."
The team solved this problem by guiding the radiation mode via a thin, highly doped GaAs layer inserted directly underneath the active core of the laser. "We tailored the doping concentration and thickness of the layer to achieve a high overlap of the mode with the active core, at the same time as minimizing absorption losses," said Köhler. "This is quite exceptional as we are able to confine the optical mode to within 10 µm, below the diffraction limit."
Since May 2002, Köhler and team have optimized processing procedures, reduced waveguide losses and improved laser performance by depositing a high-reflection coating onto the back facet of the laser. This allowed an increase in the maximum operating temperature to 75 K, with peak output powers of about 5 mW. Continuous-wave (CW) operation was obtained with similar output power up to a maximum temperature of about 50 K.
In the meantime, other groups have joined the game. Across the ocean at MIT in Cambridge, Massachusetts, US, Qing Hu had been working in the field for several years, when his student Ben Williams also succeeded in making a semiconductor-based terahertz laser.
Since December 2003 they have held the record for the highest reported operating temperature - 137 K at 3.8 THz. Hu explained: "We decided to go for a metal waveguide, in which one can confine radiation to almost 100%, just like with microwaves."
Meanwhile Köhler and his colleagues have very recently fabricated distributed-feedback lasers with high spectral purity, and are also working on the implementation of an external cavity for wide tunability. Köhler said: "At these high temperatures the levels have broadened due to lattice vibrations, and they will start overlapping so that you lose control. Another problem is that it will be much harder to cool the electrons in between the active regions. But I am confident we can reach temperatures of 200-210 K, at which you can operate with a thermoelectric cooler in a portable device."