Skyphos was forged from necessity. In 2013, Elliot was a graduate student and on a PhD track, he jumped out of Materials Science to take a few courses in biomedical. What he found was surprising to an engineer with 15 years experience, he was floored:
A prototyping and fabrication process that took months to transform an idea to reality, and only slightly shortened with the right equipment
After prototyping and prove-out, the process would take 4-6 months more to make the requisite molds for high throughput fabrication.
All this industry was based on a 30 year old process re-appropriated from making printed circuit boards
No one ever looked at it and said - maybe there is a better option
Any design not based on a geometric shape like a circle or square can’t be made
Device size, aspect ratios and limits based on materials - Why aren’t we looking for new materials or developing them?
We can’t make devices with enough throughput for certain processes - so solutions for patients never get to help them
So he did what any graduate student would do, he wrote a paper on it and solved it..theoretically. Then one year later after leaving the PhD track, starting a new company - he was faced with the exact same problem but it wasn’t theoretical anymore. So he created Skyphos, a company dedicated to solving the bottle-neck.
To fabricate 3D models, existing in both the micro and macro worlds,
Skyphos needed to accomplish three critical metrics:
Reducing pixel size &
maintain build area:
A word about resolution, most 3D printer companies define their resolution as the smallest pixel defined on the build platform. In the case of DLP based printers it is incorrectly stated as a single pixel pitch or size. It is an error to claim this because in all cases we are aware of, it requires 2 pixels sharing a side to create a solid, and 3 pixel spacing to accurately create a void.
Reviewing LCD and DLP displays, the minimal pixel aspects (XY) are 50 um, meaning the smallest solid would be 100 um, the smallest void 150 um. In the SLA category (Form Labs) the smallest Gaussian full beam half width is around 75 um, meaning a 150 um solid and 225 um void. Both are limited in application to the upper end of microfluidic regime.
The Skyphos pixel aspect range is single microns and up, meaning we can create features below 10 um and voids or channels down to 15 um.
& high precision resins:
One of the main hurdles for 3D printing has been the availability of photoinitiators which are bio-compatible; most are cytotoxic.
A second hurdle on this front was finding one which can create optically clear and transparent devices while maintaining a low auto-fluorescence so that cell tagging may be achieved. In addition, microfluidics requires surface roughnesses below 1 um.
For 3D printing this creates a challenge in that each layer printed has some artifacts which degrade this ability. With the Gillespie Algorithms (TM), these are all but eliminated resulting in, on average Skyphos achieves below 200 nm surface roughness for all exposed areas.
Computer processing power
& slicing engines:
The final hurdle has been the implementation of a slicing engine which can operate at minimal feature sizes, while maintaining proper data capture and eliminates computer crashing.
There were two steps: STLs and large arrays:
STLs: loose data, and at the sizes microfluidics requires for operation - the slicing engines available simply were not built to provide proper transfer. We developed our own.
Large arrays of pillars - on the order of 15k or more begin to cause issues arising from the processing power or computers and the method STLs use to create layers. We created the program which allows transfer from large 2D array designs to complete 3D renders.
If you can’t process your design, call us we have been there, know the pain and are ready to help