There are several variables like: concentration of electron deficiencies/holes or electrons (n,p) in p/n-type of semiconductor, concentration of donor or acceptor atoms (Nd,Na) in doped semiconductors, length of pn-juction of diode, electron or hole mobility constant (Dn,Dp), etc.

But all of these variables are unreachable for me. At the beginning I thought at least datasheets should include such data but they don't. Just the ones used practically.

Where/How can I access this kind of information?

  • BTW: All these variables matter for calculating electron/hole average speed (v), current flux(Φ), density of current (J), which are more theoretical calculations. There are also more practical ones like junction capacitance (matters when reverse-biased) and diffusion capacitance (matters when forward-biased).
  • $\begingroup$ Also doping profiles are important for the ideality factor. $\endgroup$ – Pieter Jan 6 '17 at 23:02
  • $\begingroup$ For commercially available components, there is no guarantee that any two devices are actually made the exact same way. You can still buy 2N2222 transistors - the original part is now more than 50 years old (I believe). They are not made the same way. 2N2222's from different manufacturers are made differently. The only ways to get at what you want is (1) obtain proprietary info from the fab, or (2) deconstruct and reverse engineer a device (be sure to buy a large, single lot in either case so your answer applies to the other devices you have). $\endgroup$ – Jon Custer Jan 9 '17 at 15:00

The closest you can get is the equivalent SPICE model for transistor, but not for diode, to my knowledge. In this kind of data junction capacitance will be given explicitly.

Furthermore, doping profile is not uniform and depends on the tecnological process. You can saturate top layer with dopant and then heat the wafer to 900-1200*C to let dopants diffuse. Peak doping level will be at the surface then. You can do ion implantation instead, and peak of the doping level will be buried.

Technological processes differ very much indeed, because the ultimate goal is to make the device useful. In other words, you do not make some diode and then think about its applications, but rather make the diode with certain characteristics.

To some extent you can do "reverse engineering" mathematically.

  1. Frequency response is connected to junction capacitance and carrier mobility. The former is the largest when doping level is the highest, because you have enough impurities to screen electric field on small distances. Mobility is usually limited by scattering on impurities, so you can estimate it based on doping level.
  2. Tunnel breakdown voltage is connected to doping level, since the barrier thickness is important.
  3. Sometimes you can find injection efficiency parameter, which gives you the ration between direct and reverse injection (when you open the diode, the current flows both ways, but due to different doping level, i.e. different number of electrons and holes, it will be quite different).

In principle, resolving the above self-consistently could allow you to estimate doping levels and mobility. The only problem is the geometric sizes of the junction, but they are sometimes published. If not, you can roughly estimate them from the maximum current.

Sorry for not giving you the detailed formulas for each of these steps. They are pretty well summarized in "The Physics of Semiconductor Devices" by S. Sze. Please, feel free to ask for details if I have missed something.


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.