Simple salt-water detector, go/no-go type discrimination or better?

Any idea why? NaOH shouldn't have any effect on a stainless steel oxide coating at room temperature. Or is it simply that there are more sodium ions in solution with NaOH than with Na2CO3, and that accounts for the increase in conductivity (at least, I assume that's what you found. Or was the NaOH solution /less/ conductive?).

i wondered if there was any possibility of using carbon rods from the centre of old zinc chloride batteries as the electrodes? They will be inert enough, but would they be physically strong enough for the intended purpose?

--

Jeff

:

the

dal

a

hen

t the

vel

t to

ches

,

tube

ly clean a problem. A flat surface with both conductors on is best.

r conductivity measurement.

it isn't non-conducting. I had to try to get a conductivity meter working with stainless steel electrodes, and it more or less worked, but the conduc tivity I measured was different if the solution was made conductive with Na OH or Na2CO3.

The chromium oxide layer isn't very thick, and hydroxyl ions getting into i t presumably acted as charge carriers.

The hydroxly ion (-OH)is more mobile in water than the the carbonate ion so less NaOH than Na2CO3 was required to get the conductivity I was setting u p.

The concentrations were chosen to give me series of solutions of predictabl e conductivity, and I'd checked them out with a regular conductivity meter with standard platinum electrodes.

For high conductivity solutions the stainless steel electrodes gave me less conductivity than I should have been seeing, due to the resistance of the passivating layer of chromium oxide on the electrodes, and the fact the the error was lower, but still obvious with NaOH solutions rather than with Na

2CO3 solutions made what was going on even more obvious.

Platinum is a lot more mechanically stable, and a lot easier to make connec tions to.

My eventual preference would have been to use electrodes plated onto a prin ted circuit board - a couple of patches of a thin layer of platinum (severa l microns thick) wouldn't have been particularly expensive - but I came lat e into the project and a lot of time and money had been spent on a set-up t hat the mechanical engineers had liked.

--
Bill Sloman, Sydney

If the conductivity measurement is with an AC signal, shouldn't a thin oxide layer not matter, even if it was an insulator? What about the double layer, would using a 10kHz measurement signal take care of that? Platinized platinum is still preferred in such sensors for high accuracy, but what about for modest-accuracy?

--
 Thanks, 
    - Win

I have , bought at a car boot sale out of interest, a couple of packs of carbon-arc lamp graphite rods from a hundred years back maybe, perhaps I'll give them a go.

--
Monthly public talks on science topics, Hampshire , England

With relatively conductive aqueous solutions the resistance of the oxide la yer did matter, and it was lower with NaOH solutions than with Na2CO3 solut ions made up to be of the same conductivity.

Our standard conductivity meter ran at 10kHz. I'd worked out a scheme which gave out a frequency that increased with increasing conductivity, from abo ut 2kHz in tap water (300uS/m) to about 150kHz in 2% NaOH at 85C (about 300 mS/m). There was a perfectly predictable built-in non-linearity to keep the maximum frequency low enough for the circuit to work.

--
Bill Sloman, Sydney

But my question was about accuracy. If the goal is simply to determine if there's a significant salt-water intrusion, wouldn't a lower-accuracy sensor scheme be OK?

--
 Thanks, 
    - Win

I stuck two SS optical posts (1/2" diameter) into water and measured AC conductivity just fine... ~a few kHz IIRC, I sourced voltage and measured the current. The optical posts were nice, tapped holes top and bottom, one to hold in place and the other for electrical connection.

George H.

..

If the intent is to probe a sump occasionally, zinc would do fine (yeah, it'll rot away eventually, but a quarter pound of slug-stop zinc strip will last a lifetime). The carbon would be easy to wash off after use, but so would a lot of other options. And, the carbon is useful if you want to arc-melt something.

I'm partial to the PVC pipework 'fork'; stick the fork into the pond and freshwater makes a large gap - tine separation size- between electrodes, while saltwater, conductive, makes a two-thicknesses-of-tubing effective gap. A CMOS '555 oscillates it, and a multimeter reads out the frequency.

The important thing, is that the thing has to be at/near clean, and of constant geometry, so that the liquid composition is the sensed variable.

Depends on exactly what conductivity the OP was looking for.

Sea-water and sewage are a lot more conductive than tap water (which is in turn about two orders of magnitude more conductive than distilled water).

The problem that I see is that the conductivity of the oxide layer depends on the nature of the contaminants in the water that it is exposed to, which does make interpreting the observations more complicated.

--
Bill Sloman, Sydney

The great virtue of the stacked-toroid conductivity sensor is that it measures the current induced in the solution that threads the stacked toroids without introducing any electrode surfaces to make life more complicated.

Getting two toroids wound with a non-progressive winding on each of them is going to be a performance, but once you've done that you've avoided a whole lot of other potential problems.

--
Bill Sloman, Sydney

Same only more so. Titanium like aluminium is a very reactive metal with an incredibly thin self healing oxide barrier coat on the surface.

Silver foil might be OK in a wet chloride environment as a cheaper alternative to platinum. Failed bursting disks from pharma plants is a cheap source of such material if you know someone in the industry.

--
Regards, 
Martin Brown

Silver would certainly be close to ideal, but why not copper? It will slowly corrode, but might still last a few decades and it doesn't have the problem of an insulating oxide layer. It is also readily available and fairly cheap.

John

Copper corrodes in moving water with both oxygen and chloride ions readily available. Copper chloride is just about soluble enough to be a problem (hence ferric chloride PCB etch baths).

In the open air on roofs it quickly develops a fairly thick greenish blue carbonate surface coating that prevents any further corrosion.

You might get away with it and as you say copper pipe is cheap!

Silver chloride is not really soluble in brine at all and so survives and makes quite a good biologically inert conductive sensor.

--
Regards, 
Martin Brown

Silver/silver chloride electrodes are widely used for physiological measurement as they are reversible and give low noise.

John