editor's blog
Subscribe Now

Will There Ever Be Cake?

It’s as if there’s this great party coming, and everyone is working at breakneck speed to make sure that all the details are in place to make it the best party ever. And the highlight is some amazing cake that everyone’s been hearing about. Let’s call it an Extreme Ultra Velvet cake.

This cake is so special that everything needs to change. The flatware must be swapped out so that no surface ions contaminate the exquisite flavor. The dishes must be made out of the most perfect reflective material instead of crystal so that your eyes take in the delicate crumb as seen from as many angles as possible.

Those and a million other details are being readied. But there’s only one problem: where’s the cake? We’ve been waiting for the cake for a while now, and the party has been postponed a couple times. A few people have had samples, and they swear it’s everything promised and more, but scaling this amazing pastry up to the size necessary to feed increasingly hungry partygoers has proven rather more difficult than expected.

And at this year’s SPIE Litho conference, faces seem just a bit longer. You can feel the doubt creeping in: will the cake ever show up? Or will it arrive too late, after everyone has given up and ordered pie instead?

We’ve watched over the last couple years as companies like Cymer report their results. If numbers are to be believed, they’re leading the bake-off so far (although it seems as if numbers should be viewed cautiously in this kitchen), but they’ve gone from 50 W to 55 W over the last year, if pure power is what you’re gauging by. As a reminder, 100 W has been the bare-minimum target for high-volume production.

It will be noted that many of the other critical issues we reported on last time have improved, at least with respect to Cymer:

  • At 40 W, dose stability is now below 0.2% (shouldn’t that be dose instability at a low number?)
  • This is for six one-hour runs (one lot per run) over a total wall time of 8.5 hours
  • Duty cycle is 100% for one die, 92% within a wafer (the gap being step time – note that as power goes up, the exposure time will go down, while step time remains constant, meaning that duty cycle will actually decrease with increasing power); there’s a longer gap between lots to simulate the transit time as one lot exits and a new one enters. In other words, there are no pauses simply for the machine itself to cool off or catch its breath (which it certainly gets to do during the stepping and inter-lot breaks).
  • 196 equivalent wafer exposures yielded at 99.99%
  • At 55 W, yield is at 97.5% over a one-hour run
  • Availability has been at or above 65% over the last three months (that’s actually a bit down from the 70% reported last time)
  • All of this was done with new collector protection; they continue to bombard the collectors, above 75 billion pulses so far and continuing – this corresponds to about 8 months of life with no loss of reflectivity.

In other words, the 50 W reported before is now 50 W (OK, 55 W) with much of the process stability smoothed out. One can hope that fewer issues crop up as power is increased, but clearly the 55-W numbers, as compared to 40 W, reflect the fact that higher power may not come without additional challenges.

But… all of that aside… the question remains: when do we get to 100 W? In fact, Cymer is targeting 250 W as a realistic target (as opposed to a bare-minimum number). As before, there are two ways to do this: increase laser power and conversion efficiency.

With respect to the drive laser, currently at 15 kW, Cymer will target 80-W power with a 24-kW laser, 125 W with a 31-kW beam, and will use 43 kW to get to 250 W. That last step will also require that they go from their current 2% conversion efficiency to 3%. Not clear to me if they know exactly how they’re going to do that. And note that there are no dates associated with these milestones. 250 W certainly won’t be happening in 2013.

So, yes, there has been useful progress over the last year. Nonetheless, watching the various presentations and hearing some of the discussions amongst attendees, the mood is palpable: will we in fact ever get to eat any cake?

Leave a Reply

featured blogs
Oct 23, 2020
Processing a component onto a PCB used to be fairly straightforward. Through hole products, a single or double row surface mount with a larger center-line rarely offer unique challenges obtaining a proper solder joint. However, as electronics continue to get smaller and conne...
Oct 23, 2020
[From the last episode: We noted that some inventions, like in-memory compute, aren'€™t intuitive, being driven instead by the math.] We have one more addition to add to our in-memory compute system. Remember that, when we use a regular memory, what goes in is an address '...
Oct 23, 2020
Any suggestions for a 4x4 keypad in which the keys aren'€™t wobbly and you don'€™t have to strike a key dead center for it to make contact?...
Oct 23, 2020
At 11:10am Korean time this morning, Cadence's Elias Fallon delivered one of the keynotes at ISOCC (International System On Chip Conference). It was titled EDA and Machine Learning: The Next Leap... [[ Click on the title to access the full blog on the Cadence Community ...

featured video

Demo: Inuitive NU4000 SoC with ARC EV Processor Running SLAM and CNN

Sponsored by Synopsys

See Inuitive’s NU4000 3D imaging and vision processor in action. The SoC supports high-quality 3D depth processor engine, SLAM accelerators, computer vision, and deep learning by integrating Synopsys ARC EV processor. In this demo, the NU4000 demonstrates simultaneous 3D sensing, SLAM and CNN functionality by mapping out its environment and localizing the sensor while identifying the objects within it. For more information, visit inuitive-tech.com.

Click here for more information about DesignWare ARC EV Processors for Embedded Vision

featured paper

Designing highly efficient, powerful and fast EV charging stations

Sponsored by Texas Instruments

Scaling the necessary power for fast EV charging stations can be challenging. One solution is to use modular power converters stacked in parallel. Learn more in our technical article.

Click here to download the technical article

Featured Chalk Talk

TensorFlow to RTL with High-Level Synthesis

Sponsored by Cadence Design Systems

Bridging the gap from the AI and data science world to the RTL and hardware design world can be challenging. High-level synthesis (HLS) can provide a mechanism to get from AI frameworks like TensorFlow into synthesizable RTL, enabling the development of high-performance inference architectures. In this episode of Chalk Talk, Amelia Dalton chats with Dave Apte of Cadence Design Systems about doing AI design with HLS.

More information