Digital Optics – a Discipline at the Crossroads of Digitization

April 5, 2016 Kerstin Maier

Digital Optics – a Discipline at the Crossroads of Digitization

Digital Optics – a Discipline at the Crossroads of Digitization

Think back to your days as a student in high school: you’d probably agree with the statement that optics is an integral part of physics. Of course you’d be right, so long as we treat light as photons or an electromagnetic wave. Things change, though, when it comes to describing optical systems. We can conceive of them like an analog computer. Think about the traditional description from Fourier optics on optical imaging. When optics are digitized, parts of this analog computer are replaced by digital algorithms. And right now a lot is happening at this interface.

Here exponential growth is the norm

Let’s get one thing straight: digitization is only possible thanks to optics. Driven by Moore’s law – according to which the number of transistors on an IC doubles approximately every two years – a state of the art NAND chip consists of more than 100 billion transistors . Gordon Moore proposed his famous law in 1965 on the basis of a 64 transistor chip, meaning the number of transistors has on the long run actually doubled every 1.6 years. The sheer consistency of this exponential growth – jumping from 2 to the power of 6 = 64 to 100 billion in 50 years – continues to astound me.

Optical lithography or photolithography is one of the main process-steps in the production of electronic circuits. It has consistently reduced the size of the smallest structures on the IC reaching today the 14nm node and the 10nm node at development (http://www.eetimes.com/document.asp?doc_id=1328866). This is currently achieved in volume production by using immersion lithography at a wavelength of 193 nanometers. And you might be asking yourself: how can someone create 14 nm structures with a wavelength of 193 nm? Doesn’t Abbe’s resolution limit apply? Here’s a simple calculation: the minimum achievable structure width is 0.25 x wavelength / NA and thus 38 nm for 193nm wavelength and water immersion. However, the resolving power is increased by an additional factor by using tricks similar to those in super resolution microscopy, such as by a factor of 2 in double patterning lithography. But the number of necessary photomasks – the templates for the circuits – also increases by the same factor. In the next step in the game, making it without the need of multipatterning lithography, is  EUV lithography, which is working at 13.5 nanometers. EUV requires mirrors and a vacuum, i.e. this optic is constructed in a fundamentally different manner.

Four digital trends and a symposium
But let’s get back to the consequences of digitization. We’re seeing four obvious trends in digital optics, which we deal with in the following four blog entries:

  1. Computational Imaging
    Ptychography and Lightfield Methods
    Digital Microscopy
  2. Computer Vision and Machine Learning
    Image Registration and Segmentation
    Object and Anomaly Detection
    Deep Learning
  3. Large Data in Optics
    Data Transfer and Storage
    Large-scale Data Analysis and Visualization
  4. Virtual and Augmented Reality
    AR / VR Glasses
    Light Field Displays

In order to better illuminate these trends, their technical prerequisites and their implications, ZEISS is hosting a symposium in summer 2016: But this isn’t going to be your typical scientific conference. Following the keynote lectures on current trends (Laura Waller from Berkeley on computational imaging, Ingmar Posner from Oxford on machine learning and David Bohn from Microsoft on AR optics), the conference participants will discuss the current state of technology and the need for further research in three break-out sessions. The groups will then record their ideas in a white paper.  Together the white papers will be the subject of an accompanying podium discussion.

But the event’s grand finale will take place after the symposium has ended when the ZEISS Research Awards 2016 are presented to Fedor Jelezko from the University of Ulm and Jörg Wrachtrup from the University of Stuttgart. They’ll receive these awards “for their outstanding work on quantum technology with optically addressable spins in diamond.”

Addressing a different but equally exciting development in optics, Bob Byer from Standford University will give a speech at the awards ceremony. The topic will be the ‘ultimate in optics,’ namely the recent evidence of gravitational waves obtained using the LIGO interferometer. The three areas – quantum physics, gravitational waves and digitization – are currently a hot topic in optics which we’ll definitely still be working on in the future. We’ll have to wait and see what non-linear interactions these three thematic focal points reveal on 23 June in Oberkochen.

Dr. Michael Totzeck

 

The post Digital Optics – a Discipline at the Crossroads of Digitization appeared first on Optics Beyond.

Previous Article
Computational Imaging
Computational Imaging

The theme of today's blog post is computational imaging. Computational imaging is the systematic enhancemen...

No More Articles