Issues with measuring FemtoAmp DC from a device

[Idli]

Hello all,

I am making a device which produces small direct current around 100 femto amp (1 fA = 1e-15 A) and am trying to measure it directly using a Keithley 6485 Picoammeter.

It is a device with two metal foils sandwiching an insulator (Alumina) foil. Very thin foils, 0.5 to 1 mm thin. I need to attach connections to the sides of the foils (0.5mm sides) and want the best solution for such small current measurements. Please suggest. I cannot however use destructive methods like soldering etc. I tried using silver based conductive glue to attach copper wire, but the set up did not work.

For shielding, I put the whole device inside a grounded aluminum box.


Other than that, I am using a coaxial cable with BNC connector to connect the device and the picoammeter.

The last time I did a measurement, I saw just noise, random noise. No signal.


I do not have a background in electronics or electrical sciences, therefore descriptive answers would be helpful. I earnestly request you to avoid hi fi terminology/abbreviations because I am not used to them.

Any help, any leads will be highly appreciated. Thanks a ton in advance.

MOD EDIT!! No need for two posts on the same subject!!

 

[JimB]

It is a device with two metal foils sandwiching an insulator (Alumina) foil

From your description, the "device" sounds like a capacitor.

I saw just noise, random noise. No signal.

What sort of signal did you expect to see?
What is the device, what is it supposed to do?

JimB

 

[alec_t]

What metal?
Are the metal foils removable from the alumina?
Why would soldering be destructive (the melting point of alumina is 2072C, so it should surely be ok at soldering temperatures ~200C)?

 

[AnalogKid]

Even at the fA level, electric current comes from energy conversion. What is the nature of the device you are measuring? Years ago there was an ad for a new IC (NS? LTC? BB?) with input bias current so low that it could detect an input current as low as 25 electrons per second. At these low current levels, normal surface contamination on a pc board can wipe out most signals, probably what is happening to your measurements.

ak

 

[Idli]

JimB
From your description, the "device" sounds like a capacitor.
no. due to a reaction in them a potential difference is created between the two foils. the strength is too low and we expect a direct current of around 100 fA, when the two foils are connected across a picoammeter.

What sort of signal did you expect to see?
A quickly rising current, getting stabilized at some e-13 A. A few ups and downs okay, as the system will anyways be very noisy but not so clueless signal as I reported in my first post.


alec_t
Why would soldering be destructive?
There are multiple metals to be tested. But, some of them are toxic and cannot be handled so easily. And we need to create electrical contacts on sides of the metals, not alumina.

AK: normal surface contamination on a pc board
no pc board. there could be such contamination in my setup also, but what is the best way to remove it?

 

[atferrari]

Google for Jack Ganssle. He published very recently about embedded designs low power oriented. At the last part of the article he discuss about board conditions and current leakages.

Do not be confused; he writes a lot about programing.

 

[JimB]

Particle detector.
What sort of particles?
Alpha, Beta, Neutron, Higgs Boson, or ?
I would expect that the aluminium screening box would stop alpha and beta particles.

What sort of signal did you expect to see?
A quickly rising current, getting stabilized at some e-13 A.

I take it that this is a (continuous) stream of particles rather than individual particle events.

I don't understand the operation of the detector.
From your description I think that the detector converts the kinetic energy of the particle into electrical energy, is that correct?
Or,
Does the detector create a conducting path between the two metal foils, in a similar way that a Geiger-Muller tube creates a conducting path as the particle passes through? If this is the case, you will need some kind of polarising voltage to provide the current.

Can you give some more information please?

JimB

 

[tomizett]

Would it not be possible to simply clamp two flat electrodes to the outside of the metal/Amumina sandwich? The complete setup would then consist of the Alumina foil (inside), the metals-under-test, the electrodes (copper or whatever is easy to wire to), an insulator (to insulate from the clamp, which may be conductive) and some kind of releasable clamp (even something as simple as a cloths peg?).

Would thermocouple or battery-like effects be a problem in this kind of setup? This stuff is out of my field, so I don't know...

There aren't a lot of options really. Either you simply hold the connection wire in mechanical contact with your experiment, or you've got to look at soldering/welding/brazing - apart from anything else, if you're using different metals then they're all going to react differently to these techniques.

We'd love to know what the nature of the experiment is though!

 

[kubeek]

At such level of current, are you sure that the interconnecting coax, the connectors and everything else is up to the task? I wouldn´t trust the isolation in the coax for example, my bet is that the leakage is way too high for such levels, and that is probably what makes you measure noise instead of a useful signal.

 

[KeepItSimpleStupid]

Professionally, I did a lot of low current measurements. You HAVE to loose the coax. Connections have to be Triax. Triax cables are about $100+ USD for 3 feet.
The connections consist of ground, guard and a center conductor. Keithley used a graphite impregnated shield to minimize current generation from flexing.

BTW, Triax come in 2 lug and 3 lug. The two lug will fit BNC's but will destroy the Triax connector.

Guard has the same potential as the inner conductor, thus leakage is minimized.

You box should be grounded. Inside your box there should be a separate plate insulated by at least Teflon. I don't know if you have to go better than that. The guard terminal is attached to the second insulated plate.

I'd have to see what I'd have to attach to to make suggestions. One way is to use a metal plate, and a thin shim separated by a Teflon sheet where your device could sit. The metal plate might be sufficient. In the system, I built, there was a grid of tapped holes that we used to attach custom probes and for another case, I designed custom Kelvin probes (other types of measurements) that were initially placed by magnets and held in by clamps using the grid. The design accommodated many 1' x 1" samples.

I really don;t see any reason why silver paint would not work (Used in SEM microscopy) .wires might be wire wrap wire or gold wires used for ultrasonic bonding . In a Hall Effect set-up I designed, the samples were placed on glass slides that had conductive copper tape on them. Wiring was done to the pads prior and Pogo probes could contact the pads along the edge, so the glass acted as a carrier.

A conductive elastomer might work for you, so might a Zebra connector See **broken link removed**. The Zebra connectors are rubber connectors with fine wires in a row and are used too mount LCD displays. They were salvaged from broken equipment.

If you use substrates like glass, they have to be squeaky clean. The rinse I used was semiconductor grade trichlorofloroethane, methanol and Freon TF blown off with dry Nitrogen. Freon TF was phased out because of environmental problems. The Freon TF was a dip. I could measure an appreciable current on a piece of glass with a methanol film on them.

You can also use the coulombs mode, but you generally have to release zero check and hit zero correct shortly after. Divide by the time and you get current.

At these current, both motion (moving a wire in the earth's magnetic field), and flexing (triboelectric effect) create currents and to some extent the piezoelectric effect.

Also look here: https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CCIQFjAAahUKEwihq86M2PHGAhUMnIAKHWy8D_g&url=https://www.keithley.com/data?asset=6069&ei=VRixVeGsMoy4ggTs-L7ADw&usg=AFQjCNGjLn_QAApj-iiWtbSexwmokhaVVQ&sig2=YORClN1RGcTkDwh0clOESA&bvm=bv.98476267,d.eXY

You might need to use a quartz carrier or better.

 

[Tony Stewart]

Ultrasonically welded gold strand wire is most commonly used method similar to those used for different metals such as those in 5mm LEDs which are visible on the anode side. But that may deform the surface is it is soft.

Considering the high resistance, I think the carbon elastomer Zebra connector would be best. It must be mechanically well supported for stable readings as any change in capacitance will create a voltage from the stored charge. ΔV=Q/ΔC

edit..
At these low levels there ought be some Seebeck (thermocouple) effect too.

 

[analog1]

find out the input impedance of the keathely meter and connect in series usually 11 mega ohms so the reading will be current across a 11 mega ohms close enough?

 

[KeepItSimpleStupid]

analog1:
That Keithley meter is strictly a feedback ammeter primarily designed for low current measurement. When inserted in the circuit, it's voltage drop is tiny. It does not sport a voltage mode.