Original English comment: Originally published in Mark & Technique. (DE)
"You can overcome deficiency!"

How can we significantly improve the throughput and yield of semiconductor manufacturing? And not just by a few percent, but can a particular machine improve it by 40-60% all at once? The answers to these questions will be of the greatest interest, especially in times of product shortages.

‍

Prakash Murthy, CEO, CTO, and founder of Atonalp, answers with an advanced metrology system based on mass spectrometry, which is done directly in the process chambers of today's cluster tools. For example, more than 10 chambers can be placed on a platform for deposition or etching of specific layers, such as conductive materials or dielectrics. In-situ measurement terminology. An "in-situ" measurement.

‍

Currently, however, chipmakers generally respond with so-called "in-line" measurements, which do not measure what is actually taking place in individual process chambers. Instead of measuring what is actually taking place in individual process chambers, they determine, for example, how long it takes for a given process to be completed. These times are then monitored "in-line," including appropriate safety buffers. This is not very effective and takes time.

‍

Prakash Murthy says, "Because of these many shortcomings, we developed a measurement system based on mass spectrometry from scratch, ready to use on the spot."

‍

It takes two to three years before the increased production capacity from a new fab is felt in the marketplace, and in Aston's case, throughput could be noticeably higher in a matter of weeks.
The result is "Aston." This is in honor of Francis William Aston, the chemist and physicist who invented the mass spectrometer in 1919 and was awarded the Nobel Prize in Chemistry in 1922. Such naming is obligatory. Prakash Murthy is confident that it will live up to the high expectations that come with it." In the case of cleaning process chambers, the time savings can be more than 80%." The throughput of individual machines, especially those for deposition, can be increased by more than 40 percent with "Aston" in some cases." Prakash Murthy says, "Even a 1 percent improvement in throughput for a seemingly small fab can save tens of millions of dollars a year." We can save manpower!"

‍

This is because it takes two to three years for the market to realize the increased production capacity from the new fab. Even when a new line is built in an existing factory, it takes at least a year, if there is space for a clean room. However, Aston can significantly improve throughput in just a few weeks.

‍

Risto Puhakka, president of VLSIresearch, is also convinced of Atonarp's breakthrough. This is a completely new way of doing things," he said. A 10% increase in throughput in a NAND fab is already a big step forward," he said in an interview with Markt&Technik. For the newest fabs, this is a very attractive prospect. And the same goes for existing fabs that are being upgraded for the next process node. After all, with "Aston," existing machines at the site can be upgraded without any problems.

‍

By 2021, 26,000 process chambers are expected to be in operation, which equates to $1 billion in metrology sales. 26,000 process chambers are expected to be in operation by 2021, which equates to $1 billion in metrology sales. The plant currently has over 100,000 process chambers in operation, and is expected to generate another $3 billion+ in sales by 2024. Currently, Atonarp is primarily targeting CVD and etch process applications, which are currently growing at an average annual rate of 13%.

‍

Risto Puhakka, President of VLSIresearch. said, "Atonarp's approach is truly novel. In particular, it can significantly increase the throughput of complex etch processes, such as those required to produce 3D NAND memory ICs."