Cold box binder viscosity in foundry core production is measured at the point of sand mixing as a split of Part I and Part II, with rotational viscometry per ASTM D2196-20 using a Brookfield DV2T at 25 °C and spindle 4 at 60 min⁻¹. In amine-cured phenolic urethane systems, Part I is a benzylic ether phenolic polyol dissolved in a solvent blend composed of aromatic hydrocarbon, dibasic ester, and a polar oxygenated solvent, while Part II is a polymeric methylene diphenyl diisocyanate carried in a similar but less polar solvent fraction. The as-received viscosity acceptance window in high-speed core rooms is typically 180–350 mPa·s for Part I and 30–150 mPa·s for Part II at 25 °C, although published supplier technical bulletins indicate that formulation-specific limits may shift by 20–40 mPa·s depending on solvent flash point and oligomer molecular weight distribution. Viscosity is not a standalone specification; it is coupled to wet film formation on the sand surface, where a target coating thickness of 0.5–1.5 μm on a 100–150 μm silica grain requires sufficient capillary penetration within the 10–30 s mix-to-shoot residence time. Lower viscosity improves initial distribution but increases migration into fine sand fractions, while higher viscosity reduces resin-rich agglomerates only if high shear in the mixer reaches 500–1000 s⁻¹. In production, the practical lower limit is set by solvent flash formation rather than fluid mechanics: a Part I below 150 mPa·s at 25 °C often contains excess polar solvent that produces amine salt bloom on the core surface after gassing.

Adhesion at the binder–sand interface after amine cure depends less on bulk tensile strength than on the continuity of the boundary film and the population of covalent and hydrogen-bonded interactions. Silica sand carries surface silanol groups at a density of roughly 3.0–5.0 SiOH per nm² after thermal drying, and the phenolic polyol fraction of the binder must wet these sites within 2–5 seconds before exterior solvent evaporation immobilizes the film. Contact angle measurements on washed and dried quartz using a Krüss K100 or similar Wilhelmy plate apparatus show that cold box binders with dynamic surface tensions above 38 mN/m at 25 °C produce incomplete wetting on angular lake or bank sands with residual clay films. Sand adhesion is routinely assessed indirectly through cold tensile strength on 10 mm × 10 mm dog-bone specimens cured in a Dietert universal sand strength machine, with acceptance limits of 1.8–2.8 N/mm² for full-cured cores and 0.8–1.4 N/mm² for 30-minute bench life simulation. These values are comparable to tensile results from ISO 6892-1:2019 only in test geometry, not in numerical expectation; the core test is a brittle binder-bridge failure mode. Binder viscosity influences adhesion by controlling binder bridge geometry at contact points between sand grains: at a 0.8 wt % addition rate, a low-viscosity binder produces a thin meniscus that may fail cohesively at 1.0–1.5 N/mm², whereas a high-viscosity binder can create a bulk intergranular plug that raises immediate strength but introduces gas evolution defects.

How Does Solvent Evaporation During Blow Head Transit Produce Resin-Rich Fines Adhesion?

During horizontal split-box shooting, the mixed sand-binder mass is propelled from the blow head into the core box at 4.5–6.0 bar in less than 0.8 seconds. Solvent evaporation begins in the blow head at the venturi expansion, where the pressure drop across the sand nozzle causes local cooling followed by entrainment of ambient air at 30–45 % RH. Because the solvent blend boils across a 140–200 °C range, the higher-vapour-pressure aromatic fraction is stripped preferentially, leaving behind the dibasic ester and polar fractions that raise the remaining binder phase viscosity from the neat value of 180–350 mPa·s to an effective in-flight value above 1000 mPa·s at 25 °C by the time the sand reaches the tool surface. This in-flight thickening is intentional only up to the point of preserving pattern definition; beyond that point, resin-rich fines attach to box vents and tooling domes as a tacky skin. In a 40-L Laempe or Loramendi core machine, vent blockage becomes measurable when pressure loss across the box exceeds 1.5 bar above the clean-box baseline, and the root cause is frequently a pre-shot viscosity drift in Part I combined with less than 0.35 wt % total binder due to sand feed fluctuation. The corrective action is not simply solvent addition, because solvent addition below the manufacturer’s stated lower flammability limit may reduce viscosity but also collapses the sand–binder adhesion by increasing solvent entrapment during gassing.

High-shear mixer measurements on twin-screw continuous sand mixers with 60–120 kg/min output and rotor-tip speeds of 4–8 m/s indicate that cold box binders exhibit shear-thinning behaviour within the first 5–10 seconds of contact with sand. Apparent viscosity at 100 s⁻¹ may be 80–120 mPa·s for a Part I that reads 250 mPa·s at 60 min⁻¹ in a Brookfield spindle 4. That shear-rate sensitivity is governed by oligomer chain entanglement and solvent cluster disruption; it collapses when the rotor-tip clearance exceeds 2 mm because the binder is not exposed to the same deformation history. On production lines, the fastest degradation of sand adhesion occurs when the binder is delivered by gear pumps calibrated at 0.5–1.0 % volumetric tolerance but the static mixer elements are worn to a 5 mm clearance, reducing the energy input into the resin film. In that state, the binder remains as islands that are visible under 40× magnification as glossy patches on the sand surface, and the cured core strength drops to 1.0–1.3 N/mm² even though the as-received viscosity at 25 °C remains within specification. This mismatch is captured by measuring viscosity after a controlled shear history in a cone-and-plate rheometer at 1000 s⁻¹ for 60 s, such as a TA Instruments AR 2000, rather than using the single-point Brookfield value.

Phenolic Polyol Part I Storage Viscosity Drift at 20–35 °C

Storage of Part I in unheated bulk tanks across a 20–35 °C diurnal range produces a nonlinear viscosity drift of 10–25 mPa·s per 5 °C, with the direction determined by oligomer condensation and atmospheric moisture uptake. At 20 °C, the material thickens due to reduced solvent mobility, while at 35 °C, slow advancement of the phenolic polyol can raise average molecular weight by 5–10 % over 30 days. Plant audits show that batch-to-batch viscosity variation of ±8 % is sufficient to change the required Part I pump speed by 2–3 % at a constant Part II ratio, and if the Part II is maintained at 30–150 mPa·s, the resulting ratio error remains below 0.02 mass fraction. For ISO 9001 traceability, the mixing line should record Brookfield viscosity at 25 °C per ASTM D2196-20 for every tote or drum, with a 24-hour pre-conditioning period in the mixing room. Binder heating to 30–35 °C is sometimes used to reduce viscosity, but this practice shortens bench life and accelerates solvent loss; production-scale experience in European automotive foundry core rooms indicates that a 5 °C increase in sand preheat from 25 °C to 30 °C may reduce Part I viscosity by 12–18 mPa·s but also increases amine consumption by up to 15 % due to faster pre-gassing evaporation.

Sand Grain Fineness and Wetting Envelope in Cold Box Adhesion Failure

Sand grain size distribution determines the total surface area available for binder adhesion and the spatial density of intergranular contact points. A 50–60 AFS grain fineness sand has a theoretical specific surface area of 120–160 cm²/g and a typical binder demand of 0.8–1.2 wt %, while a 65–80 AFS sand may require 1.0–1.4 wt % for equivalent tensile strength due to the increased surface area. With a 1.0 wt % addition on a 55 AFS silica, the calculated wet film thickness is approximately 0.6–0.8 μm if the binder density is 1.08–1.12 g/cm³ and the sand surface area is 140 cm²/g; any viscosity above 300 mPa·s at 25 °C tends to localise this film at the contact points rather than spreading it uniformly. The result is a dual failure mode: cohesive strength of binder bridges may reach 2.5 N/mm², but the regions between bridges remain uncoated and become crack propagation paths when the core is ejected at 2–4 bar. Angular sands with high crush resistance values of 25–35 N per grain, as measured by AFS crush resistance procedures, provide better mechanical interlocking but expose more fracture planes; their adhesion is more sensitive to binder viscosity because the wetting front must penetrate sharp-edged microgrooves of 5–15 μm depth.

When Reclaimed Sand Alkalinity Reaches pH 9.5 Before Binder Addition

Mechanical and thermal reclamation of sand after core shakeout produces a recycled sand fraction that is mixed with new sand at 60–85 % and must be controlled for pH, loss on ignition, and residual amine salt. When the reclaim stream exits the cooler at pH 9.5 or higher, the alkaline surface neutralises the acid catalyst components in the Part I phenolic polyol and shifts the cure reaction from the intended 10–20 s amine cycle to a slower, incomplete network. The adhesion deficit appears first as a soft core centre at 15–25 s after gassing, with immediate tensile strengths below 0.7 N/mm² and a friable edge that cannot hold a 0.5 mm radius. pH is measured by slurrying sand in deionised water at a 1:1 mass ratio per ASTM D2976-22 or equivalent; the acceptance limit for cold box reclaim is typically pH 6.8–8.2. If the reclaim pH exceeds pH 9.0, the binder viscosity at the sand surface is not the primary variable because the sand consumes available acidity faster than the binder can wet; the recommended correction is pre-neutralisation with dilute acetic acid or additional new sand dilution before mixer entry. A similar incompatibility occurs when the reclaimed sand retains sodium silicate residues above 0.3 wt %; these residues absorb moisture from the amine stream and create localized alkaline microenvironments that reduce binder–sand adhesion even when the bulk pH remains within specification.