Measuring Suspended Solids in Copper Electrolyte: Why a Filtrate Clarity Spec Needs a Test Method

The short answer

Suspended solids in a copper electrorefining electrolyte are normally measured in single milligrams per litre. Researchers at the Institute of Non-Ferrous Metals in Gliwice described a normal content as "usually not more than 5-6 mg/l", with anything above 20 mg/L treated as an alarm. That is a hard place to measure, for two reasons that have nothing to do with the filter.

US EPA Method 160.2, the long-standing gravimetric method for total suspended solids (TSS), states a practical range starting at 4 mg/L, and it warns that samples high in dissolved solids give a positive error. At 5 mg/L suspended, copper electrolyte carries, by our arithmetic, roughly 50,000 to 65,000 times more dissolved material than suspended material.

The standard optical shortcut fails in a different way. The published fix for coloured samples is to measure turbidity at wavelengths above 800 nm, and the copper aqua ion, measured in dilute solution, has its visible to infrared absorption peak at about 805 nm.

So a filtrate clarity figure in a filter specification, "5 mg/L" or "2 NTU", is only half a specification. The other half is the method that produced it.

How small the numbers are

Two independent published sources put the working range of copper electrolyte suspended solids in the single to low tens of mg/L. Our earlier article on suspended anode slimes reported that we had found no published mg/L figure; these two papers are the ones we have found since.

The Gliwice paper (Gajowski, Smieszek and Szota, published by TMS in 1989) gives the operating view: a normal content "in the order of several mg/l, usually not more than 5-6 mg/l, while contents exceeding 20 mg/l are referred to as an alarm state." The same paper names the usual suspects when the number rises: a crack in the filter cloth of a slime filter, turbulent electrolyte flow, anode passivation, aeration of the electrolyte, and poor cell servicing. It describes the suspended material as mainly very small particles of lead sulphate and antimony complexes, and links it to higher lead, silver, bismuth, tin, antimony and arsenic in the cathode.

A 1970 SME paper by Judson Brown gives the cathode-side view. It quotes a 1966 Chemical Engineering review of Swedish tests: at 1,000 A/m², satisfactory cathodes needed electrolyte flow past the electrodes of at least 3 to 4 metres per minute and slime content no higher than 20 to 30 mg/L, while at 500 A/m² the slime content could reach 50 mg/L without occlusion at the cathode. That is a quotation of a review of tests reported in 1966, and it should be read as an order of magnitude, not as a modern limit.

Read together, they agree on scale: a few to a few tens of milligrams in a litre of solution that also carries a few hundred grams of dissolved salts and acid.

Glencore Technology's published guidance on electrorefining lists the core measurement set for a refinery as electrolyte copper and acid concentration, temperature, cell voltage, and dissolved impurities by routine sampling and ICP analysis. Suspended solids are not on that list. Separately, the Gliwice group wrote in 1989 that the laboratory method returned results "with considerable time delay. Therefore, it is useless for the process control." Their answer was a continuous optical measurement in the flowing electrolyte.

For what the suspended fraction is made of and why it stays in suspension, see our article on suspended anode slimes in copper electrorefining. This article is about how the number is measured.

Why gravimetric TSS struggles at 5 mg/L in electrolyte

The reference gravimetric method filters a known volume through a glass fibre filter, washes it, dries it at 103 to 105 °C and weighs the residue. US EPA Method 160.2 is the long-standing version, and Standard Methods 2540 D follows the same principle. Three features matter for electrolyte.

The method is at its floor

EPA 160.2 states a practical range of 4 mg/L to 20,000 mg/L. A revision of the Chesapeake Bay Program's version of the method puts its floor at 2.5 mg/L, based on 2.5 mg of residue from a 1,000 mL maximum sample. The normal electrolyte figure from Gliwice sits right at that floor.

The sample volume follows from it. EPA 160.2 starts with 100 mL on a 47 mm filter and requires the volume to be increased if less than 1.0 mg of residue is captured. At 5 mg/L, 100 mL yields 0.5 mg, so by our arithmetic at least 200 mL is needed. Procedures based on Standard Methods 2540 D ask for at least 2.5 mg of residue, which at 5 mg/L means 500 mL. The method accepts a filter as dry when successive weighings change by less than 0.5 mg; on a 200 mL sample that tolerance alone is equivalent to 2.5 mg/L, half the value being measured. At the 20 mg/L alarm level the same 100 mL gives 2.0 mg, which is workable.

The dissolved load

EPA 160.2 is direct about this. Samples high in dissolved solids, "such as saline waters, brines and some wastes, may be subject to a positive interference," and the wash step is there to minimise it. The Chesapeake method repeats the warning in the same words.

Electrolyte sits at the heavy end of that warning. Using Glencore's typical figures and our own arithmetic, copper at 40 to 50 g/L is about 100 to 126 g/L of copper sulphate (anhydrous basis), and sulphuric acid runs at 160 to 200 g/L. That is at least 260 to 330 g/L of dissolved material, before nickel and the other dissolved impurities, against perhaps 5 mg/L suspended.

The consequence for the wash is easy to put in numbers. Each microlitre of electrolyte left in the filter after washing carries 0.10 to 0.13 mg of copper sulphate alone, before counting any acid. Four to five microlitres, a fraction of a drop, would match the method's entire 0.5 mg constant-weight criterion. The method specifies about 30 mL of wash water in three portions for a 47 mm filter, a procedure written for waters and wastes, not specifically for a solution at a few hundred grams per litre. If your laboratory has validated its wash on electrolyte, for example by running blanks of filtered electrolyte through the full procedure, that validation is part of the result and belongs in the specification.

The test filter decides what counts as suspended

TSS is defined by the filter, not by the particles. EPA 160.2 defines non-filterable residue as the solids "retained by a glass fiber filter", and notes that for glass fibre filters "the absolute pore size cannot be controlled or measured." The Chesapeake method permits 1.5 µm or 0.70 µm filters and says "pore size should be recorded". One university laboratory procedure we read specifies glass fibre filters of 2.0 µm and 1.0 µm nominal pore size.

The published work covered in our slimes article reports suspended slimes as mostly below 9 to 10 µm. A test filter rated between 0.7 and 2 µm therefore sits inside the size range of the material it is meant to count. Two laboratories using different filter grades on the same electrolyte can report different TSS figures, and both can be following the method correctly.

Why turbidity struggles, and why the standard fix points the wrong way

Turbidity is the obvious alternative: fast, and possible in line, which is what the Gliwice group built. It measures scattered light, reported in NTU, not mass. The scattered light depends on particle size and shape as well as on how much solid is present, so any conversion to mg/L is an empirical correlation for one suspension. We found no published NTU to mg/L correlation for copper electrorefining electrolyte.

White light: EPA 180.1

EPA Method 180.1 specifies a tungsten lamp at a colour temperature of 2200 to 3000 K and a detector with "a spectral peak response between 400 nm and 600 nm." It warns that "true color, that is the color of water which is due to dissolved substances that absorb light, will cause turbidities to be low," and separately that "finely divided air bubbles can cause high readings." Its applicable range is 0 to 40 NTU, and it provides for dilution above that.

Both warnings apply to electrolyte: it is strongly coloured, and a sample drawn from an aerated or turbulent point can read high for reasons that are not solids.

Near infrared: ISO 7027

ISO 7027 is the standard answer to colour. Its introduction states that measurements "can be affected by the presence of dissolved light-absorbing substances (substances imparting colour)" and that "such effects can be minimized, however, by performing measurements at wavelengths greater than 800 nm." Instrument suppliers present this as the method's main advantage. One describes "the reduction of color interference" as "the primary benefit of the ISO method over the EPA method."

The same introduction then names its exception: "Only carbon black and a blue colour, which can be found in certain polluted waters, slightly affects measurements of turbidity in this region of the spectrum."

Copper electrolyte is a blue solution. ISO's "slightly" was written for blue colour in polluted waters; copper electrolyte's absorbance in that region is of a different order. A 2015 study from Lawrence Livermore National Laboratory, published in Physical Chemistry Chemical Physics, puts the absorption maximum of the copper aqua ion, measured in dilute solution, at 1.54 eV, which is 805 nm, with an average band width at half height of about 0.66 eV across the copper complexes it tested. Wavelengths just above 800 nm therefore sit at the band's peak. Treating the band as symmetric (the paper reports that it skews, with a broader infrared edge), our arithmetic puts absorption above half its peak height from roughly 660 nm out to roughly 1,025 nm. ISO 7027's near infrared region sits inside copper's visible to infrared absorption band. In a sulphate electrolyte the band may shift and strengthen, since the same study found that other ligands around the copper ion raise its absorption; the dilute-solution figures are a guide, not a measurement on electrolyte.

The size of that absorption is large. A tabulated literature value for the molar absorptivity of the copper aqua ion is about 11 litres per mole per centimetre. At 40 to 50 g/L of copper, which is 0.63 to 0.79 mol/L, that gives an absorbance of roughly 7 to 9 per centimetre of light path at the band peak. Those are our figures, and the absorptivity comes from a single tabulated source. The conclusion does not depend on its precision: even at a tenth of that value the absorbance would still be about 0.7 to 0.9 per centimetre. For comparison, a white-light instrument built to EPA 180.1 may have up to 10 cm of total light path in the sample, so even short paths in copper's band lose a large fraction of the light.

What this means for an electrolyte reading

Our reading of the published data is this. EPA 180.1 places the detector's peak response at 400 to 600 nm, below the half-height edge of copper's band, though a tungsten source and a detector without a filter still respond to some longer wavelengths. The infrared method, chosen elsewhere to escape colour, measures inside copper's absorption band. Both are subject to the colour error EPA warns of, and we found no published comparison of how much each is affected on copper electrolyte. Ratio-mode instruments are designed to reduce the effect of colour; we found no published test of how well they compensate at electrolyte copper levels. Without such a test, the reading has to be treated as depending on the copper concentration as well as on the solids, and copper concentration moves through the operating range.

None of this makes turbidity useless on electrolyte. A trend read on one instrument at one sampling point is valuable. But an NTU figure is specific to the instrument's light source, the copper level and the site's own correlation, and cannot be carried from one plant to another or into a supplier's guarantee without those details.

What to put in the specification

A filtrate clarity line that another engineer can test needs the method written beside the number. For a copper electrolyte filter, we suggest stating:

1. The parameter. Gravimetric TSS in mg/L, or turbidity in NTU. They are not interchangeable without a site correlation.

2. For TSS: the test filter grade. 0.7 µm, 1.5 µm or another, recorded as the Chesapeake method asks.

3. For TSS: the sample volume and the residue it typically yields. At single-digit mg/L, the volume decides whether the result is above the method's floor.

4. For TSS: the wash procedure and how it was validated on electrolyte. A filtered-electrolyte blank run through the full procedure is the simplest check.

5. For turbidity: the instrument type and light source. White light to EPA 180.1 or near infrared to ISO 7027, ratio or non-ratio mode, and the copper concentration at which the correlation was made.

6. The sampling point and the sample temperature. Our slimes article covers why a settled sample and a circulating sample are different materials, and the Gliwice paper lists aeration among the causes of rising suspensions.

7. Units on a stated basis. "ppm" is often used for mg/L; the Swedish figures quoted by Brown write "50 mg. per liter (50 ppm)". At a specific gravity of about 1.2, 5 mg/L is about 4.2 mg/kg, a difference of a sixth. Our viscosity article covers why density keeps turning up in electrolyte arithmetic.

The same discipline applies upstream. Our article on suspended solids in Copperbelt PLS asks for TSS at the point the filter will see it, with the sampling point stated. For electrolyte, add the method.

What we could not establish

• A published NTU to mg/L correlation for copper electrolyte. We found none. A site correlation, made on its own electrolyte against a validated gravimetric method, is the only basis we know of.

• The wavelength of the Gliwice in-stream instrument and its full results. We read the paper's abstract and introduction only.

• Whether a cooled or diluted electrolyte sample gains or loses suspended solids before it is measured. EPA 180.1 provides for dilution above 40 NTU, and laboratory samples cool. We found no published study of either effect on copper electrolyte, so we have not estimated one.

• A modern, independently published cathode-quality limit in mg/L. The figures above are from 1989 and from Swedish tests reported in 1966. Refineries set their own.

• The full text of Standard Methods 2540 D. It is paywalled. We worked from EPA 160.2 and the Chesapeake Bay Program method, which share its basis.

Where Sharpenn fits

Sharpenn builds pressure filters for electrolyte clarification, running in leaf (body-feed) or precoat mode. When a clarity target comes with its method, sampling point and the electrolyte's chemistry, we can size and configure a filter against something measurable, and the acceptance test afterwards measures the same thing the specification meant. Our record: ~1,000 units installed worldwide across 460+ installations (all applications); 32 units in copper refining; capacities 5 to 350 m³/hr; repeat orders after 20 to 25 years. For the wider picture of the duty, see the guide to electrolyte filtration in copper refineries.

If you are specifying an electrolyte filter, send us the clarity target together with the method behind it, and we will tell you what we would build and how we would expect it to be tested.