Opportunities and challenges for laser technology in the electronics sector
The electronics industry employs lasers for a wide variety of purposes, from cutting and scribing to drilling and structuring, and – according to a recent report from management consultant firm McKinsey & Co – its use of these applications is likely to grow steadily over the next few years.
A current shift from mechanical processing means to laser-based methods is particularly noticeable in the semiconductors and advanced packaging sectors of the electronics industry, according to laser manufacturer Coherent’s director of marketing, Dirk Müller.
He says that this shift is being driven by a number of factors. Firstly, consumers and original equipment manufacturers (OEMs) alike want the latest electronic devices – such as smartphones, wearables, virtual-reality devices, sensors and home-automation equipment – to be smaller and more energy-efficient than their predecessors. This means that components, such as printed circuit boards (PCBs), need to be correspondingly smaller, thinner and have more compact, densely packed features. Müller explains: 'As features get smaller, mechanical means hit their limit of capability. For example, drilling mechanically is displaced by drilling with a laser, because making a hole that is 50 or 80μm in diameter is much easier to do with a laser beam than it is with a mechanical drill.’
Secondly, companies operating in the semiconductor and advanced packaging sectors are becoming increasingly keen to reduce the environmental impact of their operations. Müller explains that these companies 'are looking to replace certain processes that, for example, utilise a lot of water, whether it be maybe a lapping process to make something flatter, or a cutting process that is using a dicing saw. Laser-based technologies can be an alternative’. The use of laser-based processes might also be able to reduce power consumption in comparison with mechanical methods, or eliminate the need for potentially dangerous chemicals, such as the hydrofluoric acid used to etch holes in glass.
Finally, laser-based technologies are easily adaptable, both in terms of the design features they can be used to create and the materials – including metals such as aluminium, stainless-steel, iron, and titanium, and non-metals such as plastics, composites and ceramics – that they can be used to process.
Despite these advantages, laser technologies are often seen as a last resort for companies that have run into problems that cannot be solved otherwise, according to Müller. He says that in order to appeal to companies used to the simplicity and reliability of mechanical methods, suppliers of laser-based technologies need to prove that their systems are dependable and cost-effective.
The electronics industry has very stringent requirements for minimising downtime. Müller says: 'So, as you might imagine, in a $20 billion semiconductor fabrication plant, if a line is not running because a system is down, it is very expensive'. Suppliers of laser-based technologies must therefore have strategies in place to mitigate this risk. They might need, for instance, to ensure that they have trained maintenance personnel near the plant, together with a safety stock of spare parts. They might need to rehearse for worst-case scenarios, so that if a failure happens in the field, their personnel can react quickly and efficiently.
The design of a system might have to be tailored, says Müller, so that 'it will run without interruption for 20,000 hours, or two and a half years, before it has to be touched’.
Proving the cost-effectiveness of laser-based technologies can also be challenging. Müller says that while a mechanical tool for drilling holes in PCBs might cost $100,000 to $150,000, a laser-based system for carrying out the same job might cost $650,000 to $750,000.
The high capital expenditure (CAPEX) associated with the laser-based system, however, can be offset by its increased efficiency and lower cost of ownership compared with the mechanical option. The drilling tool, for example, might have up to six spindles, each able to drill twenty holes per second. The laser-based system will be able to drill 2,500 holes per second, increasing throughput significantly. Further, the spindles on the drilling tool might have to be replaced every one hundred seconds after drilling 2,000 holes, driving-up costs related to consumables. By contrast, the laser-based system might need only the bare minimum of maintenance over a two-to-three-year period.
These benefits are easy to sell to companies that know they will be shipping to large chip manufacturers for extended periods of time. But, says Müller, for companies that only have an order horizon of six-months-to-a-year, it is more difficult to convince them that an investment in a laser-based system will pay-off.
It is therefore incumbent on developers of laser technologies to continually refine their systems so that they can provide more benefits to their customers.
Take the depannelling of PCBs, a task for which lasers are widely used. The miniaturisation of PCBs has led to them being more densely packed, with components located closer to their edges. Further, to increase yield, PCBs are being placed increasingly closer together on panels.
The miniaturisation of PCBs has led to them being more densely packed, meaning that the process used to cut them from panels needs to be highly accurate and must create a narrow kerf (Image: Coherent)
This means that the process used to cut them out needs to be highly accurate and must create a narrow kerf. Additionally, the process must not generate excessive mechanical stress or heat that might affect the surrounding circuitry, or debris that would need to be cleaned-up. These demands are driving manufacturers away from mechanical depannelling methods to laser cutting, but many of the technologies used are reaching the edges of their technological limits.
Conventional carbon dioxide (CO2) lasers emit infrared (IR) radiation and cut by heating the material, but they generate a large heat-affected zone (HAZ), which can carbonise the core of the PCB, and this longer wavelength creates a larger spot-size than alternative ultraviolet (UV) wavelengths, creating a wider kerf. UV-based diode-pumped solid-state (DPSS) lasers also create a much smaller HAZ and produce substantially less debris than CO2 lasers. They are not as fast, however.
Looking to solve these problems, Coherent has explored the use of its ‘Avia’ nanosecond pulse-width, high-pulse-energy UV DPSS laser for cutting a variety of PCB materials. Based on this work, the company claims to have developed a PCB-cutting method that delivers a small HAZ, a high-quality cut edge, a narrow kerf and high production throughput.
For the production of microLED displays, Coherent’s UV transfer process can be used for laser lift-off, laser-induced forward transfer and the trimming and/or repair of defective pixels (Image: Coherent)
A key element of this technique is a proprietary method for controlling the laser pulses so that they are delivered to the work surface in such a manner that heat build-up is avoided. Because thermal damage is absent in this approach, it is possible to use a laser with substantially higher pulse energy when cutting thicker materials (of 1mm and above).
The company has also worked to ensure that the new method is reliable. Solid-state lasers emit IR light, so a third harmonic generation (THG) crystal is incorporated into them to convert this to UV. These crystals absorb UV, however, causing them to degrade over extended periods of time. To mitigate this issue, Coherent previously incorporated mechanisms that would move the THG crystal periodically before it failed in a particular location. Such mechanisms increase the cost and complexity of the laser, however, and create subtle changes to the output power and other parameters of the beam that can affect the process each time the crystal is moved. By contrast, Coherent says that is has developed a THG crystal ‘PureUV’ of such high quality and low UV absorption that it has a maintenance-free lifetime of 20,000 hours at a single spot. This ensures process consistency and cuts downtime.
Developers of laser systems must also keep their eyes out for new applications for their technology in the electronics industry. Müller provides microLEDs (µLEDs) as an example, which can be used to produce large, brilliant high-resolution displays. In contrast to miniLEDs, where the sapphire growth wafer might stay with the LED, microLEDs need to be released from the wafer on which they are grown and are only a few micrometres in thickness. As a result they are very fragile, meaning non-mechanical processing techniques are required to lift them off. Deep-UV (DUV) excimer lasers, with their large available pulse energies and micron precision, are suitable candidates for facilitating these transfer steps in mass production (see Oliver Haupt and Jan Brune’s article for further details).
In laser-induced forward transfer, a large-area laser beam passes through a photomask so that only specific dies are released and pushed onto the display substrate. A uniform so-called top-hat beam is critical to the perfect placement of the dies (not to scale) (Image: Coherent)
Müller says that another growth area for laser processing, being driven by the electric vehicle market, is the annealing of semiconductors to produce power chips. The annealing process lowers the resistance on the backside of these chips and increases their performance. Using lasers, higher annealing temperatures can be generated in shorter spaces of time than would be possible using flash lamps or ovens – increasing throughput.
But what of the future? Unsurprisingly, Müller expects that laser technologies will have to continue to evolve to keep up with the need of the electronics industry to miniaturise its products. He adds: 'In some sense, quantum computing is also enabled by lasers. All the modalities that use neutral atoms or ions, or nitrogen vacancies in diamonds, they all use lasers. And so, if at some point, the silicon-based computer runs out of steam, there is still quantum ahead for us.'
Lead image: Shutterstock/raigvi