durusmail: mems-talk: Microlens fabrication
Microlens fabrication
2009-11-04
2009-11-06
2009-11-05
2009-11-09
2009-11-09
2009-11-10
2009-11-11
2009-11-12
2009-11-13
2009-11-12
Microlens fabrication
Dean Hopkins
2009-11-12
Hi Alasdair,

I haven't deliberately reflowed bulk glass onto silicon wafers.

I know from bitter experience that it can be done, and that the resulting
surface retains most of the original surface features.

Once the process is tuned in (I'd start with long times at temperatures only
slightly above the softening point) it should be repeatable.

BUT as Josh pointed out, at some point if the features are small enough, surface
tension will overwhelm any topography.

Perhaps something more like depositing a thin layer of BPSG before applying the
glass beads, and only heating enough to reflow the BPSG, but not melt or soften
the beads would work? This might even work better as the lenses remain more-or-
less optically isolated for less crosstalk. Probably lower surface stresses as
well, since the lens array isn't a solid surface, and thermal expansion is
broken up into short segments rather than a single long path...

Hope this helps,

Dean

Dean A. Hopkins, Jr.
MEMS Process Engineer
Silicon Valley, USA
(408) 429-0501

-----Original Message-----

Reflow method can be very helpful for creating microlenses. See "Variable
Focusing Microlens Chip for Potential Sensing Applications" from IEEE
Sensors Journal VOL. 7, NO. 1, JANUARY 2007. The array has been made. The
only problem I see is how small lens it can create. Less than 100 um
probably is a problem.

UV curable adhesive is more difficult to handle than photoresist.

I am glad to see more application from it.

Cheers,
Josh



-----Original Message-----
From: Alasdair Rankin [mailto:arankin@groupivsemi.com]
Sent: Monday, November 09, 2009 6:08 AM
To: 'General MEMS discussion'
 Subject: Re: [mems-talk] Microlens array

Hi Dean,

I appreciate the feedback.  That's an interesting approach that I hadn't thought
of.  I was planning to align to a pattern underneath.  It sounds like what
you're proposing would be a more uniform array of lenses, if I am understanding
you.

The surface already does have a topography to it, and this would probably be an
ideal approach for a proof of concept.  Do you think this could be repeatable?
Any experience with the suggested chemical?

Thanks guys for your feedback,

Alasdair

reply