Silicon-on-sapphire
-
- Posts: 38
- Joined: Wed Apr 12, 2023 3:08 pm
Silicon-on-sapphire
As I've gotten better at decapsulating chips, I began to find strange devices that look like little transparent cuboids with a printed circuit on one face. I don't have a macro image on hand to show this, only 2 epi-illuminated examples paired with their part codes:
Before looking up the company 'Peregrine" and finding that they specialize in Si-on-sapphire technology: https://www.psemi.com/newsroom/blogs/15 ... n-sapphire, I wondered if the substrate was glass or quartz (since the clocks used in electronics are quartz crystal oscillators) and decided to look conoscopically, since that would tell me if there was a crystal structure or not.
Here's the conoscopic view, one of the devices is held in between the Leica 100/1.47 HCX PL APO and a Zeiss 1.3 condenser (double immersion):
The brighter central disc occurs because the NA of the objective is larger than the NA of the condenser. Clearly, the material is a uniaxial birefringent material (so, not glass). I wondered if I could get quantitative information from this image based on the relationship between the radial coordinate 'r' on the back pupil plane and the angle of an object ray (with respect of the optical axis) 'u', r = sin(u).
Given the NA of the objective lens, I can determine the angle of the substrate crystal axis with the respect to the optical axis by measuring the distance from the center to the melatope; I calculated that angle to be 54 degrees. As it happens, the crystal structures of quartz and sapphire are the same. Checking a sapphire wafer manufactuer site, I learned that circuits are printed on the 'r-plane' (https://melleroptics.com/primary-planes ... e-crystal/) which is oriented at 57.6 degrees from the c- axis.
Science!
Before looking up the company 'Peregrine" and finding that they specialize in Si-on-sapphire technology: https://www.psemi.com/newsroom/blogs/15 ... n-sapphire, I wondered if the substrate was glass or quartz (since the clocks used in electronics are quartz crystal oscillators) and decided to look conoscopically, since that would tell me if there was a crystal structure or not.
Here's the conoscopic view, one of the devices is held in between the Leica 100/1.47 HCX PL APO and a Zeiss 1.3 condenser (double immersion):
The brighter central disc occurs because the NA of the objective is larger than the NA of the condenser. Clearly, the material is a uniaxial birefringent material (so, not glass). I wondered if I could get quantitative information from this image based on the relationship between the radial coordinate 'r' on the back pupil plane and the angle of an object ray (with respect of the optical axis) 'u', r = sin(u).
Given the NA of the objective lens, I can determine the angle of the substrate crystal axis with the respect to the optical axis by measuring the distance from the center to the melatope; I calculated that angle to be 54 degrees. As it happens, the crystal structures of quartz and sapphire are the same. Checking a sapphire wafer manufactuer site, I learned that circuits are printed on the 'r-plane' (https://melleroptics.com/primary-planes ... e-crystal/) which is oriented at 57.6 degrees from the c- axis.
Science!
-
- Posts: 4
- Joined: Thu Jun 29, 2023 8:26 am
Re: Silicon-on-sapphire
Really fascinating.
I have never seen a Silicon on Saphire integrated circuit before. Great detective work investigating the substrate.
I have never seen a Silicon on Saphire integrated circuit before. Great detective work investigating the substrate.
Located in Australia.
Interested in a TEM or related equipment such as coating equipment, grids, ultramicrotome accessories etc. Please send me a PM if you can assist.
Interested in a TEM or related equipment such as coating equipment, grids, ultramicrotome accessories etc. Please send me a PM if you can assist.
-
- Posts: 399
- Joined: Mon May 16, 2022 3:44 pm
Re: Silicon-on-sapphire
Where did you find exposed substrate?
-
- Posts: 2787
- Joined: Sat Mar 03, 2018 9:09 pm
Re: Silicon-on-sapphire
Because quartz is chiral it gives a unique 'bullseye' pattern in the conoscope which is diagnostic.
-
- Posts: 38
- Joined: Wed Apr 12, 2023 3:08 pm
Re: Silicon-on-sapphire
If I understand you, I have other ones that aren't in such good condition- most of the circuitry 'washed' off- and I used one of those.
-
- Posts: 542
- Joined: Mon Sep 26, 2022 3:13 am
Re: Silicon-on-sapphire
Thanks for this fantastic post.Andy Resnick wrote: ↑Sun May 21, 2023 7:22 pmAs I've gotten better at decapsulating chips, I began to find strange devices that look like little transparent cuboids with a printed circuit on one face. I don't have a macro image on hand to show this, only 2 epi-illuminated examples paired with their part codes:
-
- Posts: 399
- Joined: Mon May 16, 2022 3:44 pm
Re: Silicon-on-sapphire
This is all pretty cool. Plus, doing doubly immersed optical minerology with 100x objectives is kind of a PITA.
So, is the substrate quartz on sapphire or just sapphire? And what's the purpose of printing an IC on sapphire vs. say, GaS., or basic Si?
I suspect that, based on the (barely visible) 4th order interference colors, if it is indeed sapphire, that it is about 250µm thick.
Science indeed!
So, is the substrate quartz on sapphire or just sapphire? And what's the purpose of printing an IC on sapphire vs. say, GaS., or basic Si?
I suspect that, based on the (barely visible) 4th order interference colors, if it is indeed sapphire, that it is about 250µm thick.
Science indeed!
-
- Posts: 38
- Joined: Wed Apr 12, 2023 3:08 pm
Re: Silicon-on-sapphire
The substrate is just sapphire. Why sapphire instead of other materials? I think it has to do with the superior insulation property when designing RF circuits (no stray induced currents, for example), but I don't have a definitive answer for that.Sure Squintsalot wrote: ↑Wed Aug 30, 2023 11:47 pmThis is all pretty cool. Plus, doing doubly immersed optical minerology with 100x objectives is kind of a PITA.
So, is the substrate quartz on sapphire or just sapphire? And what's the purpose of printing an IC on sapphire vs. say, GaS., or basic Si?
I suspect that, based on the (barely visible) 4th order interference colors, if it is indeed sapphire, that it is about 250µm thick.
Science indeed!
Good catch regarding thickness- I should check!