Wielding the increasing average power of ultrafast lasers
Ultrafast lasers are rapidly revolutionising manufacturing. The ultrashort pulses they generate – with durations ranging from femtoseconds to picoseconds – enable the extremely precise micro- and nanofabrication of a wide range of materials, due to their extremely low thermal penetration and the non-linear absorption effects they induce. As described by the Fraunhofer ILT’s Dr Arnold Gillner in an article for Laser Systems Europe earlier this year, this has led to these sources being adopted in a wide range of applications – including structuring, drilling, engraving, welding and cutting – across numerous sectors – including displays, automotive, electronics, photovoltaics and medical.
Between my visits to AKL last year and the Laser World of Photonics this year, a clear development trend underway is that the average power of ultrafast lasers is on the rise. At AKL, for example, the Fraunhofer ILT announced its intentions to develop ultrafast lasers with average powers ranging from 5kW to 20kW by 2022. Meanwhile, at the Laser World of Photonics, ultrafast laser manufacturer Amplitude revealed and demonstrated a new 300W femtosecond laser that has been developed within the framework of the TresClean project. Tresclean is one of several EU-funded projects ongoing in which ultrafast lasers with average powers up to the kilowatt range are being developed for a wide range of innovative micro- and nano-structuring applications.
According to Dr John Flemmer, however, business development manager at Scanlab, a manufacturer of scanning systems for ultrafast lasers, these increasing average powers will actually be much higher than is needed to perform the majority of applications that ultrafast lasers are currently used for.
‘We heard from an integrator recently that for most micromachining ultrafast applications, 10 to 20W is enough,’ he confirmed. ‘So if you are a company looking to buy an ultrafast laser machine, you’d be good to go with this amount of power.’
If this is indeed the case, then why are multiple businesses and projects seeking to increase the average power of ultrafast sources?
The answer, as explained by Dr Jose Antonio Ramos, director of research and development at Lasea, an integrator of ultrafast lasers, is so that the throughput of ultrafast laser processing can be dramatically increased, and so that new applications – currently deemed impractical due to the lack of throughput associated with the technology – can be enabled.
‘One application that this higher throughput could be of interest for is the structuring of moulds and tools for creating textured plastic surfaces – for example car dashboards,’ he noted. ‘Another application currently in high demand of ultrafast laser technology is the creation of hydrophobic surfaces, however this is still currently unfeasible due to the lack of throughput available for most applications.’
Creating functionalised surfaces such as those exhibiting hydrophobicity is the current focus of the EU project ‘LAMpAS’ that Lasea is a partner of, which began this year. In addition to developing a 1.5kW picosecond laser with a repetition rate up to 10MHz, the project seeks to combine direct laser interference patterning – creating periodic surface structures using beam interference patterns – with a high-speed polygon scanner to enable the mass-production of functional micro- and nanostructures on a range of surfaces, at a rate of 1 to 5m² per minute. These structures will provide antibacterial and self-cleaning properties, in addition to friction reduction, optical security functions and decorative effects.
The microstructures of this metal surface were created using an ultrafast laser and provide self-cleaning properties. (Image: LAMpAS project)
Combining beam inferencing with high-speed scanning technology is one of two strategies that Ramos explained can be used to wield the increasing average power of ultrafast sources to increase processing throughput. The other involves splitting the beam into multiple, lower-power beams using diffractive optical elements (DOEs) or beam splitting optics, which enables more surface area to be covered on a workpiece, or for multiple workpieces to be processed in parallel simultaneously. This approach has previously been used by Lasea when integrating 100W average power ultrafast lasers from Amplitude into its commercial laser machines.
‘Here, in order to make use of the higher power, we can split the beam in two, however in theory we can split it into four to improve the throughput of the machine even further. This would allow up to four parts (or a single part over a larger area) to be processed at the same time,’ explained Ramos. ‘At the SPIE Laser Beam Shaping Conference last year and the UKP Workshop hosted by the Fraunhofer ILT earlier this year, we presented the idea that, for typical ultrafast laser applications that can be performed well by the 10 to 20W average power lasers now widely available, these can be done more cheaply and faster using a 100W ultrafast source and four processing heads.’
Why stop at four beams though? Flemmer shared that some of Scanlab’s customers are looking to split the beams of higher average power sources even further, and deliver them to between 10 and 20 scanners at one time to achieve parallel processing. ‘Most of the applications we are hearing about for this increased throughput and use of multiple scanners is in the electronics market, for example for cutting and ablation in the creation of PCBs or flat panel displays,’ he said. ‘This market is growing particularly well in Asia.’ He also explained that while multiple beams can be deflected through a single scanner to achieve higher throughputs, this option can result in a limited field of view and is not as easy to manage as using multiple scanners.
Bettering beam delivery
It was also discussed at the UKP Workshop this year that in order take advantage of the increasing average power of ultrafast lasers, the most important developments that need to take place have to be carried out in the field of system technology. Ramos and Flemmer expanded on this by explaining that in the ultrafast regime, the laser sources have developed faster than the beam delivery technology used to wield them. Flemmer added that this is a good point, however, as the need for further system technology development is an excellent driver for innovation.
The development of new beam delivery solutions is therefore an important part of the many EU projects currently underway in which higher average power ultrafast sources are being developed.
In addition to LAMpAS, Lasea is also a partner of the projects ‘PoLaRoll’ and ‘HIPERDIAS’, both of which have required the development of new beam delivery solutions. Within PoLaRoll, for example, in order to cope with the higher pulse energy and repetition rate of the new femtosecond laser that has been produced, a new high-speed polygon scanner has also been developed. The resulting, high-throughput laser micromachining unit will replace the current masking method used in a continuous lithography etching process for micro-structuring stainless steel reels.
Lasea has to overcome new beam delivery challenges in order to equip its machines with the latest ultrafast sources of increasing average power.
Also, due to the higher intensity of energy that needs to be handled when using higher average power ultrafast sources, optics with higher damage thresholds – beam splitters, beam expanders, DOEs – need to be used to deliver their beam. ‘We are continuously finding better solutions for wielding higher average powers through our involvement in R&D projects,’ said Ramos. ‘In HIPERDIAS, we have successfully collaborated in the integration of an experimental femtosecond laser with 1kW average power within one of our machines. In doing this we did come up against limitations where the optical components we were using could be damaged, however we were able to find solutions to these issues. This demonstrates that collaborative projects such as this will be key for developing the optical solutions required to wield the higher average power ultrafast sources emerging.’
HIPERDIAS and PoLaRoll will be coming to a close this year, with the prototype systems under development expected to be turned into commercial solutions for industry in the next couple of years, according to Ramos.
In addition to the advancements made to polygon scanners to improve the capabilities of ultrafast laser processing – as is being done within the LAMpAS and PoLaRoll projects – galvanometer-based scanning technology is also being developed further to meet modern application demands.
‘As a manufacturer of scan heads for ultrafast lasers, we are looking to move towards more complex technology, as we can see that processes involving these sources are getting more and more complex,’ confirmed Flemmer. ‘For example, we have a very fast galvo-based 3D shifter called Excellishift that can be used for 3D-scanning – most applications are currently based on 2D-scanning. This scanner is as fast as our 2D-scanning systems and enables the fast structuring of 3D surfaces. This has applications in the tool making industry, for example for texturing plastic moulds to improve the aesthetics of the final product.
Scanlab’s five-axis Precsys system can deflect ultrashort pulses in the x,y plane, in addition to being able to tilt them for complex drilling applications.
‘We also offer another complex system which is heading in the right direction for ultrafast applications. Our Precsys five-axis drilling system is not only able to deflect the beam in the x,y plane, but can also tilt it, enabling it to be used to drill holes which are smaller at the top, for example for parts of a car where special holes are required. This has proven to be an attractive system for many customers, and we are now seeing competitors with similar systems.’
Looking forward, Flemmer explained that galvo-based scan technology will continue to be developed further, and will be optimised close to the physical mechanical limits of the technology to reach maximum speed, precision and stability.
‘In parallel, the complexity of systems will increase further and the integration of external peripherals is getting more and more relevant,’ he added. ‘These must not only be devices to deflect, shape or split laser beams… the integration of sensors for process analysis or process control is also an important topic. All of this is favoured by trends such as Big Data and the Internet of Things.’
Integrating an ultrafast laser into a system involving a moveable scan head usually requires a complex beam path of mirrors to be positioned on an articulated arm. The mirrors are placed on linear stages, enabling them to be moved towards and away from the laser source as the head is moved. While many companies use ultrafast lasers this way, the complex setup creates difficulty for machine integrators.
‘One of the hurdles to making ultrafast lasers capable of fulfilling industrial requirements and achieving strong penetration across multiple vertical markets, is the question of whether beam delivery can be dealt with in a secured and flexible manner,’ said Fetah Benabid, founder and chief science and technology adviser of GLOphotonics. ‘The answer to this question 15 to 20 years ago, was no, and that need has become increasingly pressing over the past five years due to the dramatic upscaling of average power of ultrafast laser sources.’
You might think that an optical fibre would be a far easier and likely safer way of achieving moveable scan head technology, as is already commonplace in other areas of laser processing, however the exotic nature of ultrashort pulses prevents system integrators from being able to use standard optical fibres to deliver them to a workpiece.
‘While it is possible to use solid fibres with nanosecond lasers, putting a standard glass fibre at the output of an ultrafast source is impractical due to glass’ low damage threshold and the dispersion and non-linearity effects that occur when the ultrashort pulses interact with the glass,’ explained Benabid. ‘The shorter the pulse, the higher the peak power gets and the more sensitive the pulse is to dispersion and the material is to pulse energy.’
However, GLOphotonics is now working with numerous integrators of ultrafast lasers because it claims to have developed a unique solution to these issues: the first fibre optic capable of delivering ultrashort laser pulses of high energy and high average power, while keeping their duration as short as possible over distances spanning tens of metres.
GLOphotonics’ hollow-core photonic crystal fibres enable ultrashort pulses to be transported without the need for complex mirror beam delivery paths
‘Our fibres are hollow core photonic crystal fibres (HCPCFs), meaning the light travels through either a vacuum, through air or through another gas within their core,’ said Benabid. ‘These features lend nicely to a number of functionalities in ultrafast lasers such as beam delivery, pulse compression and frequency conversion.’ He expanded on this by explaining that the vacuum and materials incorporated in these hollow core PCFs enable the controlling of the non-linearity effects of ultrashort pulses.
This ability to use fibres to deliver ultrashort pulses will enable a multitude of new benefits and possibilities for the integration of ultrafast lasers: ‘Integrators are now able to have a beam that is secured nicely in a flexible hose, which can then be integrated into a machine and directed towards a workpiece,’ Benabid said. ‘The use of a flexible hose will reduce the overall cost of ownership for an ultrafast laser, and in terms of engineering it will hugely simplify the design of ultrafast machines, making life much easier for system integrators.’
While Benabid has been working on the development of low-loss HCPCFs for almost 20 years, he told Laser Systems Europe that it is only recently that the technology is in its industrialisation phase. As a result, his company is now working directly with a lot of tier one ultrafast laser manufacturers to get the technology into fully functioning laser systems and into the hands of integrators and end users.
Benabid concluded: ‘From the market’s point of view, ultrafast lasers have reached a certain level of maturity. They have been demonstrated as an excellent tool for applications ranging from micromachining to surgery. Saying this, the technology is still yet to consolidate completely in terms of that maturity. One of the areas that still needs to develop further is the transporting of the ultrafast pulses, which is where GLOphotonics’ technology will come in.’
Catering to the customer
At the Laser World of Photonics this year, a panel discussion took place during the ‘International Laser Marketplace’ forum, in which attendees were provided with an update of the state of ultrafast laser development.
Two current key applications for ultrafast lasers were identified by the panel to be glass cutting and the dicing of silicon wafers – both of which are prime examples of how lasers are replacing conventional mechanical methods. According to Thomas Merk, executive vice president and general manager for industrial lasers and systems at Coherent, wafer dicing in particular has recently become particularly difficult to perform mechanically due to the wafers having become increasingly thin – down to below 100µm thickness. He explained that ultrafast lasers are now being used as a suitable alternative to these mechanical methods, and that this application of ultrafast laser technology is still in its early stages.
In addition to the continued trend towards higher average powers in ultrafast lasers, the key takeaways for me from the panel discussion were that for the uptake of ultrafast lasers to be increased – it was said that the ultrafast laser market volume is still less than $500 million, less than 10 per cent of the total laser market volume – manufacturers need to reduce the cost and increase the flexibility of their systems. They also need to further develop the applications of ultrafast laser technology, and provide close support to the integrators and users working with their systems.
Panellists discussed the ongoing developments and trends in the ultrafast laser regime at the International Laser Marketplace this year. (Image: Messe Munich)
‘The ultrafast laser industry is at an early stage still, and I think it’s beneficial to develop systems that have high flexibility and high average power, which can be adjusted more easily to the demands of the customer,’ said Berthold Schmidt, managing director of research and development at Trumpf. ‘I think what is happening in this industry is that single use lasers are going away, and I expect customers will work more closely on reducing capital spending in the future. Therefore, having lasers that after one production cycle can be readjusted and reused for a slightly different application is probably the way to go, so we have to go there.’ Schmidt added that he’s already starting to see a tendency toward variable pulses and adjustable repetition rates emerging in ultrafast laser technology.
Merk remarked that because industry is still transitioning from working with classical mechanical applications to working with lasers, in addition to reducing the cost of ultrafast lasers, laser manufacturers should also be investing money into further developing the applications of the technology. He added that creating a package around these applications, including the required beam delivery technology, would help system integrators and machine builders deliver a proven solution to end users, which would help increase the uptake of ultrafast lasers considerably.
In response to a question as to whether there is a current limitation in the know-how of ultrafast laser customers, Merk and Schmidt agreed that there was – with both explaining that they often have to work very closely with their customers to guide and teach them about the different applications – and help them select an optimised ultrafast laser solution. ‘I think that’s something I see as an important step forward... this more active and more involved business development,’ Schmidt remarked. On the application development side, he also emphasised that, for Trumpf’s own activities in this area, the firm currently has the same number of employees involved in ultrafast laser application development as it has working on laser welding – emphasising the importance and potential of ultrafast laser technology.
Other trends mentioned by the panel included the increasing variety of wavelengths being offered in the ultrafast regime, such as UV, which Schmidt noted will be important for future applications. Martynas Barkauskas, CEO of Light Conversion, added that he was happy to see more companies now providing UV femtosecond lasers, and said that this was definitely something that wasn’t happening a year ago. Lastly, Dr Qitao Lue, CTO of Chinese laser giant Han’s Laser, remarked that the emergence of 5G could bring with it new materials (more on page 20) that will require ultrafast lasers to process – specifically those offering picosecond pulse duration.