// Showcase

Optical loss budget for laser soldering through a viewport

A vacuum measuring cell is sealed by melting a solder ring on a metallised ceramic housing - with a laser, from outside, through the viewport of the vacuum chamber.

Photonics & Laser

Optical loss budget for laser soldering through a viewport

Photonics & Laser

01Starting point

A vacuum measuring cell is sealed by melting a solder ring on a metallised ceramic housing - with a laser, from outside, through the viewport of the vacuum chamber. That leaves exactly one route for the energy: light. Everything between the diode and the joint is either transmission or loss. The obvious part of the sum was quick: two lenses and a viewport cost 23.5 % of the power between them. The uncomfortable question came next - how much of what arrives does the part actually take? That is not a setting, it is a material property. And nobody knew it.

02Approach

We measured instead of estimating. Both joint partners went into the spectrometer: the ceramic housings from 500 to 2000 nm, the solder rings from 400 to 1070 nm, in 1 nm steps, five housings from two lots and solder rings in two surface states. The reference channel ran alongside: over 500 to 1100 nm the 100 % reference reads between 99.22 % and 100.91 %, mean 100.02 % - so the measurement uncertainty is settled before anyone starts arguing about results.

03Result

The result reordered the loss budget. In the wavelength band of the high-power diode, the ceramic housing reflects 43.7 to 59.5 %. Of 150 W at the instrument, 114.75 W survive the optics - and of those, only about 48 to 58 W actually enter the material. Roughly a third of the nameplate. The three optical surfaces cost 23.5 %; the last surface, the product itself, costs about twice that. Second finding, and this one saves trials: plasma cleaning, intended to improve wetting, raises the reflection of the solder ring by some 16 percentage points. A quarter of the absorbed power is gone, at an identical laser setting. Anyone comparing cleaning methods at constant power is therefore not comparing like with like.

What's inside

A loss chain worked through from the diode output into the ceramic, backed by our own spectral measurements of both joint partners rather than by handbook values. Plus the scatter between nominally identical parts - 3.5 percentage points at 800 nm, 1.1 at 940 nm - as a realistic floor for process variation. And an answer to the wavelength question: reflection rises monotonically with wavelength for both materials, and the near-infrared band of available high-power diodes sits at the unfavourable end of both curves.

And what happened next

The 150 W stayed, because at that power level there is hardly an alternative to the diode - but it is now an instrument figure and no longer a process parameter. The process parameter is what arrives in the ceramic. The complete optical build-up including three months of source rental came to about 15 % of the order value of the process development it served.

All showcases

Get in touch

We are ready for new challenges

You bring the challenge. We develop the solution. From the first idea through concept and development to the validated machine, we guide your project.