2
\$\begingroup\$

As I'm sure most people here know, the Shockley diode equation is

$$ I_d=I_s[e^{\frac{V_d}{nV_t}}-1] $$

Notice that none of the quantities here other than \$I_s\$ and \$n\$ are dependent on the diode itself. \$n\$ varies widely even sticking to just silicon diodes, so clearly the different voltage drop of other diode types can't be adequately explained by that.

Is the variance in \$I_s\$ alone adequate to describe Schottky junctions or diodes of other semiconductors (like GaAs or SiC or the wide range of LED semiconductors), to the same accuracy that the equation describes silicon diodes? Or is there some assumption built into the equation that makes it only valid for silicon diodes? Can \$n\$ be far outside the [1,2] range it's commonly cited as being restricted to, in non-silicon diodes? Are there dropped terms that are very small for silicon diodes that are non-negligible for other types?

\$\endgroup\$
  • \$\begingroup\$ It is definitely not a full description \$\endgroup\$ – DKNguyen Jul 15 at 17:34
  • \$\begingroup\$ @DKNguyen Hence why I asked "does it describe to the same accuracy" instead of "does it fully describe". \$\endgroup\$ – Hearth Jul 15 at 17:35
  • \$\begingroup\$ Oh, that I don't know. \$\endgroup\$ – DKNguyen Jul 15 at 17:35
  • \$\begingroup\$ Why not plug some numbers in and see? \$\endgroup\$ – Andy aka Jul 15 at 17:43
  • 1
    \$\begingroup\$ @Andyaka Few LEDs--none that I've been able to find--have \$I_s\$ listed in their datasheet. And I also feel like this would be a good question to have on here, anyway, in case anyone else is wondering the same thing. \$\endgroup\$ – Hearth Jul 15 at 17:52
3
\$\begingroup\$

The Shockley equation with minor modifications is what is used in SPICE simulations for forward current in silicon, Schottky, and LED diodes.

enter image description here

The parameter list is given as (from here):

enter image description here

The main difference between the different types (other than the number such as Is which is quite different for an LED) is the temperature dependence of the saturation current for Schottky diodes. It includes a resistance term.

The reverse bias treatment is different from the Shockley equation and includes the GMIN for voltages < -5nV and the breakdown region exponential (the latter of which isn't covered by the Shockley equation). The GMIN business is to aid convergence of the diffeq solver rather than to improve accuracy.

Note: Vj (junction potential) and M (grading coefficient) are used to model the capacitance of the diode- M is important in reverse bias and Vj for forward bias.


As an example of an LED model, the LTspice model for Lumileds LXHL-BW02 (a 350mA power LED) has Is = 4.5E-20, N=2.6, Rs=0.85

In general you will not find the saturation current on a diode datasheet (you should not use the reverse leakage term), rather you would adjust IS and N to fit the forward V-I curve.

| improve this answer | |
\$\endgroup\$
2
\$\begingroup\$

The Shockley diode equation is a model for p-n junctions. One can derive such equation without ever mentioning the name 'Silicon', because most of the semiconductor physics equations used are true for both direct band gap semicondutors and indirect ones.

As a model, it works as long as you keep the model within the assumptions you make. The equation you quote is no way close to predicting a diode's behavior due to generation-recumbination currents (often named \$J_{sc}\$) not being neglibgle (and they aren't), nor it is for describing high-injection effects, series resistance, quantum tunneling mechanisms and so forth.

Again, the equation you quote is only a good model within (rather strict) limits. But certainly these aren't unreachable approximations and that's why such model is so useful.

Finally, I can't see how that equation may be employed for M-S junctions or LEDs - they are entirely different topologies and entirely different physics, there's no way a bare exponential could cover that too.

| improve this answer | |
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.