LED Construction

This is my understanding, but I may be a bit out of date with some of it and could be wrong in some of the detail.

The substrate of most active devices (ICs, transistors, diodes, LEDs, etc) comes from the large, circular wafer of semiconductor material that the subsequent processing steps are done to. In the very early days other methods were used, but now wafer processing dominates (because it lends itself to handling the fabrication of large numbers of devices in an efficient way).

The n-type layer is formed in the substrate by gas diffusion in an oven (doping) or, for the kind of IC where the features need to be very small, ion implantation from an ion beam. Then the p-type region is diffused into the n-type. Once you have done those steps you have three layers (you'd have many more for an integrated circuit). At this point, you'd only call the base layer of original, untouched semiconductor material the substrate.

The active region forms around the join between the n-type and p-type material when the diode is forward biased. If you want to know how that occurs, then any semiconductor physics textbook will tell you [I'm not going to because I can't remember it in enough detail - sorry]. It's the same process for any p-n diode. There is nothing special about an LED other than that the radiation given off in the active region is of a frequency that is useful (which is done by choice of semiconductor and the dopants) and effort has been made to get as much as possible out of the device rather than have it reabsorbed and ending up as heat. Although getting the radiation out sounds easy, in practice it is difficult and a lot of work goes into improving the process (if you have any good ideas, run off and patent them quickly). High intensity LEDs use techniques like having arrays of holes etched into the surface that use interference effects to move the light across what is effectively an impedance mismatch between the material and air, and clever stuff like that.

Drawings don't usually give you an idea of scale - the diffused layers are very thin, so the substrate is way taller and serves as the mechanical support for the whole thing.

There is an exception to this for some kinds of LEDs. They have a non-semiconductor substrate (usually sapphire, I think) that a very thin semiconductor layer is grown on. The sapphire distorts the atomic lattice of the semiconductor material and that results in a different frequency of light emission to what you'd get with an ordinary substrate of that material.

Edit

That will teach me to post from memory without looking it up. I was making up my own physics there. The recombination is different for LEDs ('considerable amount of direct recombination') to what happens in an ordinary diode ('mostly via traps'). The difference is down to the type of semiconductor.

This is my understanding, but I may be a bit out of date with some of it and could be wrong in some of the detail.

The substrate of most active devices (ICs, transistors, diodes, LEDs, etc) comes from the large, circular wafer of semiconductor material that the subsequent processing steps are done to. In the very early days other methods were used, but now wafer processing dominates (because it lends itself to handling the fabrication of large numbers of devices in an efficient way).

The n-type layer is formed in the substrate by gas diffusion in an oven (doping) or, for the kind of IC where the features need to be very small, ion implantation from an ion beam. Then the p-type region is diffused into the n-type. Once you have done those steps you have three layers (you'd have many more for an integrated circuit). At this point, you'd only call the base layer of original, untouched semiconductor material the substrate.

The active region forms around the join between the n-type and p-type material when the diode is forward biased. If you want to know how that occurs, then any semiconductor physics textbook will tell you [I'm not going to because I can't remember it in enough detail - sorry]. It's the same process for any p-n diode. There is nothing special about an LED other than that the radiation given off in the active region is of a frequency that is useful (which is done by choice of semiconductor and the dopants) and effort has been made to get as much as possible out of the device rather than have it reabsorbed and ending up as heat. Although getting the radiation out sounds easy, in practice it is difficult and a lot of work goes into improving the process (if you have any good ideas, run off and patent them quickly). High intensity LEDs use techniques like having arrays of holes etched into the surface that use interference effects to move the light across what is effectively an impedance mismatch between the material and air, and clever stuff like that.

Drawings don't usually give you an idea of scale - the diffused layers are very thin, so the substrate is way taller and serves as the mechanical support for the whole thing.

There is an exception to this for some kinds of LEDs. They have a non-semiconductor substrate (usually sapphire, I think) that a very thin semiconductor layer is grown on. The sapphire distorts the atomic lattice of the semiconductor material and that results in a different frequency of light emission to what you'd get with an ordinary substrate of that material.

Edit

That will teach me to post from memory without looking it up. I was making up my own physics there. The recombination is different for LEDs ('considerable amount of direct recombination') to what happens in an ordinary diode ('mostly via traps'). The difference is down to the type of semiconductor.