idea to make an LED with variable light output frequency

[Jamie]

Hi,

A typical LED's output light frequency seems to be proportional to the
semiconductor's bandgap voltage, with wavelength decreasing as the
voltage increases:

http://en.wikipedia.org/wiki/Light-emitting_diode#Colors_and_materials

So if the bandgap voltage is changed variably, then the LED should be
able to output variable frequency light. If the LED wafer is sandwiched
between two metal screens, ie. to put it inside a capacitor, then the
capacitor could modify the electric field of the bandgap, to adjust the
output light frequency. Anyone ever try this?

cheers,
Jamie

 

[Tim Williams]

A semiconductor junction makes a pretty good capacitor. You can try
forcing a lot of voltage across it, unfortunately those darn electrons
keep slipping past, and you need exponentially greater currents to
increase the bias (plus linear voltage to overcome resistive losses in the
bulk semiconductor and interconnects).

That said, green LEDs (GaN) turn yellow and orange nicely when pushed well
beyond rated current.

Tim

 

[Jamie]

Hi,

A typical LED's output light frequency seems to be proportional to the
semiconductor's bandgap voltage, with wavelength decreasing as the
voltage increases:

http://en.wikipedia.org/wiki/Light-emitting_diode#Colors_and_materials

So if the bandgap voltage is changed variably, then the LED should be
able to output variable frequency light. If the LED wafer is sandwiched
between two metal screens, ie. to put it inside a capacitor, then the
capacitor could modify the electric field of the bandgap, to adjust the
output light frequency. Anyone ever try this?

cheers,
Jamie

For a diode bandgap of 2V over nanometers, it would require a kV
electric field from the capacitor to have much effect on the bandgap
voltage, but if the bandgap voltage was only be modulated 1% (+-0.02V),
that would still allow a multi-GHz RF modulation to be put directly into
the light which could be good for fiber optic data transfer.

cheers,
Jamie

 

[Jamie]

Can the bandgap of a semiconductor be changed with an applied electric
field?

I've seen red LED's turn Org when the ambient temperate around it gets
hot!!!!!!!!!!! Maybe its a design of the LED, who knows.

JAmie

 

[Spehro Pefhany]

I've seen red LED's turn Org when the ambient temperate around it gets
hot!!!!!!!!!!! Maybe its a design of the LED, who knows.

JAmie

Try dipping them in LN2.

http://www.ap.smu.ca/demos/index.php?option=com_content&view=article&id=181&Itemid=78


Best regards,
Spehro Pefhany

 

[Jamie]

Yeah the band gap is pretty much set by the atomic spacing. So, I
think, you might be able to squeeze on it and change the wavelength.
Or change the temperature as Spehro suggests. I've heard that most
LED's move to higher wavelengths as you cool (shrink) them, but some
go the other way. Yellow's become orange, or something like that.

Hi,

Instead of using a capacitor to try to change the bandgap properties,
it would be better to use a miniature high-gain antenna that has its
beam focused onto the bandgap directly, and then modulate the antenna
RF for the desired light modulation.

cheers,
Jamie

 

[Martin Brown]

I have some Nichia blue LEDs here (you have seen them in my radiation counter AVI movie perhaps),
and the datasheet gives a forward current versus wavelength graph on page 10 (of 15).
It changes from about 474 nm at 1 mA, to 486 nm at 100 mA.
It also changes from 469.5 nm at -40 °C to 472 nm at 80 °C for 20 mA constant current.

So it seems the color change is more for a delta I.
google (is on conrad.nl site):
180763-da-01-en-LED_3MM_BLAU_15_8_200_MCD_NICHIA.pdf
or look up the LED type directly:
NSPB300B

LEDs are not particularly monochromatic with an output bandwidth of
typically 50nm either side of the nominal peak frequency. Some of them
have very wide wings - the first (pale) green ones originally contained
a lot of yellow as any basic spectroscope would quickly show.

I'd be quite interested to find a 20nm or narrower high intensity
amber/yellow LED centred on 590nm or as close as possible to it. There
is a Cree XP-C part that isn't too far off the mark still a bit too much
of a tail into the green for what I want to do though.

 

[Martin Brown]

Maybe add an interference filter on the output? You can get ~80%
transmission in a 10nm window.
Or does that 'blow the budget'?

Basically yes. I use a Schott colloidal glass OG590 low pass filter at
present which sees off the unwanted green very well, but it would be
nice to avoid any extra optics (and associated cost).

 

[Jamie]

Hi Jamie, Do you have any solid state physics books? It's not easy
to change the band gap of a material. Just reading quickly in Sze's
"physics of semiconductor devices" That the bandgaps of Ge and GaAs
increase with pressure and for Si it decreases (go figure).
We temperature tune laser diodes, but it's only about 0.25 nm/deg C

Hi George,

Here is another way to change the bandgap properties:

http://micro.magnet.fsu.edu/primer/java/filters/aotf/index.html

cheers,
Jamie

 

[Jamie]

The nomenclature is a bit confusing; that is a description of
modulating the field of the depletion region, and the BANDGAP
voltage is a pressure-modified variant of the electronic orbital
energies. You can change the bandgap with pressure (very high
pressure), but not with an electrode.

Hi,

The stark/zeeman effect show that the bandgap electron orbital energies
can be changed with an electromagnetic field:

http://en.wikipedia.org/wiki/Stark_effect
http://en.wikipedia.org/wiki/Zeeman_effect

cheers,
Jamie

 

[Jamie]

Hi,

The stark/zeeman effect show that the bandgap electron orbital energies
can be changed with an electromagnetic field:

http://en.wikipedia.org/wiki/Stark_effect
http://en.wikipedia.org/wiki/Zeeman_effect

cheers,
Jamie

Hi,

Is it correct to think of the bandgap as an "electron accelerator" that
doesn't actually have any electron orbital transitions during operation?

cheers,
Jamie