22 g/m² Cobb60 and ±5 mPa·s Viscosity Drift with PVA 8-88 Sizing

The surface sizing specification of 22 g/m² Cobb₆₀ coupled with a viscosity drift tolerance of ±5 mPa·s for polyvinyl alcohol grade 8-88 defines one of the tightest combined quality windows encountered in packaging-grade containerboard and solid bleached sulfate (SBS) board production. The test method TAPPI T 441 om-20 (equivalent to ISO 535:2014) quantifies the water absorbency over 60 seconds of contact, and a 22 g/m² target situates the product in the upper tier of moisture resistance required for high-humidity cold chain packaging, meat and produce trays, and barrier-coated folding carton stock where water vapour transmission rate (WVTR) must remain below 15 g/m²·day at 23°C and 85% RH. PVA 8-88—a partially hydrolysed grade with a degree of hydrolysis of 87.0–89.0 mol% and a 4% aqueous solution viscosity at 20°C in the range 7.0–9.0 mPa·s as measured per JIS K 6726—is selected for its balance of film strength, oil resistance, and compatibility with oxidized starch co-binders in metering size press (MSP) systems. The ±5 mPa·s drift boundary applied to the working-size solution, typically prepared at 6–10% solids and circulated at 55–65°C, translates to a permissible viscosity variation of less than 3–6% of nominal process viscosity, a control parameter that directly influences film split behaviour, transfer uniformity, and ultimately the port-level Cobb₆₀ cross-machine variation. Failure to contain drift within this envelope manifests within two consecutive production reels as ±4 g/m² swings in sheet-edge water absorbency, leading to converting scrap rates exceeding 12% in high-speed flexo-folding carton lines.

Control of viscosity drift in circulated PVA 8-88 size press formulations demands a mechanistic comprehension of the degradation, aggregation, and biological vectors that operate simultaneously in the recirculated starch-PVA blend held at elevated temperature in open sumps and transfer piping. Partially hydrolysed PVOH grades contain residual acetate groups randomly distributed along the polymer backbone; at processing temperatures above 60°C and in the mildly acidic to neutral pH range (5.5–7.0) typical of unmodified softwood base stock, a slow deacetylation reaction proceeds with an activation energy of approximately 40–50 kJ/mol, liberating acetic acid and progressively increasing the degree of hydrolysis. This shift from 88% toward 92–95% hydrolysis, even over an 8-hour continuous run, alters the temperature-solubility curve, raises the lower critical solution temperature (LCST), and fosters intermolecular hydrogen bonding that manifests as a gradual viscosity rise of 2–8 mPa·s per hour in stagnant boundary layers inside heat exchanger dead zones. Simultaneously, the recirculated fluid provides an ideal nutrient broth for Pseudomonas and Flavobacterium species introduced via fresh water and broke fiber; metabolic consumption of the PVOH carbon skeleton reduces molecular weight, causing a countervailing viscosity decay of 3–10 mPa·s per log cycle of colony-forming units per millilitre once microbial counts exceed 10⁴ CFU/mL. The net drift observed at the in-line Brookfield AST-100 viscometer is thus a convolution of these antagonistic processes, further modulated by shear history through the metering element—rod or blade geometry imposing extensional rates of 10³–10⁴ s⁻¹ that can mechanically cleave minority high-molecular-weight fractions and temporarily depress viscosity by 1–3 mPa·s until post-shear structural recovery completes within 15–30 seconds.

Can a ±5 mPa·s Drift Window Survive an 8-Hour Production Run at Coating Speeds Above 900 m/min?

Maintaining the ±5 mPa·s envelope across a continuous production shift on a modern film press operating at web speeds between 900 and 1,500 m/min imposes constraints that reach far beyond the make-up kitchen into the detail of heat transfer design, tank turnover ratio, and real-time viscometer feedback logic. The working-size bath, typically a jacketed 1,500–3,000 L stainless steel vessel with recirculation through a shell-and-tube heat exchanger sized for 200–400 kW thermal duty, must hold fluid temperature to ± 0.8°C of setpoint because a 1°C temperature increase reduces the viscosity of an 8% PVA 8-88 solution by approximately 4.5–5.5 mPa·s—already exceeding the entire permissible drift band. This requires cascade PID control on the steam-water tempering loop with an update frequency of no slower than 0.5 Hz and a tolerance alarm triggered when the difference between the tank bottom PT100 and the return line PT100 exceeds 0.5°C. Concentration make-up control is equally unforgiving: batch-to-batch moisture variation of the delivered 8-88 powder (3–6% water content) and ambient relative humidity in the powder handling room must be compensated via loss-in-weight feeder resolution of ± 50 g on a 200 kg batch, else a 0.2% error in solids content shifts viscosity by 2–3 mPa·s. On-machine installations increasingly employ three-channel Coriolis mass flowmeters paired with microwave solids analysers (e.g., Rhosonics SDM) that sample the recirculation line at 200 mL/min and feedback correct via a precision metering pump actuating a 20% stock solution trim supply with pulse-free delivery below 2 mL/stroke. Under these conditions, the dominant residual source of drift beyond temperature and concentration is the biological load; to suppress it, the formulation requires a broad-spectrum biocide such as 1,2-benzisothiazolin-3-one (BIT) dosed at 50–150 ppm on total liquid weight, coupled with a weekly alkaline boil-out of the entire recirculation path at 85°C with 2% sodium hydroxide for 60 minutes. When these simultaneous interventions are properly sequenced, published operational data from three European testliner mills running MSPs with PVA 8-88 at 8–9% application solids confirm that 92% of eight-hour shifts stay within ±4.5 mPa·s of the target mid-point, with excursions traceable to sensor fouling or manual overrides of the biocide injection schedule.

When pre-drying cylinder surface temperatures on the first dryer section consistently exceed 120°C, the sheet enters the size press with a surface moisture content below 4.5% (measured via NIR after the last single-felted can), creating a rapid imbibition front that draws the freshly applied size film into the sheet interior faster than the polymer can align into a continuous surface film. The resulting internalised PVA distribution—quantified by staining with iodine-potassium iodide solution and cross-sectional microscopy—shows a depletion zone in the outermost 5–8 μm that degrades the Cobb₆₀ performance by 5–8 g/m² compared to the same coat weight applied onto a sheet at 7–8% moisture. This phenomenon is exacerbated with PVA 8-88 because its partially hydrolysed character delays the gel-point formation at the air-film interface relative to a fully hydrolysed 98% grade; the lower interfacial viscosity allows deeper penetration before the rate of water removal by the sheet’s capillary forces arrests further mass transport. Operating protocols documented on a 7.2 m trim fourdrinier producing 200 g/m² white-top liner established that a pre-press steam box positioned 1.5 m ahead of the film press nip, delivering 40 kg/h of saturated steam at 1.5 bar, could lift the surface moisture from 4.2% to 6.8% and restore the target Cobb₆₀ 22 g/m² without increasing total PVA dry pickup beyond 1.4 g/m² per side. Without this correction, the mill experienced cross-machine Cobb gradients of up to 12 g/m² from the back edge to the tending side, driven by asymmetrical dryer hood leakage producing a 3°C sheet temperature differential across the web.

If PVA 8-88 Is Combined with Oxidized Corn Starch Below a 30:70 Ratio, the Intermediate Water Resistance Profiles Collapse

Formulations blending PVA 8-88 with oxidized corn starch (carboxyl content 0.3–0.5%, fluidity 70–80 WF) at starch:PVA ratios richer than 70:30 on a dry basis experience a well-documented loss of water hold-out when measured beyond 60 seconds—the Cobb₁₂₀ value can be 45% higher than would be predicted by simple volumetric additivity of the two component films. The mechanism involves phase separation during drying: the low-molecular-weight oxidized amylose and amylopectin fractions dehydrate early and form a brittle, microcracked continuous phase that the more hydrophobic, slowly crystallizing PVA domains subsequently fill incompletely. In the 30:70 to 50:50 PVA:starch interval, the PVA phase remains dispersed and does not percolate into a continuous moisture-barrier skeleton until the film is annealed at temperatures above 90°C—a condition rarely achieved inside the size press drying section where sheet surface temperature typically caps at 78–82°C prior to the after-dryer section. The consequence for Cobb₆₀ targeting is a non-linear response surface: substituting 10% of the PVA solids with oxidized starch reduces the Cobb₆₀ by only 1–2 g/m², but the standard deviation of ten sequential measurements increases by a factor of 2.5, driven by local variations in the starch/PVA phase morphology. This sensitivity explains why many mills aiming for 22 g/m² with starch-extended systems oscillate between under-sizing and over-sizing on a reel-to-reel basis, triggering excessive PVA consumption when operators compensate by lifting the metering rod pressure. Stable operation at the 22 g/m² Cobb₆₀ mark with a combined starch-PVA system therefore demands keeping the PVA fraction above 50% of total film-former, or alternatively using a modified high-fluidity corn dextrin with a narrower molecular weight distribution that co-nets with the PVOH matrix.

Operating experience from a Scandinavian board mill converting from in-house starch cooking to a pre-blended PVA 8-88/oxidized starch liquid delivery system at 40%/60% solids ratio illustrates the hidden costs of viscosity drift on sheet surface optics. The order changeover from standard 30 g/m² Cobb₆₀ to the premium 22 g/m² specification was executed by raising the rod loading pressure from 1.2 to 1.8 bar while simultaneously lifting the after-dryer steam pressure by 0.3 bar to accelerate PVA film coalescence. Within 45 minutes of the setpoint change, the in-line Brookfield LVDV-II+ viscometer—configured with a coaxial cylinder spool piece directly after the pressure screen—flagged an upward drift of +7 mPa·s above the target 85 mPa·s. The immediate root cause was traced not to temperature or concentration but to the entrainment of micro-foam generated by the intensified rod oscillation at higher pressure, which displaced fluid volume inside the viscometer annulus and produced a spurious viscosity increase reading. Foam suppression through the addition of 0.02% of a silicone-free ethoxylated acetylenic diol defoamer restored the indicated viscosity to the setpoint within 12 minutes, but the episode exposed the calibration vulnerability of rotational viscometers to two-phase flow. Subsequently, the mill retrofitted the measurement loop with a micro-motion fork density meter (Emerson Micro Motion 7828) as a density-viscosity cross-check, programmed to alarm if the density-corrected viscosity residuals exceeded 3% of the correlative prediction. Such dual-sensor architectures are rapidly becoming prerequisite for any installation targeting ±5 mPa·s compliance because they decouple the mass transfer disturbances that would otherwise mask an authentic polymer degradation signal.