AAC's reputation for low lambda is earned at oven-dry or air-dry state: Jin et al. measured the effective thermal conductivity rising by a factor up to 3.8 across the studied moisture range, with the high-to-low slope change at roughly 15% moisture by mass for 415, 520, and 630 kg/m³ samples [S2].
For process engineers, the practical takeaway is that every 5% gain in moisture content can push thermal conductivity up 30–50% [S3], and EN 771-4 performance requirements are stated at air-dry reference conditions per ISO 10456, not at production moisture [S1].
Reference values: dry vs moist lambda for typical AAC density classes
Commercial AAC spans roughly 250–600 kg/m³ bulk density, with dry thermal conductivity typically 0.07–0.13 W/m·K across that range [S4]. When that range is read alongside the Jin factor-up-to-3.8 envelope [S2], the worst-case wet lambda at 100% moisture by mass lands near 0.50 W/m·K for the lightest density classes, a value that effectively eliminates AAC's insulation advantage and pushes the wall into a regime dominated by conduction through the liquid phase filling the pore space.
The 415 kg/m³, 520 kg/m³, and 630 kg/m³ samples tested with the transient plane source technique all show the same breakpoint structure, and a two-section piecewise linear correlation was proposed to capture the dry-side and wet-side slopes [S2]. That breakpoint near 15% moisture by mass is the practical inflection point at which a wall crossing from "construction dry" to "wetted service" begins to lose insulation performance disproportionately per unit of added water.
Why water changes lambda: pore-scale physics, not bulk chemistry
AAC's dry lambda is low because its 50–80% porosity is filled with air at roughly 0.026 W/m·K, and the matrix is a tortuous network of micropores, macropores, and artificial air pores that scatter conductive heat flow [S6]. When water replaces air in those same pores, the conductivity of the pore-filling medium jumps by an order of magnitude, and the matrix's three-phase composite (solid + water + air) collapses toward a two-phase solid-water system that conducts heat far more efficiently than the solid-air baseline [S2].
The fractal model built on a self-similar Sierpinski carpet construction captures this by adding a water phase to the dry two-phase structure, with the geometric configuration of how water occupies the pore network (film vs full-pore) controlling the predicted slope [S2]. The practical conclusion is that any factor that increases liquid water contact with the solid AAC matrix, including capillary uptake, condensation, and driving rain, degrades insulation at a rate the European harmonised calculation route historically overpredicted, with the Schoch work showing actual production-moisture surcharges far below EN ISO 10456 tabulated values [S5].
Comparison: dry reference vs production moisture vs in-service wet
Three regimes are useful for spec work, and they map onto different calculation and warranty assumptions. [S5]
Dry reference (air-dry, ISO 10456): the 0.07–0.13 W/m·K range used in declared values across 250–600 kg/m³ density classes [S4]. This is the figure that goes into U-value calculations, CE/UKCA declarations under EN 771-4 [S1], and any warrantee lambda.
Production moisture (freshly cut, often 25–35% by mass): the European standards theoretically impose "immense surcharges on λ" at these levels [S5], but measured behaviour on factory-fresh AAC shows the actual surcharge is far lower than tabulated, and the typical-use moisture conversion factor Fm stays at or under 5% in several independent investigations [S5].
In-service wet (wetted by capillary action, condensation, driving rain, or construction water): the high-slope region above 15% moisture by mass in the Jin correlation, where lambda rises faster per unit of added water and can reach the 3.8× envelope ceiling [S2]. A 5% moisture excursion inside this regime adds 30–50% to lambda, the figure most relevant for retrofit diagnosis and for AAC block failure-mode analysis when walls perform worse than their declared value.
Workmanship and joints: where moisture enters real walls
Moisture rarely enters an AAC wall uniformly. The Wesołowska study on single-layer AAC walls showed that vertical contact width and horizontal joint thickness control the local thermal micro-bridge zones, and that maximum 3 mm vertical contact and 2 mm horizontal joint thickness keep a 42 cm wall inside European thermal requirements at air-dry state per ISO 10456 [S1]. Gaps beyond those tolerances create local high-conductivity zones that produce a non-linear U-value penalty, and the same gaps also act as preferential paths for moisture to enter the wall during construction and service [S1].
Worldwide, approximately 30% of buildings use AAC, with the highest single-layer-wall share in the 300–400 kg/m³ density range where the moisture sensitivity of lambda is steepest per unit of added water [S1]. For that reason, thin-bed mortar systems (1–3 mm joint thickness) and tight dimensional tolerance are not only a thermal-bridge control measure, they are a moisture-control measure, since thicker cementitious joints are the most common route for construction water to penetrate the wall and drive the in-service moisture above the 15% breakpoint [S1].
Testing methodology and how to read the numbers
Steady-state guarded hot plate testing per EN 12664 or ASTM C177 with 38°C hot plate and 23°C cold plate is the standard route, conditioned to 5–8% equilibrium moisture to represent real-world dry service [S3]. The transient plane source (TPS) method used in the Jin study is faster and well-suited to mapping a full moisture curve, but the values must still be tied back to the guarded hot plate reference for declared values [S2][S3].
Spec engineers should request the moisture condition at which lambda was measured, not just the W/m·K number, since a 0.10 W/m·K figure at air-dry is not the same product as a 0.20 W/m·K figure at 20% moisture, even if both come from the same density class [S2][S3]. For on-site diagnosis, a moisture analyzer reading above 15% by mass is the signal that the wall is operating in the high-slope region of the lambda-moisture curve, and any calculated U-value built on declared dry lambda will under-predict in-service heat loss.
Limits and what the data does not cover
The 3.8× envelope is bounded by the 415, 520, and 630 kg/m³ samples measured by Jin and may not extrapolate linearly to the 250–300 kg/m³ super-light grades now entering the European market [S2][S4]. Likewise, the Xella Fm of approximately 5% is an empirical in-service conversion factor for typical wall conditions, not a worst-case figure, and conditions with active water ingress or freeze-thaw cycling can push the in-service moisture well above the test envelope [S5].
The Schoch experiments that produced Fm ≤ 5% were run on Xella Ytong-class AAC in the 350–500 kg/m³ range, so transferring the same conversion factor to other manufacturers or to very low-density grades requires independent verification [S5]. For thermal-imaging surveys of existing walls, a thermal imager reading that shows cold bridging concentrated at joints and reveals an asymmetric wall temperature profile is a strong indirect signal that workmanship gaps have admitted enough moisture to push local lambda above the declared value, and the same gap geometry that drives the thermal micro-bridge also drives the moisture pathway [S1].
Operational signals to track
A second trackable signal is the adoption of super-light grades below 300 kg/m³, where the moisture sensitivity of lambda is highest per the Jin breakpoint and where European EN 771-4 declarations are still being validated against long-term hygrothermal field data [S1][S4].
Related analysis: Busway and Bus Duct Lead Times for 2026 Data Halls: Spec, Schedule, and Sourcing.