Post-Installed Anchors in Early-Age Concrete: Why Young Concrete Breaks a Correct Anchor

Introduction: the failure nobody can see coming

Installing an anchor into concrete that has not yet reached its design strength is one of the few engineering errors that leaves no visible trace. Unlike a material defect or a manufacturing fault, here the anchor and the adhesive are entirely sound. The failure originates in the base material itself, which no longer matches the assumptions behind the manufacturer’s approval and behind every design equation the engineer used.

This article explains exactly what happens, physically and chemically, in concrete that is a few days old — and why both the American code and the European approval route converge on the same practical answer: wait.

We cover six connected subjects: (1) how concrete actually gains strength with age, and why an anchor cares about far more than the cylinder result; (2) what a rotary hammer does to a young matrix; (3) why diamond coring in early-age concrete is a double violation rather than a workaround; (4) what happens to the resin itself; (5) what the published test evidence shows; and (6) what the codes and approvals literally require, which is not the same thing in Europe and in the United States.

1. How concrete gains strength — and what the cylinder does not tell you

Characteristic concrete strength (f_ck) is defined at an age of 28 days in EN 206. Up to that age the cement hydrates, reacting with water to build calcium silicate hydrate (C-S-H), the phase that carries essentially all of the load. Eurocode 2 gives the development explicitly in EN 1992-1-1 §3.1.2, equations (3.1) and (3.2):

The coefficient s depends on the cement class: s = 0.20 for class R (rapid — CEM 42.5R, 52.5N, 52.5R), s = 0.25 for class N (CEM 32.5R, 42.5N) and s = 0.38 for class S (slow — CEM 32.5N, and mixes with high slag or fly-ash replacement). This single coefficient is the difference between a pour that is usable at 14 days and one that is not usable at 21.

Values computed from EN 1992-1-1 eq. (3.2) at a constant 20°C. Lower curing temperatures shift every column to the right — see the maturity discussion in section 8.

The number that actually decides: is the concrete even inside the approval?

Percentages of f_ck are reassuring and slightly misleading. What matters for an anchor is the absolute characteristic strength on the day of installation, because every European Technical Assessment states a validity range, and for practically all anchors that range starts at C20/25. Below C20/25 there is no approved resistance at all — not a reduced one.

Take an ordinary C25/30 pour with a class N cement. Eurocode gives f_cm = f_ck + 8 = 33 N/mm², so f_ck(t) = β_cc(t) · 33 − 8:

What the compressive result still hides

Even once the compressive strength is inside the range, three properties that govern anchor behaviour lag behind or are simply not measured by a cylinder test:

• Concrete tensile strength — this, not compression, controls concrete cone failure. EN 1992-4 eq. (7.2) gives the cone resistance as N⁰_Rk,c = k₁ · f_ck^0.5 · h_ef^1.5. The square-root relationship is the useful part: at 7 days our C25/30 is at 71% of its characteristic strength, which still leaves about 84% of the cone capacity. Cone failure is therefore the least penalised mechanism — which is precisely why engineers who check only the cone conclude, wrongly, that young concrete is a minor issue.
• Integrity of the interfacial transition zone (ITZ) — the thin shell of paste around every coarse aggregate particle, and the weakest region in any concrete. In young concrete the ITZ is still rich in elongated ettringite needles and large portlandite (Ca(OH)₂) crystals rather than dense C-S-H. No cylinder test sees this, and as section 2 shows, it is the ITZ that a hammer drill attacks.
• Free water in the capillary pores — water that has not yet reacted with the cement. It has no effect on a compression test and a decisive effect on the polymerisation of a resin, as section 4 explains.

2. What a rotary hammer does to young concrete

Hammer drilling works through two simultaneous motions: rotation of the bit and mechanical impacts at roughly 3,000–5,000 blows per minute. In mature concrete the intended mechanism is brittle fracture of the hardened paste into dust, with the carbide tip cutting straight through coarse aggregate. The damage zone around the drill hole is narrow — on the order of 0.5–1.5 mm — and the bore leaves a rough but mechanically sound surface. That roughness is not a side effect; it is what a bonded anchor is designed to grip.

The mechanism, step by step

a. Fracture of the interfacial transition zone

In young concrete the ITZ around each coarse aggregate particle still consists largely of elongated ettringite crystals and crystalline portlandite rather than a dense C-S-H matrix. These phases are brittle under dynamic loading. Every hammer blow sends a radial shock wave into the material: in mature concrete that wave is absorbed by the C-S-H matrix, while in young concrete it splits the ITZ and propagates micro-cracking around every aggregate particle it meets.

b. Aggregate pull-out instead of cutting

In mature concrete the bond holding an aggregate particle in the matrix is stronger than the force needed to cut it, so the bit cuts through. In young concrete that relationship reverses: the bit plucks the aggregate out, leaving an asymmetric cavity with an irregular wall. Such a hole is locally oversized, exposes immature paste that was sitting against the aggregate, and no longer matches the cylindrical geometry every bond calculation assumes.

c. Drilling slurry rather than drilling dust

In mature, dry concrete the drilling dust is dry and leaves easily with compressed air and a brush. In young concrete the cuttings contain partially hydrated C-S-H and free water, producing a sticky paste on the bore wall that standard cleaning does not remove. That layer, a few tenths of a millimetre thick, sits between the resin and sound concrete.

d. Effect on the effective embedment

Every ETA refers to a measured effective embedment depth h_ef. In young concrete, even when the hole reaches the correct geometric depth, the outer skin of the bore is damaged deeply enough that it can no longer be counted as load-bearing. The effective h_ef in practice is smaller than the one in the calculation — and since bond resistance is linear in h_ef (EN 1992-4 eq. 7.10: N⁰_Rk,p = τ_Rk · π · d · h_ef), that loss transfers directly into the resistance.

3. Diamond coring in early-age concrete — two violations at once

Diamond coring is often chosen because the concrete is young: it vibrates less, it leaves a hole that looks clean, and it feels like the considerate option next to a fresh pour. It is a classic engineering mistake. Diamond coring works by abrasion under continuous water cooling — segments impregnated with diamond grit grind the concrete away while 2–5 litres of water per minute flush the cuttings out. In mature, dry concrete this produces a clean cylindrical cut. Both of those qualities — clean and cylindrical — are the opposite of what a bonded anchor needs.

a. Water into concrete that is already saturated

A Ø14 mm hole cored to 90 mm depth puts several litres of water through the hole and its surroundings in under two minutes. In mature concrete most of that water leaves with the cuttings; the capillary system is relatively closed and the water stays mainly on the bore surface. In young concrete the situation reverses, because the capillary network is still open and connected:

• Over-saturation of the capillary pores. Young concrete already contains unreacted mix water. Coring water is added to that, raising the liquid content in the zone that will contact the resin far above what any injection system’s instructions permit.
• No way to dry it. Compressed air removes water from the bore surface, not from the surrounding mass. Absorbed water continues to migrate back out over the following hours and days — exactly while the resin is trying to polymerise.
• A permanently weaker shell. The effective water/cement ratio in the concrete immediately around the hole shifts upward. The concrete that finally hardens there reaches a lower ultimate strength than the parent pour. The weak zone does not cure away; it is built in.

b. Calcium leaching

Young concrete contains a large reservoir of free portlandite, Ca(OH)₂, a by-product of hydration. Flowing coring water dissolves it and carries it away, leaving a decalcified surface layer on the bore wall that is more porous and appreciably weaker than the concrete behind it. The effect is visible to the naked eye: the bore surface is lighter than the interior, and a whitish film of re-precipitated portlandite often appears as the water evaporates. The resin then bonds to that layer, not to sound concrete.

c. Slurry that becomes part of the structure

In mature concrete, coring slurry — water plus ground concrete fines — is flushed out or remains loose and can be cleaned. In young concrete something considerably worse happens:

• The slurry contains partially hydrated C-S-H and unhydrated cement grains.
• Those particles are pressed into the open pore structure of the young concrete during coring.
• As the surrounding concrete continues to hydrate over the following days, the slurry hydrates with it and bonds chemically to the matrix.
• The result is a skin of defective, low-strength concrete on the bore wall that cannot be removed, because it has become the structure.

This is fundamentally different from hammer drilling dust, which stays loose and can be brushed and blown out. Diamond slurry in young concrete becomes permanent, and the resin bonds to it.

d. A smooth surface removes the mechanical component

Hammer drilling leaves a bore roughness on the order of 100–300 micrometres — coarse sandpaper. Diamond coring leaves roughly 5–15 micrometres — a polished floor. Bonded anchors are qualified on the hammer-drilled surface. On a cored surface the resin has no asperities to key into, so the mechanical interlock component of the bond is largely lost and the connection depends on chemical adhesion alone. Combine that with a decalcified, slurry-coated wall and the two deficiencies are multiplicative rather than additive.

4. What happens to the resin itself

A bonded anchor relies on two parallel load-transfer mechanisms: chemical adhesion of the cured resin to the bore wall, and mechanical interlock of the resin into the roughness of that wall. Early-age concrete undermines both, through independent routes.

a. Free water interferes with polymerisation

The resins used in injection anchors are predominantly vinyl ester, epoxy-acrylate or pure epoxy. All cure by a polymerisation reaction in which initiators open double bonds to build a three-dimensional network. Free water at the bore surface interferes in three ways: it dilutes the initiator locally, so the polymer in the critical interface layer never reaches full cross-link density and stays softer and weaker in shear; part of it flashes off against the exotherm, leaving micro-voids at the resin–concrete interface; and it disturbs the stoichiometry that a two-component system depends on.

b. Alkaline attack during cure

In young concrete the pore solution is strongly alkaline — hydroxide concentrations give a pH in the region of 13. That environment attacks ester linkages in some polymers by saponification, a hydrolytic cleavage of the polymer backbone. Unsaturated polyester is the most vulnerable, vinyl ester is more resistant and epoxy is the most resistant of the three. But no resin is fully immune before it has become a dense solid: while the network is still forming it is open and exposed, and the free water of a young matrix is exactly the medium through which hydroxide ions reach it. The outcome is an interface layer that remains soft and gel-like precisely where the load has to be transferred.

c. Loss of mechanical interlock

As section 2 described, the bore wall in young concrete carries a slurry layer. Even full cleaning to the manufacturer’s procedure — brush and compressed air, repeated — does not entirely remove it. The curing resin bonds to that soft layer instead of to sound concrete, so failure occurs within the slurry layer rather than in structural concrete. In a pull-out test this is visible: a grey-black film adhering to the extracted resin plug is the signature of an anchor installed in concrete that was too young.

5. What the published test evidence actually shows

This is where a technical article has to be careful. The mechanisms above are well established; the magnitude of the penalty is product-specific, and the honest answer is that it varies enormously between adhesive chemistries. Several manufacturers have published qualification testing in early-age concrete, and those datasets are the best public evidence available.

DEWALT’s Technical Bulletin 1 reports tension tests on five adhesive systems installed in concrete at 7, 14 and 21 days, and expresses the outcome as a reduction factor α_age applied to the bond strength:

Compiled from DEWALT Technical Bulletin 1 (2018 and 2023 revisions) and the equivalent white paper published by Adhesives Technology Corporation (Hanley, 2022), which reports the same range of 0.70 to 1.00 across six products. All results are static tension tests in dry, uncracked, normal-weight concrete.

Three conclusions follow, and they are more useful than a single headline percentage:

• The penalty is real but chemistry-dependent. Slow-curing epoxies can be almost unaffected at 7-day concrete; fast-curing acrylics lose 20–30%. A blanket claim that early-age concrete halves the capacity of every adhesive is not supported by the published data, and an engineer who repeats it will be corrected.
• Time under cure partly recovers the loss. The same acrylic that scores 0.70 at minimum cure scores 1.0 if the adhesive is allowed an extended cure and the anchor is not loaded until the concrete is 21 days old. In other words, part of what is being measured is the concrete continuing to hydrate around an already-installed anchor.
• None of this is a licence to install early. Every one of these bulletins states that installation below 21 days lies outside the scope of the product’s evaluation report, is not in compliance with the code, and requires the approval of the engineer of record. The manufacturers publish the factors so that a documented, engineer-approved exception is possible — not to make early installation routine.

6. What the codes actually require — and where they differ

The “21 days” figure circulates on sites everywhere, usually without anyone being able to say where it comes from. It is worth being precise, because the European and American routes arrive at the same practice by different legal mechanisms, and an engineer who cites the wrong one will be challenged.

The American route: an explicit code rule

In the United States the requirement is written into the concrete design code itself. ACI 318-19 §17.2.2 — §17.1.2 in ACI 318-14, and D.2.2 in ACI 318-11 — requires that adhesive anchors be installed in concrete having a minimum age of 21 days at the time of anchor installation. This is adopted by reference into the International Building Code, so it is a legal requirement, not a recommendation.

Two features of the rule are frequently misunderstood:

• It is an age criterion, not a strength criterion. Demonstrating by cube or cylinder test that the concrete has reached its design strength at 10 days does not satisfy §17.2.2. Both the age and the strength have to be met.
• It applies to adhesive anchors specifically. Mechanical anchors are governed by the strength limits in their own evaluation reports.

The European route: the approval, not the code

EN 1992-4:2018 does not contain an equivalent age clause, and neither does ETAG 001 Part 5 impose a minimum age on the concrete test members. The European system works differently: EN 1992-4 is a design standard that applies to a fastener which has a European Technical Assessment, and it requires the anchor to be installed in accordance with the manufacturer’s printed installation instructions supplied with that assessment. The binding conditions therefore sit in two documents:

• The ETA fixes the validity envelope — normally concrete C20/25 to C50/60, a single nominal drill-hole diameter, the permitted drilling method (very often “hammer drilling only”), the permitted hole condition (dry, wet or water-filled) and the temperature range.
• The installation instructions carry the concrete-age condition where the manufacturer imposes one, along with the cleaning sequence and curing times.

This is why, in Europe, the strongest argument is not a quoted clause number but the one made in section 1: a C25/30 pour is simply not a C20/25 material at 7 days. It falls outside the declared validity range of the assessment, and an installation outside that range has no declared resistance. From the designer’s side, EN 1992-4 gives no way to compute a characteristic resistance for a fastener installed outside its assessed conditions.

Where the two systems agree

• Concrete below the approved strength class has no design resistance, reduced or otherwise.
• Installation outside the assessed conditions transfers the entire responsibility to the engineer of record, who has no test basis to rely on.
• Both systems treat the drilling method, the hole diameter and the cleaning procedure as part of the qualified system, not as site variables.

7. Mechanical anchors in early-age concrete

Mechanical anchors are often proposed as the way out, and within limits that is correct — but the reasoning matters, because the limits are real.

A torque-controlled expansion anchor, an undercut anchor or a concrete screw transfers load by mechanical means: expansion pressure against the bore wall, bearing on an undercut, or threads cut into the concrete. None of these depends on a chemical reaction at the interface, so the three resin-specific problems disappear entirely. Free water does not interfere with polymerisation because there is no polymerisation. High pH attacks nothing. A thin slurry layer is crushed by expansion pressure rather than becoming a bond-breaker.

What does not disappear is the behaviour of the concrete itself:

• Expansion anchors load the concrete in tension. The expansion pressure that makes the anchor work also puts the surrounding concrete into a state of splitting tension. In a young matrix with immature tensile strength, the concrete can crack under the setting torque itself — the anchor spins or pulls before it ever sees a service load.
• Concrete screws cut a thread into the material. The thread flanks bear against concrete whose compressive strength is still developing, and the thread can strip.
• Cone capacity still scales with √f_ck. Section 1 showed this is the mildest of the reductions, but it is not zero, and it applies to every anchor type.
• The approval envelope still applies. A mechanical anchor’s ETA also states a strength-class range. Some products are assessed down to C12/15, which genuinely extends the usable window; most start at C20/25 like everything else. The product datasheet decides, not the anchor category.

8. Why the drill-hole diameter is locked — neither smaller nor larger

A recurring site question: “the correct bit is binding, can I go one size up?” — or the reverse, “I only have a bit 1 mm smaller, it will hold even better, right?” The answer to both is no. The ETA of every bonded anchor defines the nominal drill-hole diameter d₀ as a single value, because it is an inseparable part of the system that was tested.

The physics of the annular gap

In an installed bonded anchor there is an annular gap between the threaded rod and the bore wall, typically 1–3 mm per side, filled with resin. That resin thickness is not arbitrary: it is engineered to transfer shear uniformly from the rod into the concrete.

Why an undersized hole fails

• No intervening resin. The rod contacts the bore wall directly and load is transferred by friction, which is neither assessed nor included in any calculation.
• The resin is extruded during insertion. Instead of enveloping the rod, it escapes upward out of the hole, leaving voids at depth and no load transfer in the deepest part of the embedment.
• The rod cannot reach nominal depth. It binds short of the designed h_ef, and since bond resistance is linear in h_ef, that shortfall comes straight off the capacity.
• Thread damage. Threads abrade when a rod is forced into a tight hole, reducing the steel resistance as well.

Why an oversized hole fails

• Bulk shear in the resin. As thickness grows, the resin behaves as a volume of material rather than an adhesive layer and fails internally, before load ever reaches the concrete. The shear strength within the resin is lower than either of the two interfaces it connects.
• Loss of confinement. In a correct annulus, the resin bearing against the bore wall develops a favourable confining pressure that raises shear capacity. Too wide a gap loses it.
• Higher exotherm. Polymerisation is exothermic. A thicker resin body reaches a higher peak temperature, which promotes shrinkage cracking within the anchor during cure.
• It breaks the design equation. Bond resistance in EN 1992-4 eq. (7.10) is computed on the perimeter π·d·h_ef with d the nominal diameter. The equation does not know the hole was changed; a diameter that differs from the assessed value simply invalidates the calculation.

9. Threshold rules for a compliant installation

• Concrete age. Under ACI 318-19 §17.2.2, adhesive anchors require a minimum concrete age of 21 days — an absolute requirement that a strength test does not replace. Under the European route, follow the age condition in the manufacturer’s installation instructions; where none is stated, 21 days remains the defensible site rule because it is roughly where an ordinary pour holds real margin over C20/25.
• Verified strength, not assumed strength. Confirm by cube or cylinder test, or by a calibrated rebound hammer correlated to cores, that the concrete is inside the strength class range printed in the anchor’s ETA — for most products C20/25 to C50/60. Winter pours, high-replacement cements and thin sections all reach that line later than the calendar suggests.
• Exact drill-hole diameter. Use the d₀ stated in the ETA of the specific anchor. Do not deviate in either direction, not even by half a millimetre.
• The drilling method the ETA permits. If the assessment says hammer drilling, diamond coring is not an alternative, whatever the site conditions appear to justify.
• Full hole cleaning. Follow the manufacturer’s sequence exactly — typically compressed air, brush, compressed air, twice over. This matters most in concrete that is young but permitted, at 21 to 28 days, where a sticky dust layer can still form.
• Temperature. Most assessments permit a base-material temperature of about +5°C to +40°C. Low temperatures slow both the concrete’s hydration and the resin’s polymerisation, so they penalise twice.
• Moisture condition. Check whether the ETA covers dry, wet or water-filled holes. Concrete under 21 days old is internally wet by definition, even when the surface looks dry.
• Documentation. Record the pour date, the age at installation, the measured strength and the batch of adhesive. If an anchor is ever questioned, this record is the difference between a defensible installation and an indefensible one.

Frequently asked questions

Summary

Standards, references and further reading

• EN 1992-4:2018 — Eurocode 2: Design of concrete structures — Part 4: Design of fastenings for use in concrete. Concrete cone resistance eq. (7.2); bond resistance of bonded fasteners eq. (7.10).
• EN 1992-1-1 — Eurocode 2 Part 1-1, §3.1.2, equations (3.1) and (3.2): development of compressive strength with age, and the cement-class coefficient s.
• EN 206:2013+A2:2021 — Concrete: specification, performance, production and conformity. Definition of characteristic strength at 28 days.
• EN 13670 — Execution of concrete structures: curing and protection requirements.
• EAD 330499-01-0601 — Bonded fasteners for use in concrete (EOTA). The assessment route behind every injection-system ETA.
• ETAG 001 Part 5 — Metal anchors for use in concrete: bonded anchors, EOTA, April 2013. §5.1.2 requires the full test programme to be repeated for each drilling technique an approval is to cover. Available from EOTA.
• EOTA TR 048 — Details of tests for post-installed fasteners in concrete.
• ACI 318-19 — Building Code Requirements for Structural Concrete, §17.2.2: minimum concrete age of 21 days at the time of adhesive anchor installation (§17.1.2 in ACI 318-14; D.2.2 in ACI 318-11), adopted by reference into the International Building Code.
• ACI 355.4 — Qualification of Post-Installed Adhesive Anchors in Concrete, and ACI 355.2 for mechanical anchors.
• ASTM C1074 — Standard Practice for Estimating Concrete Strength by the Maturity Method. The basis for converting time and temperature history into an equivalent age.
• DEWALT, Technical Bulletin 1 — Early Age Concrete and its Effect on Adhesive Anchor Performance, 2018 and 2023 revisions. Published α_age reduction factors of 0.70–1.00 for five adhesive systems installed at 7, 14 and 21 days. 2023 revision (PDF).
• J. Hanley, Adhesives Technology Corporation — White Paper: Early Age Concrete and Its Effects on the Bond Strength of ATC’s Anchoring Products, 30 August 2022. Independent confirmation of the same 0.70–1.00 range across six products. PDF.
• S. Meißl, K. Ninčević, B. E. Abali, R. Wan-Wendner — Mortar cure-dependent effects on adhesive anchor systems loaded in tension, 2023. 93 confined pull-out tests; identifies moisture uptake and loss in the cured adhesive as a mechanism affecting long-term stiffness. Full text.
• R. Eligehausen, R. Mallée, J. F. Silva — Anchorage in Concrete Construction, Ernst & Sohn, 2006. The standard reference on anchor mechanics.
• fib Bulletin 58 — Design of anchorages in concrete, 2011.
• Adit technical documentation — the ETA and installation instructions of the specific anchor always take precedence over any general guidance, including this article.

Written by Yves de Lathouwer, Adit Ltd. This article is general engineering guidance and does not replace the European Technical Assessment, the manufacturer’s printed installation instructions, or the judgement of the engineer of record for a specific project. Where a figure is described as indicative it reflects field experience rather than a published test programme, and is flagged as such in the text.