Postdoctoral Researcher in Data-driven X-ray Nanoimaging for Semiconductor Metrology

Mission

The performance of X-ray microscopy has improved significantly over the last decade, enabling synchrotron-based X-ray imaging to resolve individual transistors in modern semiconductor devices. As the industry moves to 3D transistor architectures and stacked, hybrid-bonded devices, non-destructive, high-resolution 3D imaging is becoming a necessity. Unfortunately, current 3D X-ray nanoimaging methods are prohibitively slow for practical semiconductor metrology and inspection applications. 

Unlike most samples placed under a microscope, microchips are manufactured objects that follow strict design rules. This allows us to shift from the standard imaging question of "what is the structure of this sample?" to "how does the structure differ from the known design?". The second question can be answered with far fewer measurements than the first. The newly established Laboratory for Nanoscale X-ray Metrology is looking for a motivated postdoctoral researcher to turn this idea into a working method.

This is one of two postdoctoral positions opening in the group. This position focuses on high-throughput imaging of static structural defects, while the companion position focuses on time-resolved imaging of dynamic processes. Each postdoc will lead their own independent research direction, with an opportunity for close collaborative work.

Main duties and responsibilities

You will develop an acquisition and reconstruction framework that uses the known geometry, composition, and layout regularity of semiconductor devices as prior information to detect deeply buried defects from the smallest possible number of X-ray measurements. Our pipeline reconstructs each projection from thousands of diffraction patterns, followed by volume reconstruction from projections measured at many rotation angles. These are two inverse problems, each with its own sampling requirements.

The questions we want to answer:

• What is the minimum number of diffraction patterns and projections needed to identify a defect? 
• Can we detect defects from diffraction patterns alone, bypassing projection reconstruction?
• Can we reliably identify defects without ever resolving them?
• What is the best way to incorporate the known design information into the imaging process?

The primary focus will be on algorithm development, followed by their demonstration at synchrotron experiments. You will start working with real already existing experimental datasets and will have the opportunity to extend our established reconstruction pipeline, and integrate with new data acquisition methods. You will also get a chance to perform experiments at the cSAXS beamline of the upgraded Swiss Light Source synchrotron, as well as other facilities around the world.

Profile

We are looking for a person with:

• PhD in physics, engineering, computer science, or a related field.
• Strong scientific programming in Python and experience with GPU processing of large-scale datasets.
• Experience with inverse problems and 3D reconstruction methods for tomography, laminography, or a closely related modality.
• Excellent written and oral communication skills in English.

We also value applications from people with the following experience (or similar):

• Coherent diffractive imaging, especially ptychography.
• Sparse sensing, optimization, or Bayesian experimental design.
• Machine learning for imaging.
• Synchrotron experiment experience.
• Semiconductor devices or metrology.

We offer

We offer a fully funded postdoctoral position for two years (with a possibility of an extension) in a unique dual environment. Contracts are issued for one year and renewable annually according to standard EPFL procedures. You will be employed by EPFL, one of the world’s leading technical universities. Physically, you and the team will be based at the newly upgraded Swiss Light Source (SLS) synchrotron located at the Paul Scherrer Institute (PSI). You will be a core member of the SNSF Starting Grant project NEXUS-3D, working at the intersection of applied physics, high-performance computing, and semiconductor metrology. 

The successful candidate will have the opportunity to develop their own independent research direction and work alongside leading synchrotron scientists and computational X-ray imaging experts. You will receive close mentorship and support for professional development such as attending summer schools, international conferences and conducting experiments at synchrotrons around the world.

Informations

Only applications submitted through the online platform are considered. You are asked to supply a single pdf file containing:

• A brief cover letter describing your research background and fit for this position (max 2 pages).
• A detailed CV.
• A research statement highlighting your past research achievements, current interests, and potential research directions for this project (max 3 pages).
• Contact details for 3 referees.

If you would like to also be considered for the companion position on time-resolved imaging of dynamic processes, please say so in your cover letter and explain which you would prefer and why. A single application is sufficient.

Applications will be reviewed on a rolling basis until the position is filled. The expected starting date is flexible and can be discussed. For any further information, please contact Prof. Tomas Aidukas (tomas.aidukas@epfl.ch).

Contract Start Date : 01.11.2026

Activity Rate : 100.00 

Contract Type: CDD

Duration: 2 years, with the possibility of renewal 

Reference: 2417