Strain engineered InAs/GaAs quantum dots for long wavelength emission

Lead Research Organisation: Imperial College London
Department Name: Physics

Abstract

Laser sources for third generation optical communications systems have a market value of $6bn. The target wavelength of 1550 nm corresponds to the minimum of attenuation in optical fibres and laser sources are based on InP wafers; yet GaAs is cheaper since the standard wafer diameter is 6 rather than 3 for InP. In addition the crystallographic quality of GaAs is better than InP and the GaAs/AlGaAs system is near ideal for the monolithic inclusion of distributed Bragg reflectors for optical microcavity structures. There has therefore been a world-wide effort to develop lasers at 1550 nm grown on GaAs. Dilute nitride GaInNAs quantum wells (QWs) grown on GaAs substrates is a possibility but nitrogen has a low solubility in GaAs, and the material has to be subjected to annealing procedures following growth to improve the optical quality and there have been several reports indicating poor lifetimes of GaInNAs devices. Extending the wavelength to 1550 nm is difficult and requires the incorporation of much larger amounts of N (~3 - 5%) or the inclusion of Sb neither of which appears commercially viable at present. InAs QDs grown on GaAs substrates provide an alternative route to telecomms wavelength sources. Good quality 1300 nm QD devices are commercially available and fulfil the need for sources where signal dispersion in fibres is a problem. QD devices offer some advantages over QWs: high modulation speeds, low linewidth enhancement factors (chirp) and high characteristic temperatures; all of which have been demonstrated. If some or all of these properties can be extended to 1550 nm then many future telecoms devices (both lasers and optical amplifiers) will be based on QDs grown on GaAs substrates. A key aim of this proposal will be the development of laser sources across the telecomms C-band to extract the maximum bandwidth. This will require many sources emitting at closely spaced wavelengths. These can be accessed post growth by processing gratings (Distributed Feedback Lasers) which will select the lasing wavelengths across the QD gain spectrum.Despite the technological process there are still many fundamental issues related to QD lasers which have yet to be properly investigated and understood. An ensemble of dots will have an inhomogeneous gain spectrum as evidenced by the presence of groups of lasing lines when the devices are operated at low temperatures. In addition the gain does not clamp at threshold and this can lead to dual state lasing which has been observed by us and other groups. It has been proposed that this behaviour is due to a phonon bottleneck (restriction on relaxation) but we have observed filamentation (different optical modes) which could account for the observed behaviour. By applying a novel second derivative I-V technique to monitor gain switching we can correlate this signal with the far field optical signal to investigate the inhomogeneous gain. We anticipate that this will become a standard characterisation techn ique for QD lasers.
 
Description The emission wavelength of InAs/GaAs quantum dots (QD) is typically limited to the range 900-1300nm. Since the emission wavelength is dictated by the size and composition of the dots we investigated methods of extending the wavelength towards those compatible with optical fibres. The method involved growing a lower QD template followed by a second layer of QD where the temperature could be lowered and the InAs deposition increased to form larger QD emitting out to 1520 nm at room temperature.
Exploitation Route These results would be of interest to telecommunications companies to fabricate low threshold optical fibre compatible laser sources.
Sectors Digital/Communication/Information Technologies (including Software)