How long a wavelength can a commercially available LED deliver? I mean, in LED mode, not thermally radiated because the LED is hot.
Can you get out to 2 um?
Thanks!
Gallium arsenide can put out light at around 870 nm.
Laser diodes can put out light at 2680 nm.
If you can make a laser diode at a wavelength, you can make a light-emitting diode at that wavelength.
But the question was for commercially-available LEDs. There’s not a lot of call for ones at wavelengths beyond 1 micron, since they generally are paired with a detector, and detectors get to be more complicated out beyond 1.8 microns, and you can generally do the same things (and more efficiently) at shorter wavelengths, so why bother going longer? Longer wavelengths are generally used for optical umping applications. I’m pretty sure I’ve seen them out to just beyond 1 micron.
Here’s a paper on 4 micron LEDs, but they’re not commercial:
Here are 8.6 micron diodes, but not commercial:
Here are 950 nm diodes:
http://www.nteinc.com/Web_pgs/infrared.html
A laser diode is by definition an LED. You can buy quantum cascade laser diodes from a number of vendors. Some of these extend down to THz frequencies. For example, ALPES Lasers sells an 85 micron wavelength diode (3.5 THz). Don’t forget to bring your liquid nitrogen, since it won’t operate at room temperature.
http://www.alpeslasers.ch/lasers-on-stock/lasersSPECTHz.html
Nifty, all. Thanks!
CalMeacham, I use $40 thermopile sensors with windows like zinc selenide, and they’re clean and simple. I presume you were thinking of other classes of detectors, but if I have the diode right there, the thermopile may be plenty sensitive enough.
JWT Kottekoe, I will definitely check out ALPES. I have running LN2 available for the asking. 3.5 THz? Wow. So, how high do electronic oscillators get these days?
You can get detectors a lot cheaper than $40 for inexpensive applications, though. If you want a sensor for your faucet on/off or your hand dryer there’s no point in getting a 440 detector, which is why most IR LEDs are at shorter wavelengths.
Your bartender must be thrilled. 
I’m not sure of the answer, and it depends on your definition. The THz range has been difficult to reach with sources, since it is high for purely electronic approaches, and low for optical techniques. The quantum cascade laser has been a breakthrough in this field.
With regard to electronic oscillators, transistors are available with characteristic frequencies of several hundred GHz. You can probably get a fundamental oscillator that operates above 100 GHz (0.1 THz). From there you can multiply up to higher frequencies using fast diodes, perhaps up to several hundred GHz. There are also pulse techniques that can give broad spectrum signals with components in the THz range. Another approach is to use vacuum tubes, for example Smith-Purcell sources, or free electron lasers.
Actually, yeah. There’s a surplus of nitrogen in the air and he gets giddy. Though we do have a fair bit of turnover due to nitrogen narcosis and asphyxiation…
CalMeacham, yes, you are right. I have price ranges and other details in mind and I didn’t say anything about them. My mistake!
I think parts in the tens or perhaps hundreds of dollars are workable for me. Maybe even very low thousands, if the part count at that level is very low. I want to sense attenuation through a sort of filter; this is an application similar to color measurement. I think I’d like to gage visible and NIR intensity at a few different wavelengths out to 2 um. Let’s say at least 4 wavelengths and no more than 10. I also am considering spectrometers with broadband sources, though I don’t need anywhere near the wavelength resolution they all offer. If the little spectrometer boxes you plug into your USB port, the ones selling for around $2000 and even below, would get out that far, I’d do it that way. But alas, they get something like 10X more expensive when we push the long wavelength end from around 1 um out to around 2 um.
I don’t have much experience in IR instruments, but can’t you construct a crude spectrograph using a reflective grating and a detector on a movable stage? For the resolution you’re looking for, you don’t even need any collimating/refocusing optics, just a series of apertures.
I’ve heard LEDs can also be used as detectors, would that work?
Yes. I could also get several detectors and a fixed grating, or a fixed prism. In fact there is a neat thing one can do with a single element lens: note that a chunk of the lens out near the edge looks like a prism. That chunk will both disperse the radiation on the basis of wavelength (which is why single element lenses have chromatic aberration) and also focus it. So, the image of a “white” light looks like a tight, bright spectrum. Relative to a moving grating, this has several advantages that would help my particular case. There are no moving parts, all wavelength bands can be measured simultaneously, I can choose different detector types for different wavelength bands, and in each wavelength band all the radiation within a certain numerical aperture lands on the detector.