Kulicke & Soffa wants to conquer the back-end lithography market step by step, in close collaboration with customers. That’s the way things work best in advanced packaging, argues TL Cheam, the VP in charge of strategy at the Singapore company.
How will things in the upcoming market for back-end lithography work out? Will we see consolidation? What are the chances for Kulicke & Soffa, the back-end equipment specialist that bought the Dutch litho startup Liteq, shortly after it acquired Philips spinout Assembléon in late 2014?
To discuss all this, Bits&Chips sat down with Tong Liang Cheam, vice president of corporate strategy at K&S. Cheam was also instrumental in the acquisition of Assembléon at the end of 2014. Seven years later, the takeover of the pick-and-place company resulted in the formidable market success of the Pixalux and Luminex equipment for the placement of mini and micro LEDs in the display market. Both machines are developed by the R&D team in Eindhoven.
While visiting Eindhoven to talk about pick-and-place technology, Cheam also learned about a tiny startup company that developed a back-end stepper. “It was around 2015, 2016. We just acquired Assembléon. I found Liteq very interesting. It was offering the technology of the future and not the traditional mercury lamps,” says Cheam. “It was the laser source that caught my interest. That gave the potential to deliver a litho tool with high throughput, which is very important in back-end.”
“The laser as the main differentiator really interested me because I saw advanced packaging evolving in a lot of different technologies from simple flip chip to fan-out wafer-level packaging, heterogeneous integration and chiplet technologies that are currently heavily discussed. Chiplets need to be mounted on certain substrates or even on silicon and you definitely need high-resolution lithographic tools to make accurate connections between the different dies. I see Liteq playing a big role there in the near future. We didn’t buy the company because of the current situation but really for future markets.”
Liteq – now the name of K&S’ litho portfolio – uses a 355-nm Nd:YAG laser source instead of the g, h and i-line from mercury lamps that are common in the back-end market. This choice allowed for a much cheaper aerial illuminator and lens, resulting in a significantly better cost of ownership for customers.
Liteq made this choice a decade ago. What’s the current competitive edge for its steppers in the back-end market?
“The mercury lamp has a short life span. It’s more like weeks than months. You need to change that lamp every time before it dies. If you don’t, you run the risk of a lamp explosion, which compromises the wafer being processed and also creates a health hazard. Using a laser is a big advantage because it only needs maintenance once a year. This means less downtime.”
“Lasers also allow high productivity. We have 27 to 60-watt laser versions. It gives us high productivity with 75-95 wafers per hour. Lasers will also give us the possibility for high resolution combined with a large depth of focus, in other words, a large process window. Right now, we’re looking at 2 microns and below.”
You stand out by using laser light in your Liteq steppers, compared to competitors who use mercury lamps. Yet you designate the system as i-line? Why is that?
“We call it the i-line stepper because the 355-nm laser light is very close to the i-line wavelength of 365 nanometers. The i-line resist used in the market is very sensitive to 355-nm laser light. That’s the reason why I call it the i-line stepper.”
Canon and Nikon are also eyeing the back-end litho market. What’s your view on that?
“We see Canon and Nikon moving from the front-end to the back-end market. We do see them as competition. But their systems still have a high average selling price. The Liteq systems are really designed to cater to the back-end advanced packaging market. By design, we’re more cost effective than a cost-down version of a front-end litho machine. Thus, the Liteq system has a much better cost of ownership.”
Can you tell me about your strategy in litho?
“In litho, we’re working with several customers, all under NDA. But you can compare our approach with what we did with our APAMA thermal compression bonders for heterogeneous or chiplet integration. We started to develop thermal compression bonding technology back in 2013 but didn’t get any revenue until about the end of 2018, the beginning of 2019. Right now, we’re recognized as one of the leaders in TCB. That’s because we work with key customers, from R&D to making sure that the bonder we’ve designed meets their requirements and their application. That takes time because we don’t develop for the current node but for N plus 1.”
“In a market where there’s a strong incumbent, it’s very difficult to compete in the N node. You need to leapfrog to get ahead of the competition. That was our strategy in thermal compression bonding, and we emerged as one of the recognized technology leaders in TCB by the biggest CPU manufacturer in the world. For the lithography tool, we’re working with several customers on the next-generation kind of needs, not for the current generation. So, it will take time with our customers to develop this.”
There are about ten players in back-end litho.
“A lot of interesting players are trying to get into the market, but I don’t see that they have the technical capabilities like K&S has to meet the requirements of specific customers in the advanced-packaging space. They may also not have the relationship with customers that we have. That’s important for our knowledge of future needs. We have that because we work with key customers in the advanced-packaging space.”
Regarding K&S’ technology development with customers, how do OSATs like ASE and Amkor relate to foundries like TSMC and IDMs like Intel?
“When I look at the development of advanced packaging for the next generation, foundries and IDMs have an edge over the OSATs. But the OSATs may eventually give you a higher volume. So for us, they’re equally important. It really depends on the technology phase that you’re in.”
We see an upcoming trend in panel-level integration. Onto is already selling steppers for panels.
“Certain customers prefer to stay at the wafer level because, in their whole production lines, they’re dealing with round, not square or rectangular. For big panels up to 600 by 600 mm, we definitely need to do development work. But it really depends on the customer we work with. They’re defining what we’ll be offering. We take it step by step, work with a specific customer, meet their requirements and move on to the next customers. Because of all the competing technologies in advanced packaging, we really don’t know which one will be the winner. That’s why we tend to work with specific customers rather than trying to have a generic tool to serve them all. If you try to do that, you may not win any of them.”
Is K&S considering a panel-level litho tool?
“We’re currently discussing panel-level litho with some customers, but it’s not at a point where we’ve taken a decision. The litho customers we’re engaging today are the right type of customers in the area that will eventually give us the volume we envision. They’re driving the next-generation technology. It’s very similar to the way we developed the business for thermal compression bonding. It took us a long while, and our approach in the back-end litho market is similar.”
Will there be a shakeout in back-end litho like we saw in front-end litho in the eighties?
“It will be similar in nature. Some of the players that are coming up right now will drop off because they don’t have the technology or the resources. Like in the market for wire bonding today, there will be only a few key players. I do foresee that there will be one or two or three big ones. And of course, I’d like to see K&S be the one at the top.”