Soluble Salt Contamination on Steel: Bresle Sampling and Limits

Soluble salt contamination means inorganic salts left on a steel surface that dissolve in water and conduct electricity as ions — chlorides above all, frequently together with sulphates and nitrates. It is the only one of the three pre-coating surface assessments (cleanliness ISO 8501-1, roughness ISO 8503, soluble salts ISO 8502-6/8502-9) that cannot be seen. Its difference from mill scale or rust lies in how it must be handled: abrasive blasting removes adherent material but can press salts into the valleys of the anchor pattern, or re-deposit them from the abrasive, the compressed air and the wash water. A visual Sa 2½ result therefore does not prove that the surface is chemically clean.

TL;DR: Salts left at the coating/steel interface are hygroscopic; they draw water in by osmosis and are a routine cause of osmotic blistering and early adhesion loss, while chloride ions damage the substrate separately by pitting. The assessment has two fixed steps: ISO 8502-6 describes how to extract the salts with a Bresle patch, ISO 8502-9 describes how to turn the conductivity reading into a comparable surface salt density, with the general industrial constant c = 5 kg·m⁻²·S⁻¹ (a seawater-type salt assumption). Results are reported as mg/m² or µg/cm² and the conversion factor between them is 10. Acceptance levels come from the project specification, not from the standard: NORSOK M-501 uses 20 mg/m² (as NaCl), the IMO PSPC (MSC.215(82)) uses 50 mg/m², ISO 21809-1:2011 uses 20 mg/m² for pipe coatings after blasting, and general industrial C1–C3 atmospheric work typically allows 80–100 mg/m². Once a surface fails there is only one closed-loop remedy: fresh-water rinse, forced drying, retest — and prime before flash rust forms.

Key data

• ISO 8502-6:2020 requires the water to stay in contact with the steel for at least 10 min; the 2006 edition specified no dwell time. SSPC Guide 15 allows as little as 90 s and instrument makers commonly recommend 2 min. Dwell time must be identical before results can be compared.
• ISO 8502-9 gives ρA = c·10²·V·Δγ/A, with c = 5 kg·m⁻²·S⁻¹ for ordinary industrial work. Worked example: Δγ = 20 µS/cm, V = 3 mL, A = 1250 mm² gives ρA = 24 mg/m², i.e. 2.4 µg/cm².
• The unit conversion factor is 10: mg/m² ÷ 10 = µg/cm². On the same blasted panel three inspectors reported 22 µS/cm, 28 mg/m² and 2.5 µg/cm², which normalise to 26.4, 28 and 25 mg/m² — one and the same contamination level.
• Extraction water must meet ISO 3696 grade 3 (conductivity <0.5 mS/m, that is 5 µS/cm); a blank of ≤5 µS/cm is acceptable, and every reading is blank-corrected and temperature-normalised to 25 °C.
• Patch cavities are sized A-0155 (155±2 mm²), A-0310, A-0625, A-1250 (1250±13 mm²), A-2500 (2500±25 mm²), with flexible sleeves at S-1000 (1000±10 mm²). Fill ratio is about 2.6×10⁻³ mL/mm² (±0.6×10⁻³) for adhesive patches and 10.0×10⁻³ mL/mm² for sleeves.
• Steel surface temperature during the test must be above +5 °C, recorded to 0.5 °C. A single extraction typically dissolves about 95% of the soluble contaminants; repeating the extraction approaches full recovery.
• Acceptance levels in use: NORSOK M-501 = 20 mg/m²; IMO PSPC = 50 mg/m²; ISO 21809-1:2011 after blasting = 20 mg/m²; Saudi Aramco critical immersion and linings = 20 mg/m², non-immersed atmospheric = 50 mg/m²; general industry C1–C3 = 80–100 mg/m².
• Published test work shows how wide a "safe" threshold can be: Baek et al., working across 0–660 mg/m² NaCl, settled on 115 mg/m² from cyclic corrosion testing, roughly six times the NORSOK value; Lee et al. recommended 100 mg/m²; for a two-coat epoxy at 200–300 µm dry film the literature range judged low-risk spans <10 to 164 mg/m². No single critical level is agreed.

Two ways salts destroy a coating

Osmotic blistering: driven by total concentration

Organic coatings behave as semi-permeable membranes. Soluble salts at the interface are hygroscopic; once dissolved they form a concentrated electrolyte that pulls external water along the concentration gradient and lifts the film into blisters. What matters here is the total concentration of dissolved particles, not which ion is present: even a mildly aggressive salt raises the blistering risk if there is enough of it. That is precisely why a conductivity method is a sound acceptance tool — it measures the total and is therefore conservative. Liquid inside a mature blister typically conducts several times better than the external medium, and once a blister is established it rarely stops growing on its own.

Ionic attack: chlorides and sulphates look different

Chloride ions are small and mobile; they penetrate the passive oxide layer on steel and initiate pitting and filiform corrosion under the film. Sulphates are commoner in industrial and road-side environments and take part in forming corrosion products that expand in volume and push the coating up from below. The two mechanisms usually run together, which is why a section through a rust blister typically shows both.

Why blasting does not take the salts away

Both published research and site experience point to the same conclusion: abrasive blast cleaning is not a reliable removal route for soluble salts. Salts concentrate at the bottom of pit cavities, in heat-affected zones next to welds, under mill scale and inside metal folds, and contaminated abrasive re-deposits them. A comparative trial on steel grit published in 2015 (Corrosion Protection Network, China) illustrates the effect: the same batch of panels averaged about 11 mg/m² before blasting, dropped to 9 mg/m² with a low-salt abrasive and rose to 27 mg/m² with a high-salt one — already above the 20 mg/m² of ISO 21809-1. Abrasive itself can be checked for soluble salts by measuring the conductivity of a water wash per ASTM D4940, which is worth keeping as a goods-in test.

How the ISO 8502 family is divided

ISO 8502, "Preparation of steel substrates before application of paints and related products — Tests for the assessment of surface cleanliness", is split by contaminant type. Decide what you need to measure first, then choose the method and the instrument.

The distinction matters: ISO 8502-6 only says how to collect the salts, ISO 8502-9 says how to make the conductivity comparable. Only the two together form a test result; a patch on its own, or a raw µS/cm reading, is not a verdict.

Bresle sampling: patch choice and sequence

Where to place the test

The standard advises sampling where metal loss or pitting exists and near welds, because that is where salts collect. If the measurement is taken before surface preparation, scrape the surface first until at least 50% of the test area shows bare metal. Spread the points over a large component: scatter between points on one area is normal, so read the distribution and the trend rather than a single figure.

The sequence

• Run a blank on the extraction water, record it and subtract it from the field reading to obtain Δγ; if the blank exceeds 5 µS/cm, flush the electrode or change the water.
• Choose the patch: A-0155/A-0310 for suspected local hot spots, A-0625/A-1250 for ordinary assessment, A-2500 or an S-1000 sleeve when a representative average is needed.
• Bond and de-air: press the adhesive ring flat, fill at about 2.6×10⁻³ mL/mm², inject roughly half first, withdraw the trapped air, then add the remainder.
• Keep contact and movement: insert the needle at about 30° so it does not touch the steel, hold contact for at least 10 min per ISO 8502-6:2020 and move the liquid at least three times by drawing it back and forth or by massaging the patch.
• Record temperature: the steel surface must be above +5 °C and read with a contact thermometer accurate to 0.5 °C.
• Convert: normalise the meter to 25 °C, take Δγ and apply ρA = c·10²·V·Δγ/A, or let a meter with a Bresle mode print mg/m² directly.
• Report the essentials: standard number, patch type and size, extraction volume, contact time, temperature, patch batch number and test date. Missing any of these makes the result unverifiable.

Three recurring mistakes

First, reporting µS/cm as if it were mg/m². Second, comparing points taken with different patch sizes or dwell times. Third, ignoring background contamination — the adhesive and the container can carry soluble material, so the standard requires a batch blank on a clean glass substrate, and where the blank is significant it must be subtracted.

Conversion rules and unit traps

With the common set V = 3 mL, A = 1250 mm² and c = 5 kg·m⁻²·S⁻¹ the formula simplifies to mg/m² ≈ 1.2 × µS/cm (ISO/TR 15235:2001 quotes about 1.18), which makes mental arithmetic on site possible:

The conversion is geometry-dependent. The same 20 µS/cm means something different if 25 mL was injected into an S-1000 sleeve (1000 mm²) as if 3 mL went into an A-1250 patch. A salt figure that does not state the extraction volume and the test area is not comparable with anything. Ion-specific results, such as chloride in ppm, cannot be derived from conductivity at all and need ISO 8502-2, -5 or -11.

Setting the limit: specification comparison and system matching

Why the standards refuse to fix one threshold

ISO 8502-6 and ISO 8502-9 deliberately set no pass level and leave it to the specification, because an acceptable salt load depends on at least four variables: the service environment, the coating system and its dry film thickness, the surface preparation method, and whether the structure sits in an osmosis-sensitive condition (immersion, under insulation, or under cathodic protection). The experimental record supports this. For the same class of epoxy the low-risk ceiling reported in the literature runs from below 10 mg/m² up to 164 mg/m²; Baek et al. arrived at 115 mg/m² after cyclic corrosion testing, about six times NORSOK; Axelsen and Knudsen, working between 5 and 300 mg/m² NaCl, found the effect on cathodic disbondment and on corrosion around the scribe limited, yet still recommended keeping salts as low as practicable. The conclusion: the threshold is a function of the design condition, not a universal constant, so quotations, acceptance certificates and claims must name the specification and the unit.

When to tighten the line

Three situations justify running at the 20 mg/m² level: immersion and lining service (potable water tanks, ballast tanks, vessel linings); structures where insulation and cathodic protection coexist; and heavy-duty systems designed for more than 15 years in service. Conversely, dry inland factory steelwork can sit at 80–100 mg/m² and still be sound, provided the salts are not coming from a recurring source such as a high-salt abrasive or saline cooling water carry-over.

After a failure: the rinse, dry, retest loop

Soluble salts cannot be ground out of sight; they have to be diluted and removed:

• Rinse with fresh or deionised water, then dry the surface forcibly and retest the same spot until the specification level is met. Keep the rinse-and-retest records with the inspection file.
• Prime before flash rust. A rinsed surface is highly reactive and flash rust appears quickly at high humidity or during night cooling. Most specifications cap the interval between preparation and priming, shortened further in damp or cold weather; once flash rust forms the cleanliness is void and the surface must be reworked, not simply overcoated.
• Shut off the sources of re-contamination. Re-check abrasive per ASTM D4940, look at condensate and oil carry-over in the compressed air, at sea-salt fall and rain staining during outdoor storage, and at secondary contamination from gloves and perspiration on a cleaned surface.
• Retest rather than look. Rinsing removes some salts but also brings deeper deposits to the surface, so a rise after washing is not unusual. Confirm with several points, not only the original location.

The three surface assessments belong together. Anchor depth and cleanliness logic is covered in industrial paint surface preparation: blast cleaning, roughness and dew point, visual grade comparison in surface preparation Sa 2½ and blast grades, the limits of painting over rust in rusty surface preparation: St2/St3 cleaning and conversion, the long-term consequence of interface salts in cathodic disbondment of coatings: mechanism, test standards and CP design, and how acceptance clauses are organised in anti-corrosion engineering specification for steelwork: design, preparation, acceptance.

Frequently asked questions

1. The panel is already blasted to Sa 2½ — why test for salts?

ISO 8501-1 grades are judged by eye and describe the area of remaining scale, rust and old coating. Soluble salts are invisible and are not reliably carried away by the abrasive; they sit in the anchor valleys and in pit bottoms and start drawing water after the coating is on, so failures surface a year or two into service. The three tests are not interchangeable.

2. The meter reads 20 µS/cm — is that a fail?

µS/cm is the conductivity of the solution, not the salt on the surface, so it cannot be compared with a limit directly. Convert with ISO 8502-9: ρA = c·10²·V·Δγ/A. With c = 5 kg·m⁻²·S⁻¹, V = 3 mL and A = 1250 mm² the reading is about 24 mg/m², which fails NORSOK's 20 mg/m² and passes IMO PSPC's 50 mg/m². Always establish the extraction volume and the patch area first.

3. Three inspectors got three different numbers on one area — which is right?

Five causes explain most of the spread: reporting units, patch size and fill volume, dwell time (90 s, 2 min or 10 min), whether the blank was subtracted, and whether the reading was normalised to 25 °C. Align those five and the scatter normally falls within ordinary measurement error. Comparison presupposes an identical method.

4. Does the conductivity method tell me whether it is chloride or sulphate?

No. Conductometry measures total solubles and assumes a seawater-type mixture. For specific ions use ISO 8502-2 (laboratory chloride), ISO 8502-5 (chloride by detection tube) or ISO 8502-11 (sulphate by turbidimetry); ISO 8502-12 titrates ferrous ions to judge rust activity. Near a coal-fired plant or chemical works the load may be sulphate or nitrate rather than chloride — the total still means something, but the attribution needs a separate method.

5. Factory-prepped components arrive on site — why test again?

Storage and transport keep loading the surface: de-icing salt aerosol travels far inland from coasts and roads, condensation builds up under tarpaulins, rain staining and sea spray add more, and handling leaves gloves and perspiration behind. Specifications therefore require measurement at both the shop and the site stage, with extra points at girth welds and secondary heat-affected zones.

6. Will a fresh-water rinse damage the steel?

The risk is not the water but the time it stays. Deionised water adds no salts and a short contact does not alter the substrate; the problem is that a water film lets the steel oxidise fast and flash rust form. Provide blow-off, drying heat or dehumidification and keep the interval between rinsing and priming inside the time allowed by the specification.

7. Are zinc-rich primers more sensitive to salts?

Zinc-rich systems need metallic contact between the zinc powder and the substrate to deliver sacrificial protection; a salt film and its corrosion products interfere with that contact and accelerate localised attack, which is why heavy-duty specifications normally hold such systems to the tighter level. The figure to apply is still the one written in the project specification — the coating type is not a reason to relax it.

References: ISO 8502-6:2020 (extraction of soluble contaminants, the Bresle method); ISO 8502-9 (field conductometric determination); ISO 8502-1/-2/-3/-4/-5/-11/-12; ISO/TR 15235:2001; NORSOK M-501:2022 "Surface protection and protective coating"; IMO MSC.215(82) PSPC; ISO 21809-1:2011; ISO 12944-9; SSPC Guide 15; ASTM D4940; C. A. Zimmermann et al., "The influence of soluble salts content on the performance of an epoxy coating system via accelerated corrosion tests", Matéria 25(2), 2020, DOI 10.1590/S1517-707620200002.1048; DeFelsko, "How to interpret soluble salt test results: understanding ISO 8502-6/9 (Bresle method)"; Chinese Corrosion Protection Network, 2015 study on soluble salts in steel abrasive.

Last updated: 2026-10-30 | Kexin New Materials (Guangdong) Co., Ltd.