Electrochemistry in diodes?!

[Paul]

Hi,

I'm doing some diode research, and would like to know of any possible
DC voltages produced in the diode due to electrochemical reactions.
It's a palladium-silicide diode, part number SMS7630, but also has
gold. Over time the atoms could migrate. The SMS7630 is a microwave
diode used for small signals. It is not terribly uncommon for the DC
voltages produced by small AC signal rectification from such ZBD's
(zero bias diodes) to be as low as one microvolt DC. The SMS7630 zero
bias resistance is ~ 5400 ohms. Even the slightest *possible*
electrochemical reaction is important-- e.g., diode square law
detectors.

It seems possible, for example, for silicon atoms migrating some
distance into the palladium side to bond with palladium atoms that
previously had no such bond-- diffusion.

I would like to build an experiment to measure this DC voltage
produced by electrochemical reactions. Does anyone have any idea what
DC voltages I could expect? Nano volts, pico volts, etc. The DC
voltage produced by a lead acid battery is temperature dependent,
where the electrochemical DC voltage decreases with an increase in
temperature; -0.022V/°C-- ref: http://en.wikipedia.org/wiki/Lead_acid_battery#Electrochemistry

Therefore, how about measuring the change in DC voltage produced by
the SMS7630 at say 60F versus 90F? I have a two layer metal shield
system (small and medium shields) and an electrometer that produces a
few femto amps bias current. It could detect down to 0.5uV DC, but if
there's no appreciable chance of these diodes producing up to 0.5uV DC
from electrochemical reactions then it could save a lot of research
time.


Please see page 3 of this pdf for details on this diodes structure:
http://web.archive.org/web/20071110...orksinc.com/products_display_item.asp?did=745


Regards,
Paul

 

[dlzc]

Dear Paul:

Hi,

I'm doing some diode research, and would like to know
of any possible DC voltages produced in the diode due
to electrochemical reactions.

Non-detectable directly. Your only hope is measuring conduction
carrier density over time.

It's a palladium-silicide diode, part number SMS7630, ....

but also has gold. Over time the atoms could migrate.

This does not happen over decades, unless they experience high
temperatures for extended periods.

I'd plan on looking for an secular increase in applied voltage to
drive a given current, or such a decrease in current resulting from an
applied voltage.

David A. Smith

 

[Paul]

Dear Paul:




Non-detectable directly.  Your only hope is measuring conduction
carrier density over time.


This does not happen over decades, unless they experience high
temperatures for extended periods.

I'd plan on looking for an secular increase in applied voltage to
drive a given current, or such a decrease in current resulting from an
applied voltage.

David A. Smith

Thanks for reply David. If I understand you correctly, then that's my
thought as well; i.e., that the DC voltage produced by this diodes
electrochemical reactions would be undetectable. The junction
diffusion occurring on the order of decades sounds significant. To all
intents and purposes, I'm guessing that would amount to *zero* DC
volts. Hmmm, maybe it's a few femto volts, lol. I had no idea
commercial diodes diffused at such a slow rate.

Even if it diffused at a faster rate, I'd imagine the electrochemical
potentials produced from palladium and silicon would be small, perhaps
less than one microvolt.

Correct me if I'm wrong, you're suggesting the electrochemical DC
voltage is too small to be directly measured, so it's better to
measure the long term change in resistance due to diffusion? That's an
interested method that should work. It's a bit time consuming though,
waiting for the resistance to change, but at least doable. Although it
would be interesting to know what the electrochemical DC voltage would
be in such a diode.

Many Thanks,
Paul

 

[dlzc]

Dear Paul:

Thanks for reply David. If I understand you correctly,
then that's my thought as well; i.e., that the DC voltage
produced by this diodes electrochemical reactions
would be undetectable.

You will have some sort of "thermoelectric junction" voltage, as you
would with a "thermopile", but there are some constraints on the
conductors you would use between the two junctions.

The junction diffusion occurring on the order of
decades sounds significant. To all intents and
purposes, I'm guessing that would amount to
*zero* DC volts. Hmmm, maybe it's a few femto
volts, lol.  I had no idea commercial diodes
diffused at such a slow rate.

You can google:
dopant diffusion semiconductor
.... to get a better feel for this. Until you get up near something
like 200 degC, diffusion is near zero for all intents and purposes.

....

Correct me if I'm wrong, you're suggesting the
electrochemical DC voltage is too small to be
directly measured,

You can try using two devices at different temperatures, with
identical wiring between them. See if you measure a potential
difference. But diffusion will be a *very* small signal.

so it's better to measure the long term change
in resistance due to diffusion? That's an
interested method that should work. It's a bit time
consuming though, waiting for the resistance to
change, but at least doable. Although it would be
interesting to know what the electrochemical DC
voltage would be in such a diode.

Let's see what the other folks in the groups you posted to think...

David A. Smith

 

[Marvin]

There are ion-sensitive diodes and transistors that respond
to the Nernstian potential of ions, made selective by
certain coatings. to read up on this, google with the words
"ion sensitive diode" without the quotation marks.