[time-nuts] Etching of quartz crystals

[Bob kb8tq]

Hi

> On Feb 5, 2018, at 10:31 AM, jimlux <jimlux at earthlink.net> wrote:
> 
> On 2/5/18 5:54 AM, Attila Kinali wrote:
>> On Sun, 4 Feb 2018 09:21:54 -0500
>> Bob kb8tq <kb8tq at n1k.org> wrote:
>>>> The images on this page gives a good impression about the current
>>>> skill-level in that area:
>>>> 
>>>> 	https://www.azonano.com/article.aspx?ArticleID=2740 <https://www.azonano.com/article.aspx?ArticleID=2740>
>>> 
>>> The gotcha is shown in the pictures. First point is that they are etching *very*
>>> small features. A 5 MHz 3rd overtone blank is way thicker than what they are
>>> playing with. The second issue is that even at small scale the walls are going
>>> non-parallel.
>> That's exactly the issue here. While SAW resonators benefit quite a lot
>> from the processing skills learned from semiconductor fabrication, these
>> skills do not translate into BAW manufacturing. SAW resonators are built
>> etching or depositing small features ontop of a SiO2 wafer that is supposed
>> to be as flat as possible. On the other hand BAW oscillators are 3D structures
>> by themselves. They are lens shaped (thus not flat) to keep the oscillation
>> energy trapped in the center of the slap, thus allowing the edges to be used
>> for mounting/contacting, with minimal damping of the oscillation.
>> Yes, the shapes are simple. But not only because that's the only shapes
>> we know how to build, but also because these shapes allow us to calculate
>> how the crystal will oscialate and because the simpler the structure the
>> easier it is to build it with high precision and accuracy.
>> It would be possible to use edging of surface structures into the
>> crystal to form a Bragg reflector (instead of the lense shape).
>> But I have no idea how well it works. Considering that it is easier
>> to build a slap that is flat and then etching structures on it, than
>> to form a 3D structure, I wonder why I have not read about anyone
>> doing exactly that (beside for SAW structures).
> 
> 
> Follow the money - or lack thereof - Folks are happy with the existing technology - If I'm flying a science mission that needs a space qualified Ultra Stable Oscillator - I've already budgeted my several million dollars, claiming that I'll just use what we already know how to build, and I spend no more proposal pages talking about it.  I certainly am not going to say "instead of spending $1M/oscillator for my 2 oscillators, I'm going to spend $5M on an experimental process to change how the resonator is made, and by the way, it might not work"
> 
> Would using ion milling and other modern fabrication techniques lead to an oscillator with *significantly* better performance or *significantly* lower cost?

Consider that a reasonable budget for just the gear to finish a precision resonator (not the fab side) is 
in the $3 to $5 million range these days. If you include the fab process, the question becomes “how 
far back” you go in that process. Do you start with growing the synthetic quartz? If so, the budget just
goot a *lot* bigger. 

People have indeed experimented a lot with alternative fab processes. There are alternatives out there.
So far, the precision end of things still *looks* a lot like it used to. If you scratch under the surface it’s
vastly different than it was 10 or 20 years ago. It’s nothing at all like it was in the 70’s, let alone back 
in the 50’. 

Bob


> 
> For those users for whom this is important, research focuses on looking for another qualitatively different way to get there - That's sort of what the CSAC and the Deep Space Atomic Clock (DSAC) are about - the ion trap clock for DSAC gives you long term performance BETTER than a USO. Although probably not at a lower cost, yet, there is potential for it to be so.
> The CSAC gives you "good accuracy at low power", compared to an OCXO - less than 1/10th the power.
> 
> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.