conductivity can be influenced by factors other than alinity.Mineral content can have a large effect,so conductivity alone is not enough for accurate measurement.

Absolutely, so if Steven is measuring a sample from a lake or river or other unpurified source he'll need to take this into account.

Hi Steven,

As Guest pointed out, the apparent salinity of your sample also depends on other minerals and substances dissolved in the water. They, too, can affect conductivity and can give you a false result if it is only the salinity of the water you're interested in, regardless of other content.

If you know (by some other means) that the conductivity of your sample is due to salt and nothing else, here is a formula that relates Total Dissolved Salts (TDS) to the Electrical Conductivity (EC) of your sample. Note that the formula relates TDS to conductivity, not resistivity. Not to worry, because the two are reciprocal. If you know the resistance (Ohms) you can find the conductance (Siemens) using R = 1 / S. Conversely, S = 1 / R.

TDS (mg/L) = EC (µS/cm at 25^oC) x 0.6

Note that in field work measurements of EC are typically recorded in micro-Siemens per centimeter (μS/cm) as the unit of choice. One μS/cm is equivalent to one MΩ/cm. In older texts you'll see a unit called a mho (ohm spelled backward) as the unit of conductance. The preferred unit today is the Siemen. Mhos and Siemens are interchangable -- they mean the same thing -- but Siemens are prefered over mhos in contemporary literature.

TDS is recorded in milligrams of dissolved solid in one litre of water (mg/L). Parts per million (ppm) is equivalent to mg/L but it is not a favored unit. EC measures the charge carrying ability (ie conductance) of liquid in a measuring cell of specific dimensions. It is necessary to clearly define the units of both conductance and length when talking EC.

The table* below gives you an approximate idea of how useful water is at different saline concentrations. Good, potable water has an EC between 0 - 800 uS/cm, for example. This works out to ∞ - 1,250 Ω/cm. In contrast, seawater has an EC of around 50,000 μS, or about 20 ohms.

0 - 800 μS/cm (∞ - 1,250 Ω/cm)

• Good drinking water for humans (provided there is no organic pollution and not too much suspended clay material.
• Generally good for irrigation, though above 300 µS/cm, some care must be taken, particularly with overhead sprinklers which may cause leaf scorch on some salt sensitive plants.
• Suitable for all livestock.

800 - 2,500 μS/cm (1,250 - 400 Ω/cm)

• Can be consumed by humans although most would prefer water in the lower half of this range if available.
  
• When used for irrigation, requires special management including suitable soils, good drainage and consideration of salt tolerance of plants.
• Suitable for all livestock.

2,500 - 10,000 μS/cm (400 - 100 Ω/cm)

• Not recommended for human consumption, although water up to 3000 µS/cm could be drunk if nothing else was available.
• Not normally suitable for irrigation, though water up to 6000 µS/cm can be used on very salt tolerant crops with special management techniques.
• Over 6000 µS/cm, occasional emergency irrigation may be possible with care, or if sufficient low salinity water is available, this could be mixed with the high salinity water to obtain an acceptable supply.
• When used for drinking water by poultry and pigs, the salinity should be limited to about 6000 µS/cm. Most other stock can use water up to 10,000 µS/cm.
  
• Water over 4000 µS/cm can cause shell cracking in laying hens.
  High magnesium levels can cause stock health problems in this range. Analysis recommended.

Over 10,000 μS/cm (less than 100 Ω/cm)

• Not suitable for human consumption or irrigation.
• Not suitable for pigs, poultry or any lactating animals. Beef cattle can use water up to 17,000 µS/cm and adult dry sheep can tolerate 23,000 µS/cm. However it is possible that waters below these EC levels could contain unacceptable concentrations of particular ions. Detailed chemical analysis should therefore be considered before using high salinity water for stock.
• Water up to 50,000 µS/cm (the salinity of the sea), can be used to flush toilets provided corrosion in the cistern can be controlled.

Kind regards,

TV

_{^{* With special thanks to the Department of Primary Industries, The State of Victoria, Australia.}}

The answers are, with the information supplied, not much and yes.

Thanks for the comments, just to put a bit more perspective on what I am doing, as a final year university project I am looking at generating power from osmosis using sea water and fresh water. The salinity both sides of the membrane will be measured, starting from a know concentration, therefore I only need to measure the salinity relative to my starting point.

Very useful point about the voltage, I will certainly now use sub 1.2V.

Thanks for the pointer on the electrode material, I had been pointed in the direction of platinum or graphite, but I will certainly check out stainless steel as well.

I think I will still need to use ac to avoid a net movement of ions within the solution. If you put the two probes of an DMM into a liquid solution, measuring resistance it does not give a stable answere. Which brings me back to my original question about the frequency of the ac supply, I understand that there is an optimal frequency to test salinity but as yet cannot determine what it is.

Thanks

Stephen

I understand your concern about the net movement of ions, but it is the net movement of electrons that will be your primary concern. Electron mobility far exceeds that of ions, making electrons the primary charge carrier. (This is true also of gaseous plasmas such as the Earth's ionosphere, magnetosphere, interstellar space, interiors of stars, and in neon signs.) Plasma physics are complex to say the least, because there are many additional factors to consider, such as charge-self-shielding, Debye length, electron and ion resonance, magnetic field binding, and so forth. It's a real can of worms.

At any rate, the ideal frequency will be a function of electrode spacing, electron mobility and the mean-free-path of electron movement at various concentrations. Also look up Debye Length. I'll see what more I can find out for you at my end. I haven't worked in plasma physics for more than 14 years now, but it's kinda like riding a bicycle. You never completely forget!

Platinum is the preferred electrode material by far. I didn't suggest it for obvious reasons of expense. Graphite is fine as an electrode material so long as it doesn't spall and contaminate the sample with particulates.

As you are taking measurements on both sides of a membrane -- through which electrons easily travel -- you need to keep in mind that current does not flow just between the electrodes directly, but from beside and behind them as well. This can complicate your salinity measurements considerably if you do not take steps to confine the electric field associated with each pair of electrodes.

One way to do this is to construct an immersible test cell of known electrode and cell dimensions. Surround this test cell on all sides with a perforated, grounded Faraday cage -- grounded in the sense that it is at the same average potential as the surrounding seawater. You don't want it acting like a third electrode. The fewer holes in the cage, the better, but not too few such that the sample stagnates with respect to the surrounding water.

If you go this approach, do not test the salinity on both sides of the membrane simultaneously. Do not! Alternate your measurements back and forth so that any stray electric fields set up by one electrode pair (shielded or not) does not influence the other pair's measurement. If your sensors are automated, you can switch between them, first the one, then the other, back and forth, several times a second. First and foremost, though, keep in mind that your membrane is nearly completely 'transparent' to electron flow. Not ion flow -- electron flow.

The buggers are almost as slippery as neutrinos!

A few more notes: merely placing your DMM probes in the water at random locations will give you inconsistent results even for homogenous concentrations because the test-container geometry affects the results, which will be different at different locations even if the probe spacing remains constant.

You can build a test cell that is its own Faraday cage, and it's very simple: The test 'cell' consists of an outer, open-ended metal cylinder which remains at ground potential. This is the first electrode.

Inside the cyclinder is a wire of slightly shorter length which runs down the center, forming a coaxial electrode pair. This assembly is partially submerged in the water at a constant depth, or completely submerged provided no other potential-carrying parts are exposed to the water. The cylinder should be oriented vertically so that no air bubbles remain trapped inside. Cap the lower end of the cylinder with metal screening or perforated metal to confine the electric field inside the cylinder whilst admitting the water. If the cylinder is to be completely submerged, cap the upper end with screening as well, leaving a hole for the center-electrode connection. I would use teflon-insulated wire to connect both electrodes to the outside world, as teflon won't respond to any organic pollutants in the water. The wire's conductor should be made from the same material as the electrodes (unless they are made of graphite) to prevent galvanic corrosion where they connect to the electrodes.

This construction eliminates the need for a separate Faraday cage whilst defining your test-cell's dimensions, and it is easy to build.