I'm really confused about how JFETs work. JFETS are usually significantly higher doped donor xor acceptor atoms.

My professor mentioned that MOSFETs usually have an oxide in between the the metal and semiconductor. One of the benefits of this oxide is that it prevents leakage (which I think it minority-carrier current correct?). Can someone explain why this is?

He then went on to say that suffers from leakage current. However if the JFET is so dominantly doped with one type(n or p), then this doesn't really matter, right? Can someone explain why I'm wrong please?

Any pointer to reference material is greatly appreciated as well!

  • 2
    \$\begingroup\$ A JFET has gate-to-channel junction that is a diode - usually run in reverse-bias. A MOSfet has gate that is completely insulated from channel by thin non-conducting oxide. \$\endgroup\$
    – glen_geek
    Commented Oct 6, 2020 at 21:08
  • 1
    \$\begingroup\$ SO JFETs do suffer from leakage current? Shouldn't the doping be so one sided that it doesn't matter? \$\endgroup\$ Commented Oct 6, 2020 at 21:13
  • 1
    \$\begingroup\$ @MatthewEngelstein "So one-sided that it doesn't matter" is a relative matter. MOSFETs literally have an insulator and leakage can still matter so I wouldn't be surprised if it was specified at least some of the time. \$\endgroup\$
    – DKNguyen
    Commented Oct 6, 2020 at 21:20
  • 1
    \$\begingroup\$ Gate Leakage Current = It refers to the gradual loss of electrical energy caused by electronic devices even when they are turned off. While JFETs allow the gate leakage current on the order of 10^-9 A, the gate leakage current for MOSFETs will be of the order of 10^-12 A. Two tunneling mechanisms are responsible for the gate leakage, Fowler-Nordheim tunneling and direct tunneling iue.tuwien.ac.at/phd/stockinger/node17.html differencebetween.net/technology/… \$\endgroup\$
    – CFCBazar
    Commented Oct 6, 2020 at 21:59
  • \$\begingroup\$ Forward biassing the gate will turn on the diode and yes, cause leakage current. And this also explains why there are no enhancement mode JFETs : because they are all depletion mode, the gate voltage keeps the diode junction (the J in JFET) reverse biassed. \$\endgroup\$
    – user16324
    Commented Oct 6, 2020 at 22:16

3 Answers 3


JFETs have a junction and an associated Is "leakage" current. However it's often pretty small and one can use a JFET as a low leakage diode. For example, J201 has 100pA maximum leakage vs. several nA for a 1N4148 typically and 25nA guaranteed (all numbers here are at 25°C).

The \$\text{SiO}_2\$ oxide gate insulation in a typical MOSFET has very low leakage typically, though datasheets often do not guarantee anything like the typical. Win Hill has measured typical gate leakage currents in the "few fA" for the 2N7000, for example, even though the guaranteed number is 100nA.


JFETs and Mosfets are significantly different.

The gate on a JFET is simply a diode junction, hence the name Junction Field Effect Transistor. The leakage current will be similar to that of a small signal silicon diode when reverse biased, the normal operational situation.

MOSFETs have a gate that is insulated from the device by a layer of silicon dioxide, hence the name Metal(lised gate) Oxide Silicon Field Effect Transistor. It's still leaky though.

Depending whether the MOSFET is enhancement or depletion mode, it may require a positive or negative gate drive.

Both Jfets and MOSFETS are available in n channel or p channel variants. Note that a n channel jfet will have lower on resistance when used as a switch (Vgs=0) than an equivalent p channel part.

Insulated gate bipolar transistors (IGBTs) are somewhat similar too but have the output characteristics of a transistor, not a fet.

Jfets are mostly used in signal amplification and switching (they make great switches) and MOSFETs are widely used in power electronics.


I recall, from diyAudio.com thread on "simplistic NJFET RIAA", discussion of the "hot electron" behavior that would shift the threshold voltage (I believe was the problem).

The cure was to LOWER THE drain_source voltage, using cascade devices (bipolar or fet worked fine).

  • 1
    \$\begingroup\$ Let me guess. That diyAudio circuit used no feedback ? I could speak volumes about threshold shift too. \$\endgroup\$ Commented Oct 8, 2020 at 17:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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