Semiconductor foundry CST Global is leading a UK government-funded project that could substantially reduce the cost of manufacturing high-speed laser diodes for next-generation fibre-to-the-premises (FTTP) networks.
The market for passive optical networks (PON) is expanding rapidly as consumers upgrade their broadband connections to higher speeds. Indeed, this summer, CST Global said it had shipped more than 25 million lasers into PON markets worldwide. To meet the insatiable demand for bandwidth, the next generation of PON lasers will need to be both higher speed and lower cost.
The £1.1 million project, ‘Diode Laser manufacturing process using Nano-imprint lithography’ (DiLaN), aims to establish a high-volume, commercially viable process for making distributed feedback (DFB) semiconductor lasers capable of 25Gb/s operation.
‘Our project addresses the replacement of a high-cost, nanometre-scale lithography step in the laser manufacturing process with a low-cost, high-throughput nano-imprint process to realise a cost saving of 20 to 30 per cent in the cost of manufacture of the laser chip,’ explained Andrew McKee, director of engineering at CST Global.
DFB lasers contain an integrated grating structure as the wavelength-selective element. The ultra-fine features in the gratings are usually created using electron-beam lithography, by directly ‘writing’ the pattern into a photosensitive material that can be used as a mask for an etch process.
Nano-imprint lithography is an alternative method that can be used to pattern photoresist at 100nm feature sizes. The process is simple, according to McKee. A patterned stamp is pressed against the wafer to create patterns in the photosensitive resist, which are then transferred into the wafer.
This simplicity means that nano-imprint lithography tools are less expensive and the process is rapid. E-beam machines cost millions of dollars and the ‘writing’ process can take hours; the large depreciation cost of the machinery results in high costs in the region of $2000 per wafer. In contrast, nano-imprint lithography costs could be as low as $200 per wafer, McKee claims.
‘Nano-imprint lithography is widely recognised as the most credible method of producing high throughput, high-resolution, singlemode semiconductor lasers at low cost,’ according to the project description. ‘However, to our knowledge, the nano-imprint lithography technique has not been implemented in volume semiconductor laser manufacturing, and so there is significant de-risking activity required to establish, qualify and yield engineer a new process to unlock the productivity gains.’
The UK industrial partners in this project are already significant materials and chip-scale suppliers to this market. CST Global is the project leader, supported by academic partners Cardiff and Swansea Universities and commercial partner Compound Semiconductor Centre (CSC) in Cardiff.
Now open for business, the CSC is a joint venture between IQE, a leading supplier of compound semiconductor wafer products and Cardiff University (see Cardiff cluster to commercialise compound semiconductor research). The CSC is intended to be Europe's first prototyping facility, creating a more rapid route to market for compound semiconductor based technologies.
The DiLaN project grant is valued at £821,319. Funding is provided by Innovate UK, a UK government research funding agency for engineering and the physical sciences. The DiLaN project runs from February 2017 to January 2019.