Interlayer Bonding in 3D Printed Concrete

Interlayer bonding is the mechanical strength of the interface between two successively deposited filaments of printed concrete. In extrusion-based 3D concrete printing it is, as a rule, the weakest plane in the finished element: the published literature consistently reports lower tensile and shear capacity across the layer interface than within the layer itself. This follows directly from how the material is placed. Cast concrete is poured into formwork and compacted, usually with vibration, so a monolith forms without internal boundaries. Printed concrete is laid as a sequence of filaments, each deposited onto a surface that has already begun to stiffen, with no vibration, no formwork pressure and only the weight of the extruded material to force contact. Structurally, everything that determines the quality of a printed element passes through this interface.

What it is / Why it matters

Several mechanisms combine to weaken the interface. There is no vibration, so the two layers are never fluidised together; the particles of the upper filament intermix with those of the lower one only in a thin boundary region, and sometimes barely at all. The real contact area is smaller than the nominal one: an extruded filament has rounded or slightly bulged edges, so where two filaments meet, small longitudinal voids are trapped along the boundary between them and along the sides of the layer. Imaging studies of printed specimens regularly find elevated porosity concentrated in a narrow band at the interface, and the voids there act as pre-existing flaws from which cracks start.

The lower surface also changes with time. It stiffens through thixotropic structural build-up and early hydration, so it can no longer deform to accommodate the new filament pressed onto it. And it dries: water evaporates from the exposed surface, and in a dry or warm environment a skin can form within minutes. A dried surface both reduces the hydration available to knit across the joint and can draw water out of the fresh filament placed on it. If the interval becomes long enough, the joint is a cold joint in the conventional concrete sense — the layers are in contact but not continuous.

The structural consequence is anisotropy. Strength and stiffness are direction-dependent: measured perpendicular to the layers, values are usually lower than measured parallel to them, and results depend on how a specimen was cut and loaded relative to the printing direction. This is not a defect to be reported away; it is a property of the process, and it means a printed element must be treated as anisotropic in both design and testing. Any reported strength for printed concrete is incomplete unless the orientation of the load relative to the layers is stated. A printed wall loaded mainly in compression perpendicular to the layers may perform well, while the same wall under out-of-plane bending or shear is governed by the weakest interface it contains.

How it is measured

There is no harmonised European standard for the interlayer bond of printed concrete. The methods below are adapted standard tests or research methods, and results are strongly method-dependent, which is one reason published values scatter widely.

Direct tension (pull-off) tests on sawn prisms are the most direct approach: a prism containing a single interface is glued between loading plates and pulled apart perpendicular to the layers. It measures exactly the quantity of interest but is sensitive to specimen preparation, eccentricity and glue failure. Pull-off testing of surfaces follows the general principles of EN 1542 in the repair field.

Splitting tensile tests in the manner of EN 12390-6, with the specimen oriented so that the splitting plane coincides with the interface, are simpler to perform and less sensitive to alignment, but they measure an indirect tensile strength rather than a true interfacial one.

Three-point or four-point bending on notched prisms, with the notch placed at the interface, forces the crack to run through the joint and allows fracture energy to be estimated as well as peak load. This adapts the general approach of EN 12390-5 and of fracture-mechanics test methods.

Shear tests, including direct shear and inclined-plane arrangements, characterise the interface under the sliding action relevant to out-of-plane loading; geometries vary between research groups and results are not directly comparable.

X-ray CT scanning and optical image analysis quantify the porosity, void size and void distribution at the interface without loading the specimen. These are characterisation methods rather than strength tests, and the clearest way to see why a given interface is weak.

Standard hardened-concrete tests such as EN 12390-3 for compressive strength remain relevant for the bulk material but say little about the joint unless specimens are cut and oriented deliberately. Work by RILEM TC 304-ADC and its predecessor TC 276-DFC is directed at establishing agreed test procedures for printed material.

Typical ranges

The values below are indicative and reflect trends reported across the 3DCP literature in general terms, not measurements made by Concreef. Scatter between studies is large because mixes, printers, environments and test geometries differ.

None of these figures should be used for design; bond capacity has to be established for a specific material, printer and environment.

What changes it

The time gap between layers is the variable most consistently identified as dominant. It is not an independent choice: it is set by the print path, the object's footprint and the printing speed, so path planning is bond planning. A small perimeter returns to the same point quickly; a large footprint or a long loop can leave a surface waiting far too long.

The surface moisture state of the lower layer follows from the time gap, but also from ambient humidity, temperature and air movement in the print space. A draught across a printing element can dry the top surface in a fraction of the time it would take in still air. Warm conditions accelerate both drying and hydration and therefore shorten the usable window.

Nozzle standoff and contact pressure matter directly. If the nozzle is held too high, the filament falls onto the previous layer and the contact is light; if the standoff is set so that the nozzle presses the extruded material down, the filament is deformed into the layer below, increasing the real contact area and squeezing out entrapped air. A flat or trowelling nozzle face that irons the deposited material works the same way.

Material thixotropy is a trade-off. A strongly structuring mix stands up well and allows more layers per hour, but its surface also becomes unreceptive faster; a less thixotropic mix bonds more readily at the cost of buildability. Retarders, viscosity-modifying agents, supplementary cementitious materials and the water content all shift where a material sits on this trade-off.

Mitigation measures in practice are: shorter and more uniform layer cycle times; path planning that avoids long returns; rewetting the surface with a light water mist or applying a cement-based bonding agent before the next pass; nozzle geometry and standoff set to press the filament; and control of the print-space climate. Fibres bridge the interface only partially — flow through the nozzle aligns most of them within the filament rather than across the joint — so they improve bulk toughness far more than interfacial strength.

Failure modes

The characteristic failure is delamination: a crack that initiates at an interface void and runs along the joint, separating layers under tension, bending or shear. Because the interface porosity forms a continuous weak band, such a crack propagates with little resistance once started.

Cold joints are the extreme case, produced by an excessive time gap, an interrupted print, an unplanned pause for maintenance or a mix that stiffened faster than expected. They are often visible as a clean, sharp line between layers and can sometimes be opened by hand at an edge.

Surface drying produces a related failure in which the interface looks acceptable but the bond is brittle and low in strength, because the fresh filament lost water to the dried surface under it. Excessive nozzle standoff leaves visible voids at the filament edges, producing both weak bond and a porous, water-permeable external surface; excessive pressing, conversely, can deform a filament sideways and distort geometry while improving bond.

Finally, a systematic reporting failure: giving a single strength value for a printed element without stating the direction of loading relative to the layers. This obscures anisotropy and can make an element appear stronger than it is in the governing direction.

Concreef context

Concreef is a Bulgarian 3D concrete printing company based in Sofia. It runs material trials in its Sofia workshop on a Crane WASP printer, currently printing with a commercial premix while developing its own dry mix. Concreef has no published test data, and no value in this page is a Concreef measurement.

In workshop practice, interlayer bonding is treated as a process property, not a material property alone. Layer cycle times are kept as short and as uniform as the geometry permits, and print paths are considered from the outset with the return interval in mind rather than only the finished shape. Workshop temperature, humidity and air movement are recorded with each trial, because the same mix and cycle time give a different interface in different conditions. Nozzle standoff is a deliberate setting rather than an incidental one, since it governs how firmly each filament is pressed into the layer below. Where trials compare interfaces, the loading direction relative to the layers is stated, and results from sawn specimens are treated as indicative of a specific material and printing condition rather than as a general property of printed concrete. Because a commercial premix is used for current printing, its documented properties define the baseline, and Concreef makes no claim that any material it prints is certified or tested to a standard for interlayer bond.