Measuring charge transfer efficiency in your CCD

[narrowbandpaul]

Hi Guys/gals.

I may have come up with a neat way to measure how efficiently charge is transferred between pixels.

In the beginning CCD size was limited to a few hundred pixels on a side because they had poor charge transfer efficiency around 0.98. ie 2% of electrons lost per transfer.

Now there is CTE better than 99.999%, which allows for massive arrays.

This short word doc i wrote explains a way to measure your CTE. It is fairly straightforward. i think.

I hope the word doc is easy to read and follow, but if you have questions, just ask.

I am not sure how accurate this method is, so I have emailed a guy in San Fran who knows abouth the inner workings of CCD's, incase he has some info.

If you try it, let me know how you get on.

Or just read it for fun.

The normal method for accurate CTE measurement is 5.8keV Xrays from Iron. But there are issues...firstly its not common in every household to have an Xray source. Secondly the CCD coverslip strongly absorbs this radiation, and none may actually go through to the silicon chip.

The method here just uses standard imaging techniques, so comparitavely more straightforward than ionising radiation!

Thankyou

Paul

link is below :-)

[Sam]

Paul,

Why would you do this? Is there something you can do to improve this efficiency or does it help with knowing which targets are best to go for with a particular CCD? just wondering why it's done :scratch:

Sam

[narrowbandpaul]

its one of those, because we can deals.

Its science, not really any practical use.

CCD designers can change the voltages they use to improve CTE, if they are not happy.

We the consumer cant do anything about it.

It came about when I was thinking about mine and ally's (ngc2403) astronomy project on high accuracy photometry. I was thinking we could correct for the electron loss when shifting the charge across the pixels.

Mainly a pure science thing. I find CCD's interesting, so this appeals to me.

But its not got a huge practical implications, sorry.

unless you are doing photometry of course :-)

[Sam]

oh ok, thanks for the explanation

Sam

[narrowbandpaul]

youre welcome

[dph1nm]

Hmm, as I understand it, CTE is mainly affects low intensity sources. Stars with a high number of counts should show very low CTE. Also, if you have a significant sky (or dark current) background, CTE is almost eliminated entirely. This is why, in the early 1980s, we used to preflash CCDs with a light source before exposing on the sky.

The theory seems to be that CTE is caused by 'traps' in the silicon which electrons can get stuck in as you try to shuffle them off the chip. Once these traps are full, there is no more CTE.

NigelM

[narrowbandpaul]

preflashing nowadays is used to flood RBI traps left at the interface between epitaxial and substrate silicon. The NIR flood fills the traps so no more electrons can be trapped.

i suppose this happened in the past when they still used surface channels for electron storage. ie fill em up so they dont trap anymore.

CTE is certainly limited by traps, but using very high purity silicon could reduce this.

but traps have a lower than average CTE. I believe you will always have a lower than perfect CTE regardless of whether the chip has no traps. Especially since the voltages and clock rise and fall times can be tuned to maximise CTE. You will always lose an electron or two when shifting.

Nothing in physics is perfect.

paul