Luminous Landscape Forum

Equipment & Techniques => Medium Format / Film / Digital Backs – and Large Sensor Photography => Topic started by: MichaelEzra on May 30, 2013, 10:58:17 pm

Title: Graphene Image Sensor 1000 Times More Sensitive to Light
Post by: MichaelEzra on May 30, 2013, 10:58:17 pm
Somehow these news did not yet appear in any photo websites.

“We have shown that it is now possible to create cheap, sensitive and flexible photo sensors from graphene alone,” said Wang  in a press release. “We expect our innovation will have great impact not only on the consumer imaging industry, but also in satellite imaging and communication industries, as well as the mid-infrared applications.”


http://spectrum.ieee.org/nanoclast/consumer-electronics/gadgets/graphene-image-sensor-1000-times-more-sensitive-to-light?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+IeeeSpectrumConsumer+%28IEEE+Spectrum%3A+Consumer+Electronics%29

I'd love to have this in a D800+ :)
Title: Re: Graphene Image Sensor 1000 Times More Sensitive to Light
Post by: Ben Rubinstein on May 31, 2013, 05:23:06 am
The fact it is bendable would make a big difference as well.
Title: Re: Graphene Image Sensor 1000 Times More Sensitive to Light
Post by: grzybu on May 31, 2013, 10:32:32 am
Well, if quantum efficiency  of current sensor is above 50% then how they could improve this 1000 times? Looks like they improved sensitivity 1000x compared to previous graphene sensors, not compared to current CMOS ones.
Title: Re: Graphene Image Sensor 1000 Times More Sensitive to Light
Post by: Jim Kasson on May 31, 2013, 10:43:32 am
Well, if quantum efficiency  of current sensor is above 50% then how they could improve this 1000 times? Looks like they improved sensitivity 1000x compared to previous graphene sensors, not compared to current CMOS ones.

From the IEEE Spectrum article (italics mine):

"The Singapore team, led by Assistant Professor Wang Qijie, from NTU’s School of Electrical & Electronic Engineering, has developed a graphene sensor that is 1000 times more sensitive to light than current imaging sensors used in today's digital cameras. "

Is it certainly possible to have quantum efficiencies greater than 100%, and such efficiencies have been reported in the literature. (http://www.sciencemag.org/content/340/6130/334.abstract) If, in some hypothetical sensor, on the average, one photon caused 200 electrons through some cascade effect, we'd have to say the quantum efficiency is 20,000%.  

I think it's an open question as to how useful QEs well above 100% are. I guess they could increase signal to noise ratios.

Jim
Title: Re: Graphene Image Sensor 1000 Times More Sensitive to Light
Post by: Jim Kasson on May 31, 2013, 11:08:25 am
From the article:

"But thanks to researchers at Singapore’s Nanyang Technological University (NTU), it may soon be possible to get the same low light sensitivity in a point-and-shoot camera that's now available only in today’s more expensive DSLR models."

One of the things I have to keep telling myself about Spectrum is that the people who write the articles are usually not engineers. I read the print version every month, and it gets to be pretty obvious that the people doing the writing (Bob Lucky, guest experts, and a few others excepted) are journalists first. That explains how the above quote got written.

Let's be generous and say that there's an eight to one ratio between the pixel pitch of a cheap P&S and a FF DSLR. That would translate to a 64 to one difference in light collecting ability, not 1000 to one.

But the big error is assuming that low light sensitivity (ie freedom from noise) can be improved dramatically by QEs in the 10000% to 100000% range. If we ever saw sensors like that, we'd still be limited by photon noise. In a sensor that produced 500 electrons for every photon, the standard deviation of a portion of an image with an average flux density-time of one photon would be 500 electrons, not sqrt(500) electrons.

Even with high QEs, you still need big sensels so you collect a lot of photons if you want low noise.

Jim
Title: Graphene Image Sensor [not] 1000 Times More Sensitive to Light
Post by: BJL on May 31, 2013, 02:27:07 pm
From the IEEE Spectrum article (italics mine):

"The Singapore team, led by Assistant Professor Wang Qijie, from NTU’s School of Electrical & Electronic Engineering, has developed a graphene sensor that is 1000 times more sensitive to light than current imaging sensors used in today's digital cameras. "

Is it certainly possible to have quantum efficiencies greater than 100% ...
As far as I can see, what matters is the fraction of photons that get detected, and this is already over 50%. "Reporting" each detection with an output of 100 electrons instead of one is just amplification, not an increase in detection efficiency. Such amplification can be useful up to a point, to raise the signal above dark noise, but small photosites have dark noise levels of only one or two electrons these days, so beyond that, extra "in-photosite amplification" helps very little.
Title: Re: Graphene Image Sensor [not] 1000 Times More Sensitive to Light
Post by: Jim Kasson on May 31, 2013, 03:21:15 pm
...what matters is the fraction of photons that get detected, and this is already over 50%. "Reporting" each detection with an output of 100 electrons instead of one is just amplification, not an increase in detection efficiency. Such amplification can be useful up to a point, to raise the signal above dark noise, but small photosites have dark noise levels of only one or two electrons these days, so beyond that, extra "in-photosite amplification" helps very little.

I'll buy that, but that's not how QE is defined. QE is the incident photon to converted electron ratio. Which, by the way, means that with a QE of 50%, less than half the photons could be detected, if the result of said detection is sometimes two or more electrons.

Jim
Title: New Graphene Sensor 1000 Times More Sensitive than previous _graphene_ sensors
Post by: BJL on May 31, 2013, 05:37:36 pm
I'll buy that, but that's not how QE is defined. QE is the incident photon to converted electron ratio. Which, by the way, means that with a QE of 50%, less than half the photons could be detected, if the result of said detection is sometimes two or more electrons.
Let us not get into a linguistic debate: the fact is that current CMOS sensors typically produce about one photo-electron per detected photon, and so are detecting more than half the photons, and so in practical terms, there is only room for less than a factor two of improvement in the fraction of photons detected, which is what matters for photon shot noise.

Anyway, the claim is simple misreporting: the actual claim in the paper is being 1000 times more efficient than previous graphene detectors, with even this improvement leaving graphene detectors far less efficient than CMOS sensors: http://www.nature.com/ncomms/journal/v4/n5/full/ncomms2830.html
Title: Re: New Graphene Sensor 1000 Times More Sensitive than previous _graphene_ sensors
Post by: Jim Kasson on May 31, 2013, 06:27:57 pm
Let us not get into a linguistic debate: the fact is that current CMOS sensors typically produce about one photo-electron per detected photon, and so are detecting more than half the photons, and so in practical terms, there is only room for less than a factor two of improvement in the fraction of photons detected, which is what matters for photon shot noise.

We are in violent agreement.

Anyway, the claim is simple misreporting: the actual claim in the paper is being 1000 times more efficient than previous graphene detectors, with even this improvement leaving graphene detectors far less efficient than CMOS sensors: http://www.nature.com/ncomms/journal/v4/n5/full/ncomms2830.html

Good catch. Spectrum strikes again.

Jim
Title: Re: Graphene Image Sensor 1000 Times More Sensitive to Light
Post by: thierrylegros396 on June 01, 2013, 03:42:20 am
As always, announce effects to catch investors ;) ;)

Don't trust all you read, especially today after financial crisis ::)
Title: Re: New Graphene Sensor 1000 Times More Sensitive than previous _graphene_ sensors
Post by: Wayne Fox on June 01, 2013, 04:12:13 am
the fact is that current CMOS sensors typically produce about one photo-electron per detected photon, and so are detecting more than half the photons, and so in practical terms, there is only room for less than a factor two of improvement in the fraction of photons detected,


I'm not a scientist, so I'm sure I'll be immediately corrected, from what I read with a CMOS sensor only around 10-30% of each sensel is actually light sensitive?  And I've read (somewhere) that even with current micro lenses, only about 20 to 30% of the available light reaches an actual light sensitive area?
Title: well over half of photons reaching modern sensors are detected
Post by: BJL on June 01, 2013, 09:43:21 am
... from what I read with a CMOS sensor only around 10-30% of each sensel is actually light sensitive?  And I've read (somewhere) that even with current micro lenses, only about 20 to 30% of the available light reaches an actual light sensitive area?
There is a confusion here about what that 20-30% means. The short answer is that microlenses and/or back-illumination overcome that low number.

It might be that in the very small photosites of small traditional front-illuminated sensors, only about 30% of the photosite area is unobstructed by the circuitry on top of the photosite, but
(a) modern gapless micro lenses gather light from almost all the photosite area and send it to that unobstructed part so that it can be detected, so the fraction of light that reaches than "30% window" can be (and is!) a lot more than 30% of incident light.
(b) The back-illuminated sensors now popular with small sensors avoid this problem by sending the light in from the other, unobstructed side of the chip.

The bottom line is that modern sensors (both CCD and CMOS) detect well over half of the photons that get through the color filters. In fact, even with color filters, where QE is measured as a fraction of light of all visible wavelengths that get through the color filters and are detected, the QE figures are around 30-40% ... and even higher in many compact camera sensors, which seems to use color filters that sacrifice some color discrimination accuracy for higher sensitivity.
Title: Re: Graphene Image Sensor 1000 Times More Sensitive to Light
Post by: Wayne Fox on June 03, 2013, 10:50:01 pm
Ok, thanks for that.  I new knew I would be corrected :)

(Spellcheck works wonderfully but unfortunately still doesn't handle grammatical spelling errors)
Title: Re: Graphene Image Sensor 1000 Times More Sensitive to Light
Post by: Wim van Velzen on June 04, 2013, 04:09:11 pm
You knew that you would be corrected.  ;D