Test of Sony A7s for near infrared spectroscopy

Design, construction, tuning of spectroscopes
Information and discussion about softwares (telescope remote, autoguiding, acquisition, spectral processing ...)
Christian Buil
Posts: 1431
Joined: Mon Sep 26, 2011 6:59 pm
Contact:

Test of Sony A7s for near infrared spectroscopy

Post by Christian Buil »

Evaluation of a full spectrum conversion Sony A7s digital camera for near infrared spectrography:

http://www.astrosurf.com/buil/fringes/index.html

Fringing is a serious problem with this setup...! Comments and experiences are welcome.

Christian
Peter Somogyi
Posts: 420
Joined: Sun Jul 13, 2014 8:56 am

Re: Test of Sony A7s for near infrared spectroscopy

Post by Peter Somogyi »

Hello Christian,

I'd recommend this reading:
http://stsdas.stsci.edu/cgi-bin/gethelp.cgi?defringe
Seems professionals also met with issues > 700nm. But I don't understand it deep enough (just able to execute it in IRAF).

As for me (414 EXm) in the optical, I have these fringes:
1) high frequency ripples ~ 2A and 2-3% varying with wavelength (very well reproducing - even 1 year ago the internal flat matching by wavelength! - but, to me it requires a transform on an external flat: shrink + shift it - to match on a reference hot star; no idea why the shrink necessary - still playing to see reproducability).
2) low frequency ripples (roughly ~50-100A) like yours, they seem to be removed well by regular flat, very well reproducible
3) concentric circles like this, does not move with wavelength: https://en.wikipedia.org/wiki/Fabry%E2% ... rferometer (seems better not to remove from the flat, it is correcting real)

I'm not clearly sure about your statement above the animated gif: "It is the crucial point. Our major problem is the temporal instability of the fringe pattern. " Do I understand it right: the fringe changes by time so rapidly? Maybe related to temperature perhaps, or sky position? Never heard such.

Peter
Christian Buil
Posts: 1431
Joined: Mon Sep 26, 2011 6:59 pm
Contact:

Re: Test of Sony A7s for near infrared spectroscopy

Post by Christian Buil »

Peter, I suspect temperature variation. My interrogation is about the big consequence
of a necessary low temperature variation in a short time interval (but the temperature
of a photographic camera detector and associated glass plate are not regulated..., perhaps the reason ?).

Christian
Martin Dubs
Posts: 141
Joined: Fri Jan 06, 2012 9:16 pm
Location: Maienfeld, Switzerland

Re: Test of Sony A7s for near infrared spectroscopy

Post by Martin Dubs »

Hi Christian,

the wavelength dependence of the fringe amplitude probably comes from AR coated surfaces. A typical AR coating has a low reflectance range of 350 to 650 nm for a color camera with steeply rising reflectivity outside the nominal range, similar to the curves taken from a manufacturer (Thorlabs):
AR Coating350-700nm.jpg
AR Coating350-700nm.jpg (29.49 KiB) Viewed 9048 times
The better the AR coating, the steeper the edges of the passband! The rapidly varying fringes point to an air spaced etalon, maybe between top surface of sensor and cover glass. It may help to replace the cover glass with one with a special AR coating for the infrared, but the problem may also be in the CCD itself, which would be much harder to correct.

Regards, Martin
Christian Buil
Posts: 1431
Joined: Mon Sep 26, 2011 6:59 pm
Contact:

Re: Test of Sony A7s for near infrared spectroscopy

Post by Christian Buil »

I agree Martin, complex AR coating used outside its optimization domain can turn
into a mirror. But for me the biggest mystery is the origin of equal inclination
fringes in a thickness of 50 microns only. Too thin for correspond to a cover glass
or a filter, too wide for interference into layers of a dielectric AR-coating it seems.

Christian
Martin Dubs
Posts: 141
Joined: Fri Jan 06, 2012 9:16 pm
Location: Maienfeld, Switzerland

Re: Test of Sony A7s for near infrared spectroscopy

Post by Martin Dubs »

Hello Christian,

a possible explanation for the high frequency fringes may be the back illuminated sensor itself. At around 800 nm the thin Si substrate becomes rapidly transparent, leading to interference between back (illuminated) and front (electric contacts) surface. With a refractive index of about 3.6 and a 15 micron Si substrate thickness the fringes can be easily explained. As light above 700 nm is normally blocked by filters the manufacturer probably did not bother about these internal reflections. Actually the detector sensitivity in the IR is increased with a reflective front surface, by preventing photons from escaping. The following diagram shows T, R and A of a 15 micron Si substrate, uncoated both sides
15 mu Si wafer. T: red, R: green, A: blue
15 mu Si wafer. T: red, R: green, A: blue
Si Substrat CMOS15.jpg (95.54 KiB) Viewed 9002 times
Clearly visible is the loss of absorption above 750 nm and the onset of fringes. I am afraid there is no cure against that.

Regards, Martin
Robin Leadbeater
Posts: 1924
Joined: Mon Sep 26, 2011 4:41 pm
Contact:

Re: Test of Sony A7s for near infrared spectroscopy

Post by Robin Leadbeater »

Yes the ripples in IR due to transparency of the Si substrate are well known to the professionals (if this is the cause here). Less well know is that similar problems are also possible in the visible region, not just from the full thickness of the substrate but even from the different layers in the sensor. See section 3 and 5 of this paper
http://www.aphesa.com/downloads/download2.php?id=1

The kodak KAF CCD show this problem severely but they are seen at a lower level in Sony CCD like the ICX694. I have also seen them from CMOS sensors used with the Star Analyser where a flat correction is not possible.

Robin
Last edited by Robin Leadbeater on Tue Oct 25, 2016 4:47 pm, edited 1 time in total.
LHIRES III #29 ATIK314 ALPY 600/200 ATIK428 Star Analyser 100/200 C11 EQ6
http://www.threehillsobservatory.co.uk
Christian Buil
Posts: 1431
Joined: Mon Sep 26, 2011 6:59 pm
Contact:

Re: Test of Sony A7s for near infrared spectroscopy

Post by Christian Buil »

Hello Martin,

Your interpretation is very attractive. It may be reinforced
by the presence of the same fringes in the image of H-alpha Sony at the band limit
(but low intensity here), so the problem could be concentrated in the detector.

I have permission to use your graph for illustration?

"Newton rings" aspect is also visible, may be related to the average rays angle of incidence,
variable along the spectral field?

Regards,

Christian
Martin Dubs
Posts: 141
Joined: Fri Jan 06, 2012 9:16 pm
Location: Maienfeld, Switzerland

Re: Test of Sony A7s for near infrared spectroscopy

Post by Martin Dubs »

Christian,

you may use my graph as you like. I created it with my thin film program with optical constants n and k similar to data from Palik, Handbook of optical constants for pure Si.
I am not sure about the Newton rings. I thought the EXMOR sensors are covered with small lenses such as shown here:
EXMOR pixel cross-section
EXMOR pixel cross-section
Pixel-cross-section-Back-Illuminated.jpg (24.28 KiB) Viewed 8985 times
For a thin layer the accceptance angle of the center fringe is fairly large, so this should still produce fringes of lower intensity. Just to make sure, with your camera, was the color filter and these lenses removed? I assume they were, since you say total defiltered.

Regards, Martin
Christian Buil
Posts: 1431
Joined: Mon Sep 26, 2011 6:59 pm
Contact:

Re: Test of Sony A7s for near infrared spectroscopy

Post by Christian Buil »

Martin,

Only the native IR-cut filter is removed and replaced by a clear glass of same thickness. The microlens and Bayer filters are present (very difficult to remove).

Christian
Post Reply