A Bit About The Diode

Most of you already know what a diode is, but how much do you really know about the device?

The diode is a component which allows current to pass in only one direction. Originally they were made by placing a positively charged anode plate within view of a tungsten cathode in a high vacuum. By heating the cathode to several hundred degrees, the metal’s work function is reduced enough that electrons with may leave into the vacuum using only a few volts. These electrons would then be attracted to the cold, positive anode and would flow into it and out of the tube. As the cold plate’s work function was several magnitudes higher than the cathode’s, there was a greater probability that current would flow in only one direction.

While this thermionic process works very well and very fast, the heater requirement ends up making the diode quite inefficient. As a result thermionic diodes are only used when frequencies of several hundred megahertz must be rectified at very high powers; they’ve largely been antiquated by the semiconductor diode in most applications.

Semiconductors are neat little elements. When pure, they are very good insulators and will not conduct. It’s possible though, to coax these materials to either conduct electrons or holes, simply by adding some impurities to the crystal lattice. By throwing a few atoms with more than 5 valence electrons into the lattice the semiconductor will be able to conduct electrons, creating an N-Type semiconductor. Likewise, by throwing in a few electron-few atoms it’s possible to conduct holes, creating a P-Type material. By sandwiching these two types together we can form a PN junction; a diode.

With the P-Semiconductor biased positively and the N biased negatively, electrons easily can flow into the diode, jump across the small 0.7V depletion region and leave to continue on their merry way. If the diode is biased incorrectly though, holes and electrons migrate away from the junction and a very big depletion region is formed. In fact, the junction turns into a few-picofarad capacitor, and in some cases may even be used as such.

As great as they may be, PN junctions are lossy little things. Though they do a really good job of blocking reverse current their 0.7 (now, 0.5) volt depletion region will readily burn 14 watts if 20 amps are to be rectified. Not only that, but it takes significant time for the diode to ‘recover’ after a reverse-bias, thereby limiting the speed at which it may rectify. 1N series diodes are usually no good for anything more than 400Hz! UF series diodes are much faster and may operate at 100kHz smoothly, but anything more than 500kHz is a bit much to ask for. I suppose we can’t complain though, since they sure beat the hell out of coherers!

I hope you all liked this article; in the future we’ll take a look at specialty diodes!

 

28 thoughts on “A Bit About The Diode

  1. The 0.7v/0.5v thing is not a property of age (reading into the use of the phrase “now 0.5” in the article) but a material property. Silicon diodes are typically 0.7v but germanium based ones are typically 0.5v (they both can be different values but generally characterised with these values).
    There are also diodes that are formed in the same way as the “cat’s-whisker detector” in old radios, i.e. it’s a metal-silicon interface to create the barrier, rather than being between P and N-type doped material (this is more commonly known as a Schottky diode and has a lower junction voltage compared to silicon n-p junction).

    1. That last bit might be misinterpreted.
      Just to clarify…
      He is saying that the Schottky is a metal to silicone interface, not a P to N-type interface.

      Sorry. The sentence structure bugged me. :)

  2. One of my favorite LED tricks is to connect both pins through a current limiting resistor to two MCU GPIO pins. The LED can then be reverse biased to charge it as a capacitor. Then flip the cathode to input, and count the number of cycles for the pin to transition to logic zero state. The “capacitor” will discharge as a function of photons hitting the junction, thus turning it into a light sensor. When wired this way, the same LED can be used to both produce and measure light!

    1. You *can* do that, but LEDs make atrocious photodiodes. Real men with hair on their chest use PIN photodiodes, transimpedance amplifiers, and capacitor values chosen for ultimate bandwidth.

      1. It’s worth noting that when an LED is used like this it is *not* used the same way that a photodiode is used. The use of a photodiode is going to require an op-amp and a DAC in addition to the sensor itself. This little hack requires nothing more than a timer and the cheapest LED you can find.

  3. maybe im wrong, maybe im very wrong, but either way,
    im going to say this:

    a small signal diode (1N type), if run at ~1/2 current will have a voltage of about 0.666 …

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.