A 10 Nanometre Micrometer Goes to Work on Superconducting Strip

Two applications published by Micro-Epsilon are worth reading together, because between them they describe what in-line measurement has to survive to be useful.

The first is the optoCONTROL 2700-40 LED micrometer, specified at 10 nanometre resolution, repeatability of 0.1 micrometres or better, and a 5 kHz measuring rate. It is applied to superconducting strip: multi-segment measurement across stainless steel strip 12 millimetres wide, resolving partial segments 3 to 4 millimetres across with 0.5 millimetre gaps between them. The measurement is optical and non-contact, which is not a convenience — the strip carries sensitive PVD coatings that a contacting gauge would damage.

The specification that decides whether such a gauge closes a control loop or merely logs data is the interface list, and it is stated plainly: measurement data passes to control systems in real time over EtherCAT, PROFINET, EtherNet/IP or analogue signals. A 5 kHz optical gauge that can only report over USB is a laboratory instrument. The same gauge on EtherCAT is a feedback element.

The second application puts capacitive displacement sensors inside semiconductor process chambers to measure wafer planarity and coplanarity for hybrid bonding, covering both die-to-wafer and wafer-to-wafer processes. These use a vacuum-compatible design, which is the harder engineering problem it sounds like: no outgassing from cable jackets or potting compounds, no lubricants, and a sensor body that tolerates the chamber's thermal cycling without drifting out of calibration.

Hybrid bonding is where advanced packaging currently lives, and coplanarity is one of its defining constraints — the surfaces being joined must be flat and parallel to a degree that makes conventional mechanical gauging useless. Measuring it inside the chamber, rather than inferring it afterwards from bond yield, is the difference between process control and post-mortem.

The common thread is that both applications are measurements that cannot be taken by stopping the process, touching the part, or removing it to a laboratory. That constraint, not the resolution figure, is what defines an in-line gauge.