This is the first time live control over the output spectrum of QCLs has been demonstrated, and it could lead to real-world applications – see more below.
Quantum cascade lasers are not new.
Physically, a QCL is a stack of quantum wells – in this case GaAlAs-GaAs quantum wells grown by molecular beam epitaxy.
These are etched to leave a long thin block (a ‘laser ridge’ 6mm x 180µm, for example) with flat ends and quantum well layers running form end to end – a ‘Fabry-Perot’ laser cavity (see diagram).
Power fed through the thickness of the quantum wells from top to bottom leads to THz radiation leaving the ends of the ridge.
Unlike simple laser cavities, the output is not at a single wavelength but has a spectrum of wavelengths, and this spectrum can be controlled by depositing a patterned gold film along the top of the ridge to encourage some modes (output frequencies) and discourage others.
These patterns take the form of slots (~1um wide) cut through the metal film across the width of the laser cavity. The slots create surface plasmons which modify local electric fields and affect the modes within the cavity. Effectively, they are small antennas on top of the laser cavity, Dr Subhasish Chakraborty, lead researcher on the graphene-controlled laser, told Electronics Weekly.
What is new:
The pattern of slots previously used, called a ‘grating’, is generally periodic – the slots are at regular intervals along the laser cavity – but for the last few years Manchester has been working an a-periodic holographic form of grating.
Starting with the desired output spectrum, a computer algorithm slot spacings that will deliver that output effectively, and the results are generally non-periodic.
“It is a powerful technique. You set the target, and it calculates where to put the slots,” said Chakraborty.
In the proof-of-principal device, reported in the journal Science, the gold holographic grating encourages an output spectrum extending from 2.88 to 3.0THz with four peaks at 30GHz intervals. These emerge along with a sea of largely-suppressed sidebands.
The Manchester team built two similar lasers, one with just the gold grating and an identical one modified with graphene.
In this case, it was a mono-layer graphene (one atom thick) sheet made separately on copper foil by chemical vapour deposition and transferred to the laser. The graphene sheet covers all of the slots, and is separated from the gold by a thin polymer insulator to block short circuits.
The important graphene characteristics in this case are, that while gold has fixed conductivity and fixed surface plasmon properties, graphene’s Fermi energy and carrier concentration can be changed by electrical biasing to adjust its conductivity and surface plasmon properties.
Effectively, the unmodified laser has gold on top interrupted by narrow ‘no-gold’ slots, and the modified cavity has gold interrupted by narrow transverse graphene strips.
Compared with the unmodified reference laser, biasing the graphene with respect to the gold on the modified laser has a profound effect on the level of unwanted sidebands.
With the graphene set to the same voltage as the gold, the sidebands are far worse – In effect the slots have been partially turned off because the graphene-covered slots have low plasmon density and are poorer antennas than ‘no-gold’ slots.
Set with graphene 1V positive with respect to the gold, the sidebands are suppressed far more than with plain slotted gold – high plasmon density in the graphene covered slots makes them into super-slots – better antennas than no-gold slots.
“We are the first to excite graphene plasmons directly on top of a laser,” said Chakraborty.
Universal THz lasers for industry
Demonstrating control over the purity of four peaks in a THz spectrum might not seem too exciting, but, according to Chakraborty, the full potential of the technique will be realised when technologists find ways to deposit individual graphene control gates over individual slots.
“Quantum cascade lasers have been neglected by the market because there is no way to make them relevant to industry,” he said.
“Current terahertz devices do not allow for tuneable properties, a new device would have to be made each time requirements changed, making them unattractive on an industrial scale,” said Kostya Novoselov, also on the team.
With algorithms that can turn spectral requirements into slot patterns, plus the ability to turn slots on and off on-the-fly, single THz lasers can be designed for more than one application, or to generate more than one spectrum in a single application.
Manchester University is arguing that markets too small to be addressed by THz lasers because they each need their own spectrum, or need multiple spectra in the same box, could now be addressed by one type, or a few types, of mass-produced adjustable QCL laser.
Lasers for what?
THz radiation is useful for a couple of major reasons: some chemicals canbe identified by their THz spectra, and THz radiation penetrates some way into certain materials, including skin, allowing high-resolution 3D scans to be performed. Such scanning could be detailed due to wavelength around 0.1mm, and less damaging than x-ray scanning.
Applications that could be unlocked, according to the University, include scanning and analysing materials in pharmaceutical, security and agricultural markets.
Output power should not be an issue, said Chakraborty, because the semiconductor industry has a good track record in scaling up chip lasers.