optical sweep generator

I have one ancient Sony cam (with a floppy disk!) that sees pretty deep.

That was from a laser on loan from NIF.

I think modern cams tend to have "better" IR filters.

Old CCD ICX259AL without IR-cut filter security camera, laser SFP module I directed the light from the laser to the camera lens in a dark room.

IR seems to show up blue or purple on most cams.

I experimented with infrared photography:

spontaneous emission or two-photon absorption with 1550nm laser light ?

Normally CCDs are completely insensitive past about 1200 nm. The bandgap of Si is about 1.12 eV, whereas 1 um photons are 1.24 eV, so the band gap corresponds to 1100 nm. To see 1.55 um (0.8 eV), you need about 300 meV worth of help.

Thermal spread of the Fermi level helps some, and with bright enough light you may be heating the silicon locally by enough to get some signal--400C corresponds to 0.034 eV, so the probability of seeing anything goes down by a factor of exp(-9), i.e. about 8000. At room temperature it's exp(-11.6) or 1E-5.

There's more physics than that in it, e.g. the phonon spectrum of silicon and the possibility of two-phonon interactions (which are observed in silicon Raman spectra). It might get a bit of a boost from the impurity bands, but they're generally less than 0.02 eV. (If it were much higher than that, BJTs would stop working at room temperature.)

Cheers

Phil Hobbs

Surveillance cameras are cheap, and they accept IR (but I'm not sure of the range). It's not hard to disable their illuminator IRLEDs.

So, for wavelength, you could use a transmission grating (Edmund used to sell these for a buck, in projection-slide frames) and a screen, with just a plastic lens to focus to a spot. View the spot with the camera, and scribble a few marks on the screen for reference.

Yet another trick, is the 'here's a ruler, and a laser, measure the wavelength' scenario. You can direct a beam obliquely down the marks on a (high precision Starrett) ruler, and get reflection AND diffraction spots due to the periodic marks. Your fiber output, though, won't be collimated.

Edmund still sells the cheapos You don't need high accuracy to tell the difference of those three colors.

Yeah I was going to say some shallow impurity in the Si? Can you get indirect absorptions in Si? (photon plus phonon)

Hmm a bunch of article for phonon assisted absorption.

I will assume that the laser 1550nm shines a little at a wavelength ~1000nm

Brilliant idea: after an e/o converter is tested, we could snoop it with a cell phone cam and see if it's a different color for 850/1310, or dark for 1550.

Tried it with my phone: the 850 and 1310 both look purplish.

I don't have a 1550 handy.

I think so. Longitudinal phonons have zero angular momentum, like photons, so it would just be the size of the interaction matrix element. Photodiod es do cut off further into the red at high temperatures.

Cheers

Phil Hobbs

Light-emitting diodes as measurement devices for femtosecond laser pulses :

Any working LED (from long-wavelength infrared to ultraviolet) is visible in the thermal imager))

Some of us have learned to get by with a bare grating and an upconversion p hosphor or storage phosphor or our monochromator and a Lead Sulfide detect or. I have a HENE that lases at both 632.8 and 1150 for calibration. Near I R also quenches some of the commercial ZnS:Cu phosphors out there. I've use d that method too, and have commercial viewing cards out to 12 microns that also quench.

A wonderful fellow with the unique name of Gorton Fonda came up with lead d oping ZnS:cu so it stores trapped electrons and then releases them when hit with IR. His patents and papers are an amazing read.

The alignment disks can take much higher powers then the cards as they are ceramic backed.

Don't laugh, but this is how I used to verify and align Optical Parametric Oscillators. Of course with them, due to Conservation of Energy, if you kno w the pump and idler wavelengths you know what the IR product is.

You could also die of sticker shock buying a Flame-NIR USB spectrometer at Ocean Optics/Ocean Insight, too..

Steve

Am 15.05.20 um 23:28 schrieb snipped-for-privacy@gmail.com:

...any color of the spectrum, from infra-dead to ultra-violent...

(Hitchhiker's Guide To The Galaxy)

oh yes! The glass window of a defective "black" laser will shine like a young laser in a thermal imager!

After we test an o/e converter, we know it makes light. What we're not sure is what wavelength it's making. We have got some laser diodes that were mislabeled, or that we mixed up ourselves.

I could take the light and feed it into a multimode 50:50 splitter, then into two photodiodes, Si and GaAs. The signals might have enough info to resolve 850/1310/1550. But I'd rather buy something.

Depends on what resolution you want, but the simplest way is starting with a simple diffraction grating.

Is a grating blazed for 1.6 microns, cost is 72.00.

On Line Grating Calculators are wonderful, soL At a zero degree angle of incidence, no slits, no focusing mirrors, not a Czerny Turner setup, just a collimated beam input, the deflection angle of the first order beam is

How ever you want to detect and measure is up to you, put a diode with a p inhole on a sliding stage, IR viewer card, a crude sine bar with a bearing and micrometer plus a photodiode, just a protractor, bolt it to a table and mark spots on the wall, there are a lot of possibilities.

A quick check of the NIST spectral data base shows NE-I has a few distinct very strong lines, from 1200 to 1500, so a Neon glow lamp could be used as a reference.

These folks claim to have stretched the silicon spectrum out to nearly 1550 and the price is not that bad.

If you want to measure the center wavelength accurately, Ocean Insight's fi ber coupled USB spectrometer is still your best shot.

Steve

Maybe someone has already asked about this, but do the different laser diodes draw different amounts of current? If so, can you accomplish the same goal simply by checking the supply current in the production test rig?

I like to do that in test scripts because it sometimes catches things that other tests don't. If something takes 20% more or less current than usual, chances are there's a good (bad) reason for it.

-- john, KE5FX

I'm confused. Is there a two-photon absorption in silicon from a non-pulsed source 1300 or 1550nm?