Steffen Diez (SD): Thanks a lot. I started at Heidelberg Instruments in 2000 as a software engineer. We were only around 30 people at that time and the business was just starting to pick up. Over the years, I had the chance to look into almost every part of our company. I worked in Research and Development (R&D), led several projects, did benchmarking and customer demonstrations, and then switched to Technical Sales and Product Management. In 2012, after one-year stay in Japan, I was invited to join the Technical Board, a group of experts managing all technical aspects of our company. This was when we decided to develop the first Maskless Aligner (MLA150). A few years later, in 2017, I joined the executive team of Heidelberg Instruments together with Martin Wijnaendts and Konrad Roessler. After a recent Heidelberg Instruments restructuring, I am now the Chief of Operations.

SD: The MLA150 was officially introduced with a big marketing campaign in May, 2015. After we developed the first working prototype, we decided to work together with a few selected users to evaluate and refine the tool. Main focus besides stability and bug-fixing was to optimize the usability of the system. The first beta-site was established at the CMi (Center for MicroNanoTechnology at EPFL in Lausanne). This facility was already our long-term customer and it has an outstanding reputation. Philippe Fluckiger, the head of CMi, didn't hesitate to enter a collaboration after the concept of the MLA had been presented to him. The CMi team helped us a lot to optimize the user interface and eliminated most of the critical bugs. The second beta-site was in the group of Andreas Fleischmann and Sebastian Kempf from the Kirchhoff Institute of Physics in Heidelberg. They are world leaders in the research on SQUIDS, superconducting quantum interference devices that are used as low-temperature magnetic field sensors. Here, we focused on transferring multi-layer processes from their old mask aligner to our MLA. These applications have very demanding requirements for overlay accuracy, and we had to enhance the performance of our tool in many aspects even further to match the required specification. This was a very successful beta-site testing – we were even able to increase the yield for these particular sensors by a factor of three. Harvard/CNS served as the final beta-site. Jiangdong Deng, the CNS Lab Manager, has also already worked with us previously. I really appreciate the fruitful collaboration between the beta-site users and our R&D team. Almost all requests from users were efficiently implemented. Finally, all collaboration partners purchased the tool after the beta-site period ended: A clear sign that the MLA has become a mature product.

SD: In the beginning, it was not so easy to convince the customers that the MLA was a real alternative to mask aligners. Long-time users of traditional lithography hesitated — understandably so. Many times, we heard that mask aligners expose much faster. We insisted that the overall cycle time is a more important parameter, and this time is much shorter with the MLA compared to a mask aligner, especially taking into account the mask fabrication time (needed for the mask aligner). But the tool speaks for itself, and soon MLA150 gained a very good reputation, and the news spread. The R&D community in the USA has accepted MLA150 very fast, and eventually we saw more and more interest in Europe as well. Over the time, we implemented quite some extra features following our customers' requests. MLA is attracting increasingly more attention from industrial companies as well. Most of them have special requirements which cannot be fulfilled by mask-based technologies, some just produce low volumes, others benefit from the low running costs of the maskless technology.

SD: We planned to develop an industrial version of a maskless aligner back when we began working on the MLA concept. Due to limited R&D resources, we prioritized the work on the MLA150. This turned out to be a good choice, because we could fully develop the core technology and learn from our industrial MLA150 users what the tool really needed to improve in order to fulfil their requirements. The MLA150 is optimized for R&D: The software is designed for handling small lots and for easy tracking of project status, not for high-volume production. The exposure time is excellent for research applications, but it is not fast enough for commercial production. There was no automation, either. MLA300 is a completely new machine, so we could meet all the needs of our industrial customers, which extended far beyond just higher throughput:
● A massively increased throughput as compared to the MLA150;
● The MLA300 can be fully automated, including substrate handling and loading, and interfacing to manufacturing execution systems (MES);
● It is a fully integrated and modular machine with a minimal, yet flexible footprint. This concept simplifies service and allows a fast "plug and play" installation;
● The software interface and usability concept were developed with the help and feedback from a customer, this time from industry.
Our R&D team really pushed the limits of direct-write lithography. We are very proud of the MLA300. It is the most advanced tool we have ever built. It is unique in many ways, and it is also the only production-ready solution for industrial micro-fabrication currently available in the market.

SD: After some calculations, we realized that we couldn't achieve the required speed with only one optical engine. So we embarked on one of our most demanding R&D projects ever: The development of a fully integrated optic module. Our optical designers and construction engineers created an opto-mechanical masterpiece. The MLA300 system is designed for the integration of up to four optic modules. It is even possible to integrate modules with different wavelengths or different resolutions. That means you can customize the MLA300 according to the specific application requirements in terms of resolution, write-speed and substrate handling.

SD: At the moment, we focus on applications where maskless lithography has a significant advantage, for example, warped substrates. Such deformations commonly occur in applications which use a stack of materials with different thermal expansion. The warpage can measure up to 300 µm. Our system is able to compensate warpage with the real-time autofocus system. Other niche applications demand a flexible adaption of the pattern, a requirement which is efficiently met by the MLA300 system. We aim to ultimately place the system in high-volume applications such as wafer-level packaging, MEMS fabrication, or LED production. One of our customers has already begun using the MLA300 in a pre-production environment. We are now integrating the tool into their production line to make it their standard exposure machine. Since the official launch of the MLA300 in November 2019, we have already received multiple orders for the system. The fast order intake confirms that the flexibility and superior exposure quality on challenging substrates is a large unmet need for industrial production customers.