How do they generate one (only) photon?

[Vorpal Blade]

In a Roger Penrose book I’m reading (don’t remember the title, he’s arguing against algorithmic AI) he mentions that some experiments in quantum mechanics involve a single (one, count ‘em, one) photon. In the same book he mentions, by way of contrast, that a 60 watt light bulb generates about 10^20 photons per second (!).

My question is: HTF do they generate a single photon?

[CalMeacham]

You don't usually generate a single photon -- you generally generate lots of photons and then use lots of filters to make certain that the photon flux is exceedingly low, amounting to only one photon at a time over long periods of time. That., I understand, is how they did those experiments such as Young's double-slit experiment with only one photon at a time. The result, if you held to the particle rather than the wave model, shows that, even when only a single photon at a time encounters the slits, you still get the same photon statistics and diffraction pattern as when lots of photons are there (over time, of course). So the photon doesn't need to "interfere with" another photon as it passes through the slit. *


Recently there have been other improvements, such as Mike Feld's "single atom laser", which must, of necessity, emit only a single photon at a time. I'm not up on exactly how that'sa built and prepared, but any time you have an isolated excited atom, you'll only get a single photon out. At an unpredictable time.




*The way the Double Slit experiment is described is pretty misleading. People make a big deal about the photon going through one slit or the other, as if that's the big mystery. It isn't. You'll get interference effects with three or more slits. Heck, you'll get a diffraction pattern using a single slit, but which changes its shape as the slit width changes. In that case you know which slit the photon has gone through. The question then might be how the photon "knows" how wide the slit is, when photons are much smaller than the slit. Asking which slit the photon goes through -- or even how the photon senses slit width -- is a pretty crude way of asking how you can reconcile a localized particle model with non-localized wavelike behavior.

[Smeghead]

Or just flip the light switch on and off really fast.

[Chronos]

You can have a room bright enough to read in, which still contains only one photon at a time. Yes, many are created each second, but at the speed of light, it also doesn't take long to hit a wall.

[aerodave]

Chronos, do you disagree, then, with the nomber quoted above that a 60W light bulb generates on the order of 10²⁰ photons per second? Because if that's true, then a light bright enough to read by (I'll say one to two orders of magnitude less than that 60W bulb) would generate more than 10¹⁸ photons per second. And so each photon would have only 10^-18 seconds to get out of the way before the next one was generated. that's only enough time to travel a billionth of a foot. Pretty small room. But again, that's based on the 10²⁰ number being accurate.

Regardless, you just blew my mind with that post.

[yorick73]

Originally Posted by Chronos

You can have a room bright enough to read in, which still contains only one photon at a time. Yes, many are created each second, but at the speed of light, it also doesn't take long to hit a wall.

I'd call BS on this if I knew enough about physics to argue Can you please explain this statement.

[Sunspace]

Originally Posted by aerodave

Chronos, do you disagree, then, with the nomber quoted above that a 60W light bulb generates on the order of 10²⁰ photons per second? Because if that's true, then a light bright enough to read by (I'll say one to two orders of magnitude less than that 60W bulb) would generate more than 10¹⁸ photons per second. And so each photon would have only 10^-18 seconds to get out of the way before the next one was generated. that's only enough time to travel a billionth of a foot. Pretty small room. But again, that's based on the 10²⁰ number being accurate.

Regardless, you just blew my mind with that post.

It would have to be a really small room?

[GargoyleWB]

"Bright enough to read in" could be more than several orders of magnitude less than a light bulb. With a dark-adapted eye, you could conceivably still read with a children's room glow-in-the-dark sticker held right above the page--not nearly as many photons there. That might scale you past your room volume to achieve only one-photon in the room at a particular instant.

Haven't run the numbers, but a fun speculative post from Chronos that I'd never thought of

[beowulff]

Originally Posted by yorick73

I'd call BS on this if I knew enough about physics to argue Can you please explain this statement.

I'll second that.
If the distance from the light source to the wall is 10', then it takes a photon 10.018 nanoseconds to reach the wall. Therefore, you can only have 98.208 million photons per second, which is probably not enough to read by (that's 12 orders of magnitude fewer photons than the 60W bulb generates per second).

[Ring]

Originally Posted by beowulff

I'll second that.
If the distance from the light source to the wall is 10', then it takes a photon 10.018 nanoseconds to reach the wall. Therefore, you can only have 98.208 million photons per second, which is probably not enough to read by (that's 12 orders of magnitude fewer photons than the 60W bulb generates per second).

The 60 watt bulb emits photons in practically all directions, whereas your 98.2 million photons are the number emitted in only one single direction. At least I think that's how you calculated it, I haven't bothered to check your math.

[beowulff]

Originally Posted by Ring

The 60 watt bulb emits photons in practically all directions, whereas your 98.2 million photons are the number emitted in only one single direction. At least I think that's how you calculated it, I haven't bothered to check your math.

No, I just calculated it as if the bulb was suspended in the middle of a 10' sphere, and only one photon is allowed to be in flight at any time.

[Ring]

Originally Posted by beowulff

No, I just calculated it as if the bulb was suspended in the middle of a 10' sphere, and only one photon is allowed to be in flight at any time.

Sorry. I also missed that Chronos said there would only be one photon in the entire room at a time. I hope he reenters the scene. He's very rarely wrong.

[Chronos]

OK, running the numbers, it turns out I was dead wrong. For a flight path of 3 meters, and blue light, one photon at a time would mean about 5e-11 Watts.

Dangit, I know I saw a statistic like that somewhere, but obviously, I horribly mangled it. What was it I was thinking of, then?

[Ring]

Ok, as I said above, you just can't trust Chronos he's very rarely right.

[Vorpal Blade]

OK, this has been fun (Really. That's why I ask questions on this board; I get much more than asked for) but does anyone have any answer to the OP? Is the super heavy filter the answer? Seems a rather hit-or-miss approach.

[CalMeacham]

Seems a rather hit-or-miss approach.

But it's far and away the easiest, and perfectly adequate for many of the experiments they used to do.

With the advent of quantum cryptography and other phenomena, though, people are looking for something better. Check out this recent article on the topic in Laser Focus World:

http://www.laserfocusworld.com/articles/257230

It is possible to simulate a single-photon source by strongly attenuating light pulses that initially contain many photons so only a single photon remains. That approach is widely used in quantum cryptography experiments, but is inherently very inefficient. Research is now focused on developing true single-photon sources that can do better.

(Bolding mine)

[JacobSwan]

Originally Posted by bikebloke

OK, this has been fun (Really. That's why I ask questions on this board; I get much more than asked for) but does anyone have any answer to the OP? Is the super heavy filter the answer? Seems a rather hit-or-miss approach.

For my undergraduate project, about 15 years ago, I did the double split experiment with a CCD camera. CCD was quite new and exciting then and this expensive camera with PC capture card had been purchased for the observatory but ended up in my hands.
I captured hundreds of images of a single photon and then added them together one at a time to produce a series of still images, each one having one more dot than the last. These were then fed out to VHS (remember that?) with each image lasting half a second. The final result was a 3 minute video showing the diffraction pattern appearing one photon at a time.
These days this is probably a standard teaching method to first years, but at the time I believe I was the first to use this technology in this way. Rock and roll!

Anyway, I can confirm that the laser beam was reduced to a single photon at a time by the use of a bucket load of filters.

[billfish678]

I've always wondered if you could make a holgram this way, one photon at a time.

One photon is always coherent with itself right?

Of course each photon would still have to be in a very narrow range of wavelengths as well.

[Jamicat]

What about a 10'x10'x10', or 10' Dia. Sphere, Hermetically Sealed, Mirrored room...and a single photon?


At the speed of light, wouldn't it be pretty much everywhere all at once?

How many Photons does it take to make Light visible at normal 20/20 vision?

(the visible spectrum)

[The Hamster King]

Originally Posted by Jamicat

How many Photons does it take to make Light visible at normal 20/20 vision?

I'm not quite sure if this is what you're asking, but in a darkened room the human eye can detect a single photon.

[Jamicat]

Originally Posted by The Hamster King

I'm not quite sure if this is what you're asking, but in a darkened room the human eye can detect a single photon.

But, would One Photon Illuminate an Object enough in said room?

Don't objects absorb and release photons, in their respective wave lengths?

[Jamicat]

Add in the equation One bar of gold and another of silver.

Would one photon be able to decipher one from the other, by color band?

Is a photon a carrier particle that can morph at will, depending on atoms it encounters, then changes wavelength the same way?

Stop making me think dammit...

[CalMeacham]

I've always wondered if you could make a holgram this way, one photon at a time.

One photon is always coherent with itself right?

Of course each photon would still have to be in a very narrow range of wavelengths as well.

A very interesting question. When Gaboir invented holography the light source with the longest coherent length was a highly-filtered mercury arc, and even then the coherence length was , IIRC, sub-millimeter. Holograms were boring in those days, and for the next 14 years until Leith and Upatnieks used a long coherence length HeNe laser to make a hologram that looked 3D. Could you get the same effect by using single photons and really long exposure times, without all that troublesome filtering?


I don't think so. Let's say you're trying to make a diffraction pattern of a double slit hologram like JacobSwan describes, only using a cheap mercury source. The problem is that the pattern has a wavelength dependence -- the shorter the wavelength, the smaller the pattern. Coherence is related to the range of wavelengths you've got, and an unfiltered mercury source is still going to have a broader range of wavelengths than a laser. It's true that each photon interferes with itself, but photons of different wavelength will try to make slightly different sized patterns. These are going to average out to a blurry pattern in the long run, rather than a sharp pattern made by a very narrow range of wavelengths. The net result of coherent interferences of different wavelength photons won't be the same as large-scale interference of multiple single-wavelength photons.


Sorry. Nice idea, though. Worth an undergraduate experiment, like JS's.

[billfish678]

Originally Posted by CalMeacham

I don't think so. Let's say you're trying to make a diffraction pattern of a double slit hologram like JacobSwan describes, only using a cheap mercury source. The problem is that the pattern has a wavelength dependence -- the shorter the wavelength, the smaller the pattern. Coherence is related to the range of wavelengths you've got, and an unfiltered mercury source is still going to have a broader range of wavelengths than a laser. It's true that each photon interferes with itself, but photons of different wavelength will try to make slightly different sized patterns. These are going to average out to a blurry pattern in the long run, rather than a sharp pattern made by a very narrow range of wavelengths. The net result of coherent interferences of different wavelength photons won't be the same as large-scale interference of multiple single-wavelength photons.

Thats why I said NARROW

If it was monochromatic would it work ?

[CalMeacham]

Thats why I said NARROW

If it was monochromatic would it work ?

Ain't no such thing as monochromatic -- there's only different degrees of narrow.


Yeah, it'd work. But if I'd narrowed by bandwidth down enough to the point where this would work, why bother making my hologram out of single photons? It'd go a lot faster with as many as I could hit the film/dichromated gelatine/whatever with.

another note:

even though your photon will always be correlated with itself, any source with a fjnite bandwidth will have a coherence length determined by that bandwidth. And you can't make a hologram of an object bigger than your coherence length. again, each individual photon will giv you a good data point, but your collection of points from the range of photons will be decorrelated after you exceed the coherence length. So there's no point in a hologram built out of single photon exposures, as far as I can see.

[billfish678]

Originally Posted by CalMeacham

Ain't no such thing as monochromatic -- there's only different degrees of narrow.


Yeah, it'd work. But if I'd narrowed by bandwidth down enough to the point where this would work, why bother making my hologram out of single photons? It'd go a lot faster with as many as I could hit the film/dichromated gelatine/whatever with.

another note:

even though your photon will always be correlated with itself, any source with a fjnite bandwidth will have a coherence length determined by that bandwidth. And you can't make a hologram of an object bigger than your coherence length. again, each individual photon will giv you a good data point, but your collection of points from the range of photons will be decorrelated after you exceed the coherence length. So there's no point in a hologram built out of single photon exposures, as far as I can see.

Yes..yes..yes

Why do it? First, cause its hard to do ?

Second, maybe you would like to make a nice 3D reflection hologram yet you dont have a laser. Is there a way to get a very narrow/nearly monocromatic light source that isnt a laser? If so, whats the coherence length? You are implying that the coherence length would HAVE to be extremely short. Not that you want it short, that these methods would result in short kengths.

[CalMeacham]

As I say, Gabor in 1948 used a highly-filtered mercury arc with a single line because that was the highest intensity narrow-band source available to him. You can make narrower sources with higher output out of mercury emission lines -- by building them into lasers. I'm not sure if any of the new technologies (like quantum dots) would give you a better bandwidth without making a laser

[yorick73]

Originally Posted by Chronos

OK, running the numbers, it turns out I was dead wrong. For a flight path of 3 meters, and blue light, one photon at a time would mean about 5e-11 Watts.

Dangit, I know I saw a statistic like that somewhere, but obviously, I horribly mangled it. What was it I was thinking of, then?

Okay slight hijack. I still don't get it. Regardless of the Watts, one photon in a room at one time would have to hit something and then hit your eye. There doesn't seem to be any way that I can imagine that one photon in a room at one time could carry enough information to see anything except one photon. Am I missing something fundamental?

[Chronos]

You wouldn't see an image from a single photon, just (roughly) a single pixel. But if you get one photon, and then another, and then another, one after another really quickly, you could see an image from that.

[yorick73]

Originally Posted by Chronos

You wouldn't see an image from a single photon, just (roughly) a single pixel. But if you get one photon, and then another, and then another, one after another really quickly, you could see an image from that.

Okay, that makes sense. But you would have to assume that the photons are hitting enough areas on a page to read it. Or does the speed of the photons take care of this problem? (i.e. there are so many in such a short amount of time that one can expect the entire page to be completely covered very quickly)

[Yeti08]

Originally Posted by The Hamster King

I'm not quite sure if this is what you're asking, but in a darkened room the human eye can detect a single photon.

I know that Wikipedia says that a rod cell can be sensitive to a single photon, with no citation provided, but I have a hard time believing that. For instance, even humans emit visible light up to 100 photons per cm2 (ultra-weak light emission/human biophoton emission) and I've never seen a glowing person before even in the darkest of settings.

Originally Posted by Jamicat

But, would One Photon Illuminate an Object enough in said room?

Don't objects absorb and release photons, in their respective wave lengths?

Well, unless the object is luminescent or incandescent it will only reflect, absorb, or transmit visible light. The emission will, with few exceptions like Anti-Stokes fluorescence, be in the longer wavelength of the infrared band.

Originally Posted by Jamicat

Add in the equation One bar of gold and another of silver.

Would one photon be able to decipher one from the other, by color band?

Is a photon a carrier particle that can morph at will, depending on atoms it encounters, then changes wavelength the same way?

Stop making me think dammit...

It would depend on the reflectance of each material at the wavelength of the incident light. An example of this is any sort of colored light bulb, for example a red light - it tends to make everything appear red because there is no other color to reflect, so even if you have, say, a green object and a brown object they will look more or less the same.

[Sunspace]

Originally Posted by Yeti08

I know that Wikipedia says that a rod cell can be sensitive to a single photon, with no citation provided, but I have a hard time believing that. For instance, even humans emit visible light up to 100 photons per cm²

(bolding mine) Say what? Are we talking about the upper tail of the emission curve here?

Originally Posted by Yeti08

... and I've never seen a glowing person before even in the darkest of settings.

Were you in a place with less than 100 photons per cm² emission otherwise? The human would have to be the brightest light source.

Is there a minimum photon energy needed for our eyes to receive?

[Yeti08]

Originally Posted by Sunspace

(bolding mine) Say what? Are we talking about the upper tail of the emission curve here?Were you in a place with less than 100 photons per cm² emission otherwise? The human would have to be the brightest light source.

Is there a minimum photon energy needed for our eyes to receive?

Not sure what emission curve you're talking about. This is a form of luminescence, not incandescence.

http://en.wikipedia.org/wiki/Biophoton

I'm at home so I don't have the journal articles, but I know that one estimated 100 photons per cm2 and I think one stated that it was something like a thousand times less than a human eye could perceive. This is in the visible range, IIRC something like 480-640 nm is the range for one mechanism for biophoton emission. I've been in some very deep caves before where I assume there weren't any light sources (after headlamps and flashlights were turned off), and everything was absolutely pitch black. I don't think photon energy really has anything to do with it since the energy is only a function of wavelength (E=h*c/lambda), so a red photon at 650 nm will be the same regardless of what emitted it.

[Washoe]

Originally Posted by Chronos

Dangit, I know I saw a statistic like that somewhere, but obviously, I horribly mangled it. What was it I was thinking of, then?

Maybe you were talking to Richard Feynman’s cat?

[Chronos]

100 photons per cm^2 per what time? Is that per second, per lifetime, what?

[Yeti08]

Originally Posted by Chronos

100 photons per cm^2 per what time? Is that per second, per lifetime, what?

There should have been a per second in there.

[Omphaloskeptic]

Originally Posted by Chronos

OK, running the numbers, it turns out I was dead wrong. For a flight path of 3 meters, and blue light, one photon at a time would mean about 5e-11 Watts.

Dangit, I know I saw a statistic like that somewhere, but obviously, I horribly mangled it. What was it I was thinking of, then?

For a flight path of 2 feet, one photon at a time means about 500 million photons per second. The human eye needs to receive photons at about 50-100 per second to trigger a response, or so I've heard (not-much-of-a-cite). So you could paint the retina with a 5 megapixel image (better than a Kindle!), one photon at a time, from a scanning laser source a couple feet away. From across the room you'd lose some resolution but could probably get a readable large-print page.

If my BOTE calculations are right, a star like the Sun, 4 lightyears distant, would emit about 5E-11 watts onto a human retina, so this number isn't really all that absurd.