Diode forward voltage drop at lower forward current

[vaibhavwaman]

Is making linear approximation for calculating forward voltage drop at lower current than 10uA is good option ? I want to pass current approximately 100nA via this diode and want to understand forward voltage drop in this case ? Or is their any specific reason for which datasheet dont give forward voltage value for very low forward current ?

 

[danadakk]

That graph is logarithmic on x axis, so no on a linear approximation. You
could use power curve fit or least squares to do it. Keep in mind thats a
"typical" curve, not worst case.

If you are measuring T with it one could use a cal routine to aid in improving
accuracy....

The dual current single diode method, attached.



Regards, Dana.

 

[crutschow]

I want to pass current approximately 100nA via this diode and want to understand forward voltage drop in this case ?

As noted the drop is logarithmic with current.
At 25°C the drop shown is about 0.2V for a two order of magnitude drop in current, so the drop from .01mA (10µA) to 100nA should also be about 0.2V, giving a forward voltage of about 0.2V at that current.

Below is the LTspice sim of a similar diode curve showing slightly over 200mV drop at 100nA:
(Note I had to do the sim at 15°C to reasonably match the 25°C curve of the graph you posted.)

is their any specific reason for which datasheet dont give forward voltage value for very low forward current ?

No, it's just that most applications don't operate at that low current, so they don't post it.

 

[D.A.(Tony)Stewart]

Not all diodes are the same for leakage due to the geometry and passivation of the Si dielectric barrier surface encapsulation, which is < 1um thick.
All diodes are thermal sensors with voltage when forward biased (NTC) and current sources with leakage current when <= 0V. Above 300 'C this leakage power poses a threat due to the much higher heat density and higher thermal resistance, (if applied with reverse voltage)

ref: https://www.researchgate.net/public..._surface_leakage_reverse_current/figures?lo=1

Here is a low leakage Si diode.
Plastic SMD package
• Low leakage current: typ. 3 pA @ 25'C

If you need leakage current to be constant, make an SMD oven above room temp with a thermistor, regulator,10 mW SMD resistor all attached to the diode with styrafoam insulation. I've done this to make a cheap XTAL OSC. into an OCXO.

Keep in mind if forward biased, 200 mV/100 nA = 2 Gohms equiv

I suspect the Space Charge Limited (SCL) model of this Poole-Frenkel (PF) effect advanced by (heavens to) Murgatroid and Boole are such that my hunch is the equivalent circuit is constant RC varicap. The Voltage-controlled capacitance or Varicap effect is found in all diodes created by the reverse bias E-field effect to makes the leakage resistance and C constant at a fixed Temp. Tau=VR/IR * C might be temporarily constant while reverse leakage increases and capacitance reduces with increasing reverse voltage until the temperature changes.

But you might find the answer in this thesis. (TL;DR)

The plastic SMD chip is black to block light which is another source of leakage current from the photo-electric effect in all diodes, transistors and IC's.

 

[danadakk]

Keep in mind if forward biased, 200 mV/100 nA = 2 Gohms equiv

Should this be 2 Meg Ohms ?


Regards, Dana.

 

[dick_freebird]

The logV/I characteristic breaks away from the ideal log-linear
at both high current (Rs) and low current (loss of ideality from
recombination / defectivity). Recombination can be
"modulated" post-fabrication by electrical or radiation abuse.

These latter are "high scatter" lot to lot, wafer to wafer, die to die
and go/no-go leakage @ Vr is about the best you can expect
(if devices are even probed / sorted at all).

Layout nonuniformities (like the corners of an ortho layout) can
in effect create a "two tiered justice system" where the corners
conduct early and the flat edges later and the bottom plate later
still, adding inflection-points to the I-V curve even up where
you'd expect log-linearity (think about superimposing 3 diodes
of grossly different size and doping, in parallel).

In my curve-pulling I see this below 100-200mV on a silicon
diode / transistor junction. Above this the ideality (slope) seems
to be pretty consistent.