Mould release agents and formwork oils

How a release film works

Demoulding is a fracture process, and a release agent decides where that fracture runs. It does not grease, and lubrication is not its primary job; it places a defined weak plane between two surfaces that would otherwise adhere. Everything on this page about application rates, defect patterns and inspection intervals follows from that single idea.

Adhesion, cohesion and the idea of a sacrificial layer

Three forces act against one another between the moulded part and the mould wall: the adhesion of the moulding material to the mould wall, the cohesion of the moulding material within itself, and the strength of the release layer in between. Without a release agent, chance decides where the fracture runs. Where adhesion is high, the moulding material tears within itself and part of it stays behind in the mould. That is what the plant sees as sticking, as edge break-out, or as build-up on the mould wall.

The release agent places that weak plane deliberately. It is chosen so that the release layer is the weakest level in the system: it fails first, the fracture runs along it, and both surfaces stay intact.

Sacrificial or semi-permanent: the difference sets the application cycle

At this point two design principles part company, and the distinction decides half of your shop-floor organisation.

Sacrificial systems are consumed on every cycle and have to be rebuilt on every cycle. They are the normal case in concrete, ceramic and brick production and are the focus of this page. For them, the sentence that recurs throughout applies: a release agent that leaves something behind is not working economically, it is building up a deposit.

Semi-permanent and reactive release systems behave differently. They build a more durable film on the mould surface that carries through several demouldings and is renewed by plan only after a defined number of cycles. There, a residual film is not a defect but an operating requirement.

One practical consequence follows. Application cycle, permissible residual film, and the question of whether a mould may be cleaned in between at all are product-specific. They are stated in the technical data sheet of the individual product and cannot be derived from a general rule. If you run a semi-permanent system, the consumption calculation further down has to be set up differently as well: grams per square metre and application interval, not grams per square metre and cycle.

Wetting, film formation, and why thickness does not help

For the release layer to be closed, the liquid has to wet the mould wall: on impact it must spread rather than draw together into droplets. Whether it does is not decided by any single quantity but by the balance of the three interfacial tensions involved (solid/vapour, solid/liquid and liquid/vapour). That balance becomes visible as the contact angle of the droplet on the mould wall: a small contact angle means good wetting and a spread, closed film; a large one means the droplet stays where it landed and leaves gaps behind.

The practically important term is the middle one. The interface between mould wall and liquid changes with every contamination: release agent deposit, cement laitance, grease residues from maintenance. On clean, bright steel the same liquid wets well; on a fouled mould it no longer does so reliably. Mould cleaning is therefore not cosmetic; it is the precondition for the release agent forming a closed film at all.

The film thickness required is remarkably small. Deutsche Bauchemie, the German construction chemicals association, puts it for concrete at a theoretical one thousandth of a millimetre, combined with a closed, even application on the formwork, roughly one micrometre. Everything above that is simply lying there without contributing to release.

And it does not lie there without consequence. An oil film that is too thick is sticky enough that air bubbles rising during compaction cling to the formwork wall instead of escaping; they set as blowholes in the visible face. If the agent is applied too shortly before concreting, the solvent evaporates only after placement, and the resulting gas pressure drives additional blowholes to the concrete surface. Aqueous emulsions need flash-off time for the same reason: the oil film only closes once the water has evaporated.

Applying more does not release better. In a sacrificial system, release performance sits in a film of roughly one micrometre; everything beyond that ends up in the surface, in the extracted air, or in rework.

Physical and chemical release action

Every release agent works physically: the hydrophobic oil film keeps the mixing water of the concrete, or the moist or dry ceramic body, away from the mould material and prevents direct contact. In its own taxonomy, Deutsche Bauchemie describes the action of concrete release agents expressly as primarily physical separation and classifies products by raw material base and viscosity rather than by a mechanism.

A chemical contribution is added as soon as the product contains fatty acid esters. Fresh concrete is strongly alkaline; Deutsche Bauchemie states a pH of 12.9 for the pore solution. Esters are saponified in that environment. Note that alkaline ester cleavage does not yield the free fatty acid: it yields the corresponding carboxylate salt and the alcohol directly, and the carboxylate then reacts with the calcium ions of the fresh concrete to form water-insoluble calcium soaps. Deutsche Bauchemie describes the outcome as the fatty acid component reacting with the alkaline constituents of the fresh concrete at the concrete surface to form water-insoluble metal soaps. These lime soaps form a defined, water-insoluble intended fracture plane between concrete and formwork; the alcohol is bound into the cement matrix.

This reaction explains two things. It is why a reactive release agent for concrete needs a considerably lower application rate than a purely physically acting oil. And it is why overdosing produces a defect pattern of its own here: if too much oil is saponified, so many calcium soaps precipitate that the surface zone turns powdery. That is dusting, and it interferes with any subsequent coating.

In ceramics this reaction partner is missing. Dry-pressed granulate or plastic clay body is not alkaline enough to saponify oil to any meaningful extent. There, release performance is predominantly physical: film separation plus friction reduction at the tool wall. Anyone transferring experience from the concrete plant to the ceramic press without checking is transferring only half the mechanism.

Internal or external: two fundamentally different routes

Ceramics offers a route that concrete does not. Instead of treating the mould (an external release agent), the lubricant and release agent can be blended into the granulate or the body. During pressing it migrates to the interface and acts where it is needed.

The operational difference is considerable. An external sacrificial agent has to be reapplied on every cycle and therefore constrains cycle time; with external semi-permanent systems, the longer application interval described above applies instead. An internal agent is already in the material and removes the application step altogether. In exchange it goes entirely into the component and has to come out again during drying and firing. External die wall lubrication and internal blending have long been described in process engineering as two independent principles. Choosing between them is a process decision, not a product selection.

Behaviour under temperature and pressure

The release layer does not work at room temperature. Concrete sleeper production uses accelerated hardening; ceramic and brick presses generate frictional heat. Three effects set an upper limit on any product.

• Viscosity drop. Oil thins as temperature rises. The film can run off vertical mould walls before the moulding material is placed. On upright formwork and heated moulds this is the most common cause of gaps in the upper third of the mould.
• Evaporation. Volatile fractions leave the film. In some products this is expressly intended, because only the effective residue then remains. If the dwell time between application and filling is too long, however, too little is left.
• Thermal decomposition. Above the stability limit of the base oil, cracking products form and remain on the mould as a sticky, dark deposit. That deposit is the beginning of every mould build-up and the reason why cleaning interval and temperature control belong together.

Under pressure, the film can additionally be displaced out of the contact zone. In dry pressing of ceramic granulates this is the normal case: there is no longer hydrodynamic lubrication but boundary friction. What is then effective is no longer film thickness but the ability of the molecules to adhere to the tool surface and form a shearable layer. This is why applying more does not help when pressing; only applying the right thing does.

The mould material has a say

The same film behaves differently on different mould materials, and so markedly that the material question precedes the product question.

• Bright steel is non-absorbent, wets well and is prone to corrosion. Non-absorbent formwork promotes blowholes, mottling and colour variation, because air and surplus water are not carried out of the surface zone. Release agents for steel moulds therefore frequently carry corrosion inhibitors.
• Coated steel and plastic formwork facing have low surface energy; wetting deteriorates and the film tends to dewet. Less material, distributed more evenly, matters more here than it does on steel.
• Timber and formwork panels absorb. Because of the microscopically larger surface they need more release agent for wetting, and the adhesion of the concrete is higher, with edge break-out as a consequence. Conversely, absorbent formwork facing gives faces with fewer blowholes and more even colour, because it takes up air and water from the surface zone.
• Aluminium is sensitive to alkaline attack. Fresh concrete attacks the oxide layer wherever the release film has gaps.
• Rubber and elastomers (dies, isostatic pressing moulds, structural form liners) can swell. Compatibility here is a materials question, not a release question.
• Plaster of Paris in ceramics is the special case: the plaster mould draws the water out of the slip and thereby creates the parting plane itself. The precondition for that is precisely the wetting of the mould material; a hydrophobic release film would cancel the mechanism.

Sources: Deutsche Bauchemie, “Concrete release agents: information for the user”, 2nd edition April 2020 (film thickness, formwork facing, blowholes, classification by raw material base and viscosity) [3]; Deutsche Bauchemie, status report “Concrete release agents and the environment”, 4th edition June 2015 (saponification, pH 12.9 of the pore solution, metal soaps) [4]; Holcim, “Guide to fair-faced concrete: site”, 2023 (absorbency, gas pressure from solvent evaporation) [29]; Verband der Keramischen Industrie, “Technical ceramics in practice”, 2000 (slip casting in plaster moulds, wetting) [18]; US patent 5,682,591, 1997 (external die wall lubrication as an independent process principle) [32]. The description of wetting via the balance of the three interfacial tensions and the contact angle, and the formulation of alkaline ester cleavage as producing carboxylate and alcohol, are standard physical chemistry and are stated here in the technically correct form, which is more precise than the association text.

In production: ceramics, concrete, brick

Three industries, three tasks. In ceramics it is about die wall friction and ejection force; in concrete about surface quality and application rate; in brickmaking about the compromise between clean demoulding and what the body will still tolerate in the kiln. Anyone who lumps these three tasks under one word will inevitably choose wrongly somewhere.

Ceramics: dry pressing

Spray-dried, free-flowing granulate is compacted uniaxially in a die. The low moisture content of the pressing body is characteristic: the German ceramic industry association gives below 3 % for dry pressing, against 10 to 15 % for wet pressing, 15 to 25 % for plastic forming and 25 to 40 % for casting. That residual moisture itself forms part of the sliding layer: too little of it and the layer is incomplete; too much and the incompressible fraction works against densification.

The core problem is die wall friction. It dissipates pressing pressure over the height of the part, so density in the green compact is not uniform; in double-acting pressing a zone of lower densification forms in the middle, the so-called neutral zone. Comparison with cold isostatic pressing shows the cause cleanly: pressed axially, densification is high but varies in the pressing direction; pressed isostatically it is high and almost uniform. We deliberately quote no pressing pressures here. The figures available to us could not be traced to a sound process-engineering primary source, and an order of magnitude you cannot substantiate is of no use on the shop floor.

The second quantity is ejection force. It rises with die wall friction and loads ejector and die. That makes it the most useful early-warning quantity a pressing shop has: it rises before the first green compact breaks. If you do not record it, you learn about the problem through scrap.

Pressing aids act at both points. Sliding aids reduce external and internal friction, improve pressure transmission during pressing and make ejection of the compact easier. Internal agents are blended into the granulate and reduce particle-to-particle and wall friction at the same time; external agents are applied to die and punch and act only at the tool wall. For scale: for fine powders granulated before dry pressing, a total of 1 to 5 % by mass of organic additives is documented, and that covers binders, plasticisers and lubricants together. We are not aware of a sound separate range for the lubricant alone; if you are given such a figure, ask where it comes from.

Ceramics: extrusion, isostatic pressing, slip casting

In extrusion of plastic bodies, wall slip at the die decides extrudate quality. The paste does not flow in shear layers but as a largely rigid plug surrounded by a thin lubricating film at the wall; the wall shear stress is often an order of magnitude below the yield stress of the paste itself. That explains why a wall lubricant has a disproportionately strong effect here: it changes exactly the layer that carries the entire throughput.

In isostatic pressing the situation is other than one expects. Release at the elastomer mould is predominantly a material property: the mould material is chosen so that it reacts neither with the powder nor with the pressure medium and parts readily from the compact; the patent literature names natural rubber, neoprene, PVC, butyl, silicone and, preferentially, urethane. No separate release agent is presupposed there. If you have demoulding trouble in isostatic pressing, check the mould material first rather than looking for a release agent.

In slip casting into plaster moulds the mould separates by itself, and a release agent would be counterproductive. The precondition for cast formation is that the suspending liquid wets the mould material; the capillary forces of the plaster mould draw the water out of the slip. A hydrophobic film would suppress exactly that wetting. Drying then takes care of demoulding: the cast shrinks away from the mould wall inside the plaster mould and can be removed. Release agents only come into play where non-absorbent mould materials are used, for instance plastic instead of plaster in pressure casting. Then the suction is missing, and the parting plane has to be applied.

Concrete: cast in place, precast, fair-faced

On site, solvent-free products dominate; Deutsche Bauchemie attributes the largest market significance to them in cast-in-place concrete. Application is by pressurised hand sprayers of 5 to 10 litres at 4 to 6 bar with flat-jet nozzles, by non-pressurised hand sprayers of 1 to 10 litres, or by high-pressure airless equipment between 60 and 250 bar. Excess is taken off with a rubber squeegee or a cloth, not as a concession to sloppiness, but as a planned work step.

In precast plants, solvent-containing products and emulsions are more common; emulsions there have reached a market share above 10 % according to Deutsche Bauchemie. Application runs through stationary spray systems at 4 to 8 bar with flat-jet or oscillating nozzles; for economical application, rotary atomisers are named.

For fair-faced concrete the governing order of magnitude is a single number, and it is worth the argument with the plant: for fair-faced concrete of classes SB 3 and SB 4 on non-absorbent formwork, Deutsche Bauchemie, citing the German concrete association DBV, gives an application rate of 10 g/m². That is a very fine mist, considerably less than is customary in practice. Many plants apply a multiple of it without ever having measured, and then look for the cause of blowholes and discolouration in the product.

Fair-faced concrete, incidentally, is not a product property but a result agreed between the parties to the construction contract. Which fair-faced concrete class is achieved depends on the interplay of formwork facing, concrete mix, compaction, curing and application method; the release agent is one factor among several. Classes SB 1 to SB 4 and the associated formwork facing classes are set out in the German DBV/VDZ code of practice “Sichtbeton” (June 2015), whose Annex E deals expressly with release agents. If you owe a fair-faced concrete surface, agree the class and a trial area before execution begins, not afterwards. Readers outside Germany should note that this classification is a German convention; the corresponding agreement in your own market will follow different documents.

Concrete: immediate and late demoulding in sleeper production

Concrete sleeper production sets two opposing requirements side by side. In immediate demoulding, the mould is struck directly after compaction and is available again for the next production cycle; the fresh concrete has to reach a high green strength for that. In late demoulding, the concrete hardens inside the mould, and concrete and prestressing steel stay there until the required strength is reached. We quote no specific dwell times here: they could not be traced to a sound, permanently accessible source, and in any case they depend on plant, mix design and temperature control.

For the release film these are two different load profiles. Immediate demoulding calls for a parting plane that carries straight away and leaves the mould clean for the next circulation. Late demoulding calls for a film that is still present after a long dwell at elevated temperature, that is, one that has not run off, evaporated or decomposed. A product that does one well is not automatically suited to the other.

Brick and clay products

Roofing and facing bricks are formed from plastic body; pressed roof tiles are formed between upper and lower die on a revolver press. It is worth noting that the industry solved the release problem historically without chemistry and in part still does: in hand-moulded brick, sand is the release agent and produces the characteristic rough texture; in water-struck brick it is water. If you want to understand what a release agent has to achieve at its core, you can read it off here: all it has to do is interrupt direct contact.

The mould material has a say. Plaster moulds have good surface properties and carry water away but suffer heavy abrasion; steel moulds are dimensionally stable but show poorer release behaviour. The release agent works in exactly that gap, and the task is twofold: clean demoulding of the soft compact without edge distortion, and protection of the mould tooling, whose service life co-determines manufacturing cost.

The upper limit on application rate is set here by the product itself, and from two directions. In the green state, oil that gets between the clay particles sliding against one another during pressing disturbs the bond within the body; a 1926 patent specification describes this mechanism expressly for oiled metal moulds and concludes that a properly frost-resistant roof tile cannot be obtained that way. The source is old, the mechanism is not: overdosed release agent does not stay at the surface, it migrates under pressure into the surface zone of the moulded part.

In the kiln, organic carbon has to burn out before the surface sinters. Where that fails, reducing conditions prevail inside the body: the iron there is no longer present as iron(III) but is reduced to iron(II). This conversion under reducing firing conditions is the actual mechanism of black coring; low-melting iron silicate phases may additionally form and reinforce the effect. The result is a dark core that degrades appearance and mechanical properties. The link to organic load is also supported experimentally: in tile bodies, a raised organic carbon content produced pronounced black coring, which largely disappeared once that content was reduced. What the study carries and what it does not both need saying: it substantiates the mechanism, but not that the release agent is what supplies that load; the carbon investigated came from the raw material. No permissible release agent quantity can be derived from it. What remains for the plant is the qualitative finding, and it is enough: excess release agent increases the organic load that has to burn out in the core.

The kiln exhaust takes up the rest: brick flue gases are laden with organic substances and are treated by regenerative thermal oxidation, at around 820 °C combustion chamber temperature, to water and carbon dioxide. Whatever release agent is applied in excess therefore ends up in the energy balance of that oxidation. The permitting side of these emissions depends on site and installation and has to be assessed locally.

What in this category is not a mould release agent

For completeness, and because these sit side by side in the catalogue: not every product that prevents sticking is a mould release agent. A bitumen release agent protects steel surfaces on truck bodies, asphalt pavers and rollers against adhering hot asphalt: same principle, different process, no moulded part. A saw blade release agent for industrial woodworking prevents resin, pitch and wood dust adhering to the tooth crown; it is a tool release agent, not a mould release agent. And a release agent for demoulding wood and lightweight panels, sprayed into the mould and then distributed with a brush, is a good example of a two-stage application in which the brushing is what produces the evenness that spraying alone does not achieve. Everything on this page about film thickness, flash-off time and application rate applies to mould release. It cannot be transferred to those three applications without checking.

Sources: Verband der Keramischen Industrie, “Technical ceramics in practice”, 2000 (moisture content by forming process, neutral zone, sliding aids, slip casting) [18]; “Dry Pressing”, ScienceDirect Topics (1 to 5 % by mass organic additives in total) [19]; Benbow & Bridgwater, and Advances in Applied Ceramics 108(6), 2009 (plug flow, wall shear stress) [21]; US patent 5,490,969 A, 1996 (elastomer mould materials) [20]; Deutsche Bauchemie, information for the user, April 2020 (equipment, pressures, 10 g/m² for SB 3/SB 4) [3]; Deutsche Bauchemie, status report, June 2015 (market shares) [4]; DBV/VDZ code of practice “Sichtbeton”, June 2015 (fair-faced concrete classes, Annex E) [2]; German utility model DE 20 2010 001 525 U1, 2010, and patent specification DE 198 36 320 C2, 2002 (immediate and late demoulding) [28]; German Brick and Tile Industry Association [22]; Wienerberger and Nelissen (hand-moulded and water-struck brick) [23]; patent AT 518323 B1, 2017 (plaster versus steel moulds) [24]; patent specification DE 432343 C, 1926, historical [25]; Taşkıran et al., Physicochemical Problems of Mineral Processing 61(3), 2025 (qualitative relationship between organic load and black coring) [26]; Ziegelwerk Bellenberg (regenerative thermal oxidation) [27]; Voitländer-International product catalogue, as at 9 August 2026 [30].

What the plant gets out of it

The benefit of a release agent rarely shows up where it is purchased. Price per litre is the smallest of the five quantities in which it pays off, and the only one that lies on the table in purchasing. The other four surface in production, in maintenance and in rework.

Scrap and rework

On fair-faced concrete, blowholes, voids, staining and dusting are not a question of looks but an acceptance risk. Reworking an SB 3 or SB 4 surface means paying for cosmetics, scaffold standing time and schedule delay: three items that appear in no materials calculation. In ceramics the same item shows up as green compact breakage after ejection: scrap that arises only after forming and therefore carries all preceding work with it.

Cycle time

With immediate demoulding the mould is back in circulation directly after compaction. A demoulding disturbance then costs not one component but one mould circulation. What that is worth in a given case depends on the plant and cannot be quantified in general; if you want to know, set one disturbed circulation against the number of circulations per shift.

Mould and tool life

In ceramics and precast concrete, moulds are capital goods with lead times. Die wall friction, ejection forces and deposit formation are the wear drivers. A set of moulds that lasts one campaign longer beats any saving on price per litre. Ejection force is the quantity in which that development can be read early.

Consumption itself

Here the calculation is the most direct and the easiest to verify. Between the 10 g/m² quoted for fair-faced concrete classes SB 3 and SB 4 on non-absorbent formwork and a generous manual application there is, in practice, a factor that most plants have never measured. The measurement is trivial: consumption per shift divided by the mould area treated. Whoever tracks that one figure holds the lever, and in both directions, because under-application costs too, in sticking and in additional cleaning. With semi-permanent systems the same calculation is referred to the application interval rather than to the individual cycle.

Cleaning effort

What builds up has to come off again. Deutsche Bauchemie recommends cleaning formwork immediately after every strike, because the effort then stays small. The longer cement slurry and release agent residues remain on the mould, the sooner cleaning turns into mechanical work on the formwork facing, and that costs facing life. Cleaning interval and release agent therefore belong in the same calculation, not in two separate cost centres.

Sources: Deutsche Bauchemie, information for the user, 2nd edition April 2020 (10 g/m², cleaning immediately after striking) [3]; Verband der Keramischen Industrie, 2000 (ejection force, die wall friction) [18]; German utility model DE 20 2010 001 525 U1, 2010, and patent specification DE 39 31 201 C1, 1990 (mould circulation in immediate demoulding) [28].

Risks and defect patterns

The risks fall into two groups that are rarely considered together on the shop floor: defect patterns on the component, and hazards to the people applying the agent. The most striking finding in this section concerns the second group. It is below under “Which limit value actually applies”, and it is handled wrongly across the industry with some regularity.

Defect scheme: over-application, under-application, build-up

Almost every defect pattern coming from the release agent path can be assigned to one of three directions, and that assignment decides whether you start at the application, at the cleaning interval or at temperature control. The table below is meant to be printed and pinned up at the press or the formwork bay: appearance, direction, likely cause, and the check you start with. The causes given are those that come from the release agent path; rarely the only ones, but the ones you can eliminate without changing formulation.

Skin exposure

Skin contact is the underestimated everyday risk of this product group. Release agents dissolve the skin’s natural oils; repeated contact can lead to irritation and inflammation, and the alkaline reaction of cement and concrete on skin contact reinforces the effect. BG BAU, the German statutory accident insurance institution for the building trade, puts it briefly in its data sheet C 321: skin contact leads to irritation and inflammation.

The governing rule in Germany is TRGS 401 “Hazards from skin contact”, edition October 2022, last amended 19 September 2024. It requires the skin hazard to be determined and assessed and the protective measures to be derived from that: glove selection including breakthrough time, a skin protection plan, and replenishing skin care. Data sheet C 321 names, for concrete release agents specifically, protective gloves of nitrile or butyl rubber together with safety glasses and face protection. Which gloves are suitable for a given product follows from that product’s safety data sheet, not from a general recommendation.

Spray mist and aerosols

Two statements belong side by side here, pointing in different directions, and both belong on the table. Deutsche Bauchemie reports that proper spraying with a hand sprayer produces no appreciable proportion of respirable particles below 5 µm and that the fraction between 5 and 10 µm lies well below 1 %. BG BAU nonetheless states that inhaling concrete release agents can lead to damage to health, and requires an extension tube when spraying, in order to limit the inhalation of spray mist.

The two are compatible. The particle size distribution is favourable, exposure through vapour nevertheless remains relevant, and automated high-pressure systems produce a finer spectrum than a hand sprayer. One caveat belongs with this, and we state it expressly: the measurement quoted refers to the hand sprayer. For stationary high-pressure and airless systems we are not aware of a corresponding measurement. Quoting the favourable figure alone turns a conditional measurement into a general all-clear.

Which limit value actually applies, and why the obvious answer is wrong

For hydrocarbon mixtures, the German TRGS 900 sets group limit values under its clause 2.9 following the RCP approach: 700 mg/m³ for C6 to C8 aliphatics, 300 mg/m³ for C9 to C14 aliphatics and 50 mg/m³ for C9 to C14 aromatics. It is tempting to apply these values to release agents and formwork oils. That is precisely the error.

TRGS 900 clause 2.9 paragraph 2 expressly excludes complex hydrocarbon-containing mixtures such as lubricating oils as soon as they contain olefinic fractions, more than 1 % by weight of non-hydrocarbon additives, or chain lengths above C14. At least one of these features typically applies to an additivated formwork oil; the RCP group values are then not the right yardstick.

What applies instead cannot be answered generically for “release agents”, and this is where the second half of the error is made. Occupational exposure limits are substance-specific. Which entry has to be used follows from the components of the mixture actually in use and from its safety data sheet, not from the product category. Where highly refined mineral oil is the component governing the assessment, that leads to the list entry “mineral oils (petroleum), highly refined” at 5 mg/m³, carrying remark 11: the value applies as the sum of vapour and aerosols. For a mixture of different composition a different entry may govern, and a mineral-oil-free product does not fall under it at all.

The group limit of 300 mg/m³ for C9 to C14 aliphatics is, as a rule, the wrong yardstick for an additivated formwork oil. Which occupational exposure limit applies instead has to be determined substance by substance, from the components of the mixture in use and its safety data sheet. Where highly refined mineral oil is the governing component, that means 5 mg/m³ as the sum of vapour and aerosol, lower by a factor of 60.

Two additions belong with this. First, the general dust limits apply alongside: 1.25 mg/m³ for the respirable fraction and 10 mg/m³ for the inhalable fraction. Second, where no occupational exposure limit exists for a given product, that is not a free pass but triggers TRGS 500 clause 5.1(9) no. 3: the state of the art then has to be met and demonstrated.

The hierarchy of measures is in the same technical rule: substitution is the most effective protective measure, followed by closed systems, only then extraction, and personal protective equipment last. Translated into everyday release agent practice: first examine the product and the application method (rolling and brushing generate no mist), then enclose, then extract, and only then discuss respiratory protection. Data sheet C 321 names, for spray application, A2-P2 combination filters with solvent-containing products and P2 or FFP2 particle filters with solvent-free products. Do not spray against the wind and do not spray overhead.

Fire and explosion protection, storage, aspiration

Solvent-containing products are ignitable. Data sheet C 321 requires ignition sources to be kept away, containers to be stored closed, and the quantity held at the workplace to be limited to one shift’s requirement. The German TRGS 800 draws the conceptual line: it distinguishes flammable liquids carrying H224, H225 or H226 from combustible liquids with a flash point above 60 °C.

For storage, TRGS 510 translates the classification into storage classes: storage class 3 is “flammable liquid”; storage class 10 covers combustible liquids that cannot be assigned to storage classes 1 to 8. One point is frequently short-circuited in practice: which storage class a product receives follows from the classification of the mixture as a whole and from every further hazard property it carries. The storage class cannot be derived from a single figure such as the flash point, nor does a flash point above 60 °C automatically mean that a flammability label is dispensed with. The safety data sheet governs.

One point hits precisely the most convenient products. With the move to the CLP Regulation, the governing viscosity threshold for aspiration hazard changed. Hydrocarbon-containing mixtures with a kinematic viscosity of 20.5 mm²/s or less at 40 °C are, under the conditions set out there, to be classified as aspiration toxicants and carry H304. That explains why thin, spray-friendly products more often require labelling than more viscous ones: the very property that enables an even spray pattern is the one that can trigger the classification. Whether a given product carries H304 follows from the classification of the whole mixture and is stated in its safety data sheet; it can be neither derived from nor excluded by the raw material base alone, including a renewable one.

On viscosity classification in general: Deutsche Bauchemie distinguishes thin concrete release agents below 5 mm²/s at 20 °C from medium-viscosity ones above 5 up to 25 mm²/s. Viscosity substantially determines spray pattern and consumption and thereby, indirectly, which application technique is eligible at all.

Water, soil, biodegradability

Concrete release agents are sprayed outdoors and often over unsealed ground. In the event of a leak the product can penetrate the soil and contaminate ground or surface water. In Germany, classification is governed by section 3 of the AwSV ordinance: not hazardous to water, generally hazardous to water, and water hazard classes 1 (slightly), 2 (clearly) and 3 (severely hazardous to water). Assessment is by acute toxicity, aquatic toxicity, biodegradability and bioaccumulation potential. Under section 3(4) AwSV, unclassified mixtures count as severely hazardous to water; the absence of a statement is therefore the least favourable statement.

Section 62(1) of the German Federal Water Act (WHG) requires installations for storing, decanting, producing, treating and using substances hazardous to water to be constructed and operated so that no adverse change to the properties of a body of water is to be feared. Important for the site: these provisions attach to installations, not to the act of spraying itself. We did not find a rule stating how much release agent may go astray when spraying outdoors. What applies here is the general precautionary principle and operational diligence, not a figure you could invoke. Residues are oil-containing and have to be disposed of separately.

Two terms are regularly conflated in biodegradability, and the test methods behind them differ. Ready biodegradability is tested to the OECD test guidelines of the 301 series; inherent biodegradability to the 302 series. Anyone saying “biodegradable” therefore has to say by which method, with what result, over what period and referred to what; otherwise the statement cannot be verified. According to Deutsche Bauchemie, mineral-oil-based release agents are as a rule inherently or primarily biodegradable, and release agents from renewable raw materials as a rule readily biodegradable; that is a general statement about product groups, not an assurance for an individual product. And Deutsche Bauchemie expressly recommends testing the complete formulation rather than the base oil alone: emulsifiers, preservatives and corrosion inhibitors bring classifications of their own.

A clarification, because the argument is sometimes made wrongly in sales: the German ChemVOCFarbV ordinance limits the content of volatile organic compounds in paints and varnishes for building coating and in vehicle refinishing products. Under its section 1, release agents do not fall within its scope. We checked this and set the ordinance aside deliberately.

Where the concrete will later be in contact with drinking water, section 17 of the German Drinking Water Ordinance applies. Exactly which requirements are placed on construction auxiliaries such as release agents, and which DVGW code of practice is to be used in a given case, we were unable to verify conclusively against the original text; the relevant worksheets are behind a paywall. We therefore name no rule number as binding, but the procedural route instead: this question belongs with the client and the water utility before the order is accepted, recorded in writing, and not on the building site.

Substances hazardous to water: what the law attaches to. Section 62(1) WHG addresses installations for storing, decanting, producing, treating and using substances hazardous to water. What it covers is the plant’s fixed and fixed-use equipment: storage tanks and drum stores, decanting and transfer points, stationary spray plant together with its bunding. The act of spraying itself, for instance manual application onto formwork, is not an installation in that sense, and we found no water-law quantity requirement for it; there, general diligence and the precautionary principle carry. Independently of that, for every product: know the water hazard class from section 15 of the safety data sheet, treat unclassified mixtures as severely hazardous to water under section 3(4) AwSV, and dispose of oil-containing residues separately. Whether a particular piece of equipment is an installation within the meaning of the AwSV, and which duties follow from that, is a question of law in the individual case. Outside Germany, the equivalent national water protection law applies.

Residues in the fired product

In ceramics and brick, the release agent does not leave the process as waste but through the kiln, with the two consequences described above: black coring where burnout is incomplete, and organic load in the kiln exhaust that has to be oxidised. That is the real reason why dosing in ceramics is more frugal than in concrete construction. There the excess migrates into the visible face; here it migrates through the product and afterwards through the kiln.

Sources: Deutsche Bauchemie, information for the user, April 2020 (defect patterns, cleaning, viscosity classification) [3]; Deutsche Bauchemie, status report, June 2015 (saponification, spray mist particle sizes, viscosity threshold for aspiration hazard, water hazard classes, biodegradability as a rule, testing the complete formulation) [4]; Holcim, guide to fair-faced concrete, 2023 (blowhole and void causes) [29]; BG BAU, data sheet C 321 “Concrete release agents”, 07/2021 (skin contact, respiratory protection, extension tube, workplace quantity) [5]; TRGS 900 “Occupational exposure limits”, edition January 2006, last amended GMBl 2026 p. 353 to 354 of 5 June 2026, clauses 2.4, 2.9, 2.10 and substance list [6]; TRGS 500 “Protective measures”, 09/2019, clauses 5.1(9), 5.2 and 5.3 [7]; TRGS 401, 10/2022, last amended 19 September 2024 [8]; TRGS 800, 12/2010 [9]; TRGS 510, 12/2020, Annex 2 [10]; Regulation (EC) No 1272/2008 (CLP), consolidated version [11]; AwSV section 3 [12]; WHG section 62(1) [13]; ChemVOCFarbV section 1 [14]; OECD test guidelines of the 301 and 302 series [15]; section 17 of the German Drinking Water Ordinance [16]; Taşkıran et al., 2025 (organic load and black coring) [26]; Ziegelwerk Bellenberg (regenerative thermal oxidation) [27]; Verband der Keramischen Industrie, 2000 (neutral zone, ejection force) [18].

Preventive measures

The table below is meant as a working document: one row per parameter, with interval, criterion and the source it comes from. Two things stand out. First, the most effective measure is not an inspection but a measurement that hardly any plant keeps: consumption in grams per square metre. Second, the chain of rules ends at the boundary of concrete construction; for ceramics and brick the plant has to set its own parameters.

A note on jurisdiction: the German rules cited here (TRGS, BG BAU, DIN, AwSV, WHG) are the ones we can substantiate. Outside Germany the corresponding national occupational safety, chemicals and water protection law applies; the measured quantities and intervals remain useful, the legal references have to be replaced by the local equivalents.

For ceramics and brick there is no chain of rules comparable to concrete construction. Neither the German ceramic industry association, nor the German brick and tile industry association, nor the German Ceramic Society publishes a document on release agents or mould care that we could find in the public domain. That is a finding, not an accident: in ceramics, good practice is set at plant level rather than by an association. In practice that means determining the parameters yourself: application rate per cycle, ejection force as a trend, green density and green strength as spot checks, visual inspection of the tool for deposits, and control of burnout via a section through the body.

Not a substitute for your risk assessment. This article describes the general state of the art and the applicable body of rules. It does not replace the risk assessment required under section 6 of the German Hazardous Substances Ordinance (GefStoffV) and TRGS 400, nor the operating instructions, staff instruction, or any product-specific approval. For each product, the current safety data sheet and technical data sheet prevail, together with the arrangements in force at your site. Responsibility for selection, operating conditions, monitoring and documentation rests with the employer.

Sources: given per row in the table above; the numbers refer to the list of sources at the end of this page.

Standards and rules

The shortest answer to the question of what the standards say about release agents is: remarkably little. And that is the useful answer. There is exactly one normative provision, and it is a German one. Everything else is association or accident-insurance guidance: technically valuable and soundly argued, but without the status of a standard. Confusing the two means arguing with the wrong weight when a dispute arises.

One point for readers outside Germany. The provision below sits in DIN 1045-3, the German standard for the execution of concrete structures. We did not verify whether the European execution standard, or your own national one, carries an equivalent clause on release agents, and we therefore do not claim that it does. Treat the requirement as the substance to look for in your own execution standard, not as a citation you can carry over.